From d122d5c987b8b13483190acf9838535298e585f5 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Fri, 3 Jan 2025 21:57:55 -0500
Subject: [PATCH 01/23] q6_k scale caching

---
 .../vulkan-shaders/mul_mat_vec_q6_k.comp      | 37 +++++++++----------
 1 file changed, 17 insertions(+), 20 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
index 70e13a56bd730..a2a3623869f12 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
@@ -6,22 +6,22 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
+shared FLOAT_TYPE sccache[BLOCK_SIZE/16][16];
+
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
 
     const uint num_blocks_per_row = p.ncols / QUANT_K;
 
-    // 16 threads are used to process each block
+    // 16 thread groups are used to process each block
     const uint it_size = gl_WorkGroupSize.x/16;
     const uint tid = gl_LocalInvocationID.x;
-    const uint itid = tid%16;  // 0...16
-    const uint ix  = tid/16;
-
-    const uint step = 8;
+    const uint itid = tid%16;  // 0...15
+    const uint ix = tid/16;
 
-    const uint v_im = itid/step;                            // 0 or 1. 0 computes 0..., 1 computes 128...
-    const uint v_in = itid - step*v_im;                     // 0...15 or 0...7
+    const uint v_im = itid/8;                               // 0 or 1. 0 computes 0..., 1 computes 128...
+    const uint v_in = itid - 8*v_im;                        // 0...15 or 0...7
 
     const uint l0 = 4 * v_in;                               // 0, 4, 8, ..., 28
     const uint is = v_in / 4;
@@ -46,11 +46,8 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
             const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
 
-            FLOAT_TYPE scales[4];
-            scales[0] = FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 0]);
-            scales[1] = FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 2]);
-            scales[2] = FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 4]);
-            scales[3] = FLOAT_TYPE(data_a[ib0 + i].scales[s_offset + 6]);
+            sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]);
+            barrier();
 
             uint32_t ql0_u32 =  uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 1]) << 16);
             uint32_t ql32_u32 = uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 16]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 17]) << 16);
@@ -63,7 +60,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             uint32_t qh_u32 = uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2 + 1]) << 16);
             uint32_t qh0_u32 = (qh_u32 & 0x03030303) << 4;
             uint32_t qh2_u32 = (qh_u32 & 0x0C0C0C0C) << 2;
-            uint32_t qh4_u32 = (qh_u32 & 0x30303030) << 0;
+            uint32_t qh4_u32 = (qh_u32 & 0x30303030);
             uint32_t qh6_u32 = (qh_u32 & 0xC0C0C0C0) >> 2;
 
             uint32_t q0_u32 = ql0_u32_lo4  | qh0_u32;
@@ -82,14 +79,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
                 B_TYPE_VEC4 by64 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 16];
                 B_TYPE_VEC4 by96 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 24];
 
-                FLOAT_TYPE sum = FLOAT_TYPE(0.0);
-                [[unroll]] for (int l = 0; l < 4; ++l) {
-                    sum = fma(FLOAT_TYPE(by0[l])  * scales[0], FLOAT_TYPE(int8_t(q0[l]) - 32),
-                          fma(FLOAT_TYPE(by32[l]) * scales[1], FLOAT_TYPE(int8_t(q1[l]) - 32),
-                          fma(FLOAT_TYPE(by64[l]) * scales[2], FLOAT_TYPE(int8_t(q2[l]) - 32),
-                          fma(FLOAT_TYPE(by96[l]) * scales[3], FLOAT_TYPE(int8_t(q3[l]) - 32), sum))));
+                FLOAT_TYPE sum[4] = {0, 0, 0, 0};
+                [[unroll]] for (uint l = 0; l < 4; ++l) {
+                    sum[0] = fma(FLOAT_TYPE(by0[l]), FLOAT_TYPE(int8_t(q0[l]) - 32), sum[0]);
+                    sum[1] = fma(FLOAT_TYPE(by32[l]), FLOAT_TYPE(int8_t(q1[l]) - 32), sum[1]);
+                    sum[2] = fma(FLOAT_TYPE(by64[l]), FLOAT_TYPE(int8_t(q2[l]) - 32), sum[2]);
+                    sum[3] = fma(FLOAT_TYPE(by96[l]), FLOAT_TYPE(int8_t(q3[l]) - 32), sum[3]);
                 }
-                temp[j][n] += sum * d;
+                temp[j][n] = fma(fma(sum[0], sccache[ix][s_offset], fma(sum[1], sccache[ix][s_offset + 2], fma(sum[2], sccache[ix][s_offset + 4], sum[3] * sccache[ix][s_offset + 6]))), d, temp[j][n]);
             }
         }
     }

From 6b06d1689011196ff3312277530402adefb53fbb Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 4 Jan 2025 13:32:44 -0500
Subject: [PATCH 02/23] 16 bit unpack

---
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      | 22 +++++++++----------
 .../vulkan-shaders/mul_mat_vec_q6_k.comp      |  2 +-
 2 files changed, 11 insertions(+), 13 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index cd1dd8e89c21e..225f0ce70a6fc 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -15,13 +15,11 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     // 16 threads are used to process each block
     const uint it_size = gl_WorkGroupSize.x/16;
     const uint tid = gl_LocalInvocationID.x;
-    const uint itid = tid%16;  // 0...16
-    const uint ix  = tid/16;
+    const uint itid = tid%16;  // 0...15
+    const uint ix = tid/16;
 
-    const uint step = 4;
-
-    const uint il = itid/step;                      // 0...3
-    const uint ir = itid - step*il;                 // 0...7 or 0...3
+    const uint il = itid/4;                         // 0...3
+    const uint ir = itid - 4*il;                    // 0...7 or 0...3
     const uint n =  4;
 
     const uint v_im = il / 2;  // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224
@@ -49,12 +47,12 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
             const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
 
-            uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
-            uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-            uvec4 scale0 = uvec4(unpack8(scale0_u32));
-            uvec4 scale4 = uvec4(unpack8(scale4_u32));
-            uvec4 scale8 = uvec4(unpack8(scale8_u32));
+            const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
+            const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
+            const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
+            uvec2 scale0 = uvec2(unpack8(scale0_u32));
+            uvec2 scale4 = uvec2(unpack8(scale4_u32));
+            uvec2 scale8 = uvec2(unpack8(scale8_u32));
 
             const uint32_t sc0 = (  scale0.x       & 0x3f);
             const uint32_t sc1 = (  scale0.y       & 0x3f);
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
index a2a3623869f12..e1afd55e0f540 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
@@ -14,7 +14,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
     const uint num_blocks_per_row = p.ncols / QUANT_K;
 
-    // 16 thread groups are used to process each block
+    // 16 threads are used to process each block
     const uint it_size = gl_WorkGroupSize.x/16;
     const uint tid = gl_LocalInvocationID.x;
     const uint itid = tid%16;  // 0...15

From 21c6b805c99d90332250adb969e128171d431525 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 4 Jan 2025 14:57:28 -0500
Subject: [PATCH 03/23] q4_k test (slow)

---
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      | 51 ++++++++++++-------
 1 file changed, 34 insertions(+), 17 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index 225f0ce70a6fc..ee439f5c4ca31 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -6,6 +6,8 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
+shared FLOAT_TYPE sc[BLOCK_SIZE/16][2][8];
+
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
@@ -25,6 +27,16 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint v_im = il / 2;  // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224
     const uint v_in = il % 2;
 
+	// v_im may be incorrect
+	const uint itid8 = itid%8;
+	const uint a = (itid8 > 5) ? 4 : 0;
+	const uint b = (itid8 > 3) ? 0x0f : 0x3f;
+	const uint c = (itid8 > 3) ? 0xc0 : 0xFF;
+	const uint dd = (itid8 > 3) ? 2 : 0;
+	const uint gg = (itid8 > 1) ? 1 : 0;
+	const uint g = (itid8 > 3) ? 8 : gg;
+	const uint h = itid8%2;
+
     const uint l0 = n * (2 * ir + v_in);            // 0...15
     const uint q_offset = 32*v_im + l0;
     const uint y_offset = 64*v_im + l0;
@@ -50,18 +62,23 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
             const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
             const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-            uvec2 scale0 = uvec2(unpack8(scale0_u32));
-            uvec2 scale4 = uvec2(unpack8(scale4_u32));
-            uvec2 scale8 = uvec2(unpack8(scale8_u32));
-
-            const uint32_t sc0 = (  scale0.x       & 0x3f);
-            const uint32_t sc1 = (  scale0.y       & 0x3f);
-            const uint32_t sc2 = (  scale4.x       & 0x3f);
-            const uint32_t sc3 = (  scale4.y       & 0x3f);
-            const uint32_t sc4 = (( scale8.x       & 0x0f) | ((scale0.x & 0xc0) >> 2));
-            const uint32_t sc5 = (( scale8.y       & 0x0f) | ((scale0.y & 0xc0) >> 2));
-            const uint32_t sc6 = (((scale8.x >> 4) & 0x0f) | ((scale4.x & 0xc0) >> 2));
-            const uint32_t sc7 = (((scale8.y >> 4) & 0x0f) | ((scale4.y & 0xc0) >> 2));
+
+			uvec2 scale[3];
+            scale[0] = uvec2(unpack8(scale0_u32));
+            scale[1] = uvec2(unpack8(scale4_u32));
+            scale[2] = uvec2(unpack8(scale8_u32));
+
+			sc[ix][v_im][itid8] = FLOAT_TYPE(((scale[gg][h] >> a) & b) | ((scale[g][h] & c) >> dd));
+			barrier();
+
+            //const uint32_t sc0 = (  scale0.x       & 0x3f);
+            //const uint32_t sc1 = (  scale0.y       & 0x3f);
+            //const uint32_t sc2 = (  scale4.x       & 0x3f);
+            //const uint32_t sc3 = (  scale4.y       & 0x3f);
+            //const uint32_t sc4 = (( scale8.x       & 0x0f) | ((scale0.x & 0xc0) >> 2));
+            //const uint32_t sc5 = (( scale8.y       & 0x0f) | ((scale0.y & 0xc0) >> 2));
+            //const uint32_t sc6 = (((scale8.x >> 4) & 0x0f) | ((scale4.x & 0xc0) >> 2));
+            //const uint32_t sc7 = (((scale8.y >> 4) & 0x0f) | ((scale4.y & 0xc0) >> 2));
 
             uint32_t qs0_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4];
             uint32_t qs64_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4 + 16];
@@ -104,11 +121,11 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
                 const FLOAT_TYPE sz = fma(FLOAT_TYPE(by20.x),      q4_8,  fma(FLOAT_TYPE(by20.y),  q4_9,  fma(FLOAT_TYPE(by20.z),  q4_10, FLOAT_TYPE(by20.w) *  q4_11)));
                 const FLOAT_TYPE sw = fma(FLOAT_TYPE(by232.x),     q4_12, fma(FLOAT_TYPE(by232.y), q4_13, fma(FLOAT_TYPE(by232.z), q4_14, FLOAT_TYPE(by232.w) * q4_15)));
                 const FLOAT_TYPE smin =
-                    fma(FLOAT_TYPE(by10.x), sc2, fma(FLOAT_TYPE(by132.x), sc3, fma(FLOAT_TYPE(by20.x), sc6, fma(FLOAT_TYPE(by232.x), sc7,
-                    fma(FLOAT_TYPE(by10.y), sc2, fma(FLOAT_TYPE(by132.y), sc3, fma(FLOAT_TYPE(by20.y), sc6, fma(FLOAT_TYPE(by232.y), sc7,
-                    fma(FLOAT_TYPE(by10.z), sc2, fma(FLOAT_TYPE(by132.z), sc3, fma(FLOAT_TYPE(by20.z), sc6, fma(FLOAT_TYPE(by232.z), sc7,
-                    fma(FLOAT_TYPE(by10.w), sc2, fma(FLOAT_TYPE(by132.w), sc3, fma(FLOAT_TYPE(by20.w), sc6,     FLOAT_TYPE(by232.w) * sc7)))))))))))))));
-                temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n]));
+                    fma(FLOAT_TYPE(by10.x), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.x), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.x), sc[ix][v_im][6], fma(FLOAT_TYPE(by232.x), sc[ix][v_im][7],
+                    fma(FLOAT_TYPE(by10.y), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.y), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.y), sc[ix][v_im][6], fma(FLOAT_TYPE(by232.y), sc[ix][v_im][7],
+                    fma(FLOAT_TYPE(by10.z), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.z), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.z), sc[ix][v_im][6], fma(FLOAT_TYPE(by232.z), sc[ix][v_im][7],
+                    fma(FLOAT_TYPE(by10.w), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.w), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.w), sc[ix][v_im][6],     FLOAT_TYPE(by232.w) * sc[ix][v_im][7])))))))))))))));
+                temp[j][n] = fma(dall, fma(sx, sc[ix][v_im][0], fma(sy, sc[ix][v_im][1], fma(sz, sc[ix][v_im][4], sw * sc[ix][v_im][5]))), fma(-dmin, smin, temp[j][n]));
             }
         }
     }

From b0e4ccbeb95f3987052fbc08500459773216a691 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 4 Jan 2025 14:57:40 -0500
Subject: [PATCH 04/23] revert it

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      | 16 +++---
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      | 12 ++---
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      | 51 +++++++------------
 .../vulkan-shaders/mul_mat_vec_q5_k.comp      | 12 ++---
 4 files changed, 35 insertions(+), 56 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 9342134462416..921759bc8bed2 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -14,13 +14,11 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     // 16 threads are used to process each block
     const uint it_size = gl_WorkGroupSize.x/16;
     const uint tid = gl_LocalInvocationID.x;
-    const uint itid = tid%16;  // 0...16
-    const uint ix  = tid/16;
+    const uint itid = tid%16;  // 0...15
+    const uint ix = tid/16;
 
-    const uint step = 8;
-
-    const uint v_im = itid/step;                             // 0 or 1. 0 computes 0..., 1 computes 128...
-    const uint v_in = itid - step*v_im;                      // 0...15 or 0...7
+    const uint v_im = itid/8;                                // 0 or 1. 0 computes 0..., 1 computes 128...
+    const uint v_in = itid - 8*v_im;                         // 0...15 or 0...7
 
     const uint l0 = 2*v_in;                                  // 0...15
     const uint q_offset = 32*v_im + l0;
@@ -44,7 +42,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE dall = d.x;
             const FLOAT_TYPE dmin = d.y;
 
-            uint32_t s0_u32 = data_a_packed32[ib0 + i].scales[s_offset / 4 + 0];
+            uint32_t s0_u32 = data_a_packed32[ib0 + i].scales[s_offset / 4];
             uint32_t s4_u32 = data_a_packed32[ib0 + i].scales[s_offset / 4 + 1];
 
             uint32_t s0_lo4_u32 = s0_u32 & 0x0F0F0F0F;
@@ -57,13 +55,13 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             uvec4 s0_hi4 = uvec4(unpack8(s0_hi4_u32));
             uvec4 s4_hi4 = uvec4(unpack8(s4_hi4_u32));
 
-            uint16_t qs0_u16 = data_a_packed16[ib0 + i].qs[q_offset / 2 + 0];
+            uint16_t qs0_u16 = data_a_packed16[ib0 + i].qs[q_offset / 2];
             uint16_t qs16_u16 = data_a_packed16[ib0 + i].qs[q_offset / 2 + 8];
             uvec2 qs0 =  uvec2(unpack8(qs0_u16));
             uvec2 qs16 = uvec2(unpack8(qs16_u16));
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-                B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 0];
+                B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
                 B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
                 B_TYPE_VEC2 b32 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16];
                 B_TYPE_VEC2 b48 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24];
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index 86b0159d97a89..b9a87f13a99ed 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -14,13 +14,11 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     // 16 threads are used to process each block
     const uint it_size = gl_WorkGroupSize.x/16;
     const uint tid = gl_LocalInvocationID.x;
-    const uint itid = tid%16;  // 0...16
-    const uint ix  = tid/16;
+    const uint itid = tid%16;  // 0...15
+    const uint ix = tid/16;
 
-    const uint step = 8;
-
-    const uint v_im = itid/step;                            // 0 or 1. 0 computes 0..., 1 computes 128...
-    const uint v_in = itid - step*v_im;                     // 0...15 or 0...7
+    const uint v_im = itid/8;                               // 0 or 1. 0 computes 0..., 1 computes 128...
+    const uint v_in = itid - 8*v_im;                        // 0...15 or 0...7
 
     const uint8_t m = uint8_t(1 << (4 * v_im));
 
@@ -60,7 +58,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
 
-                B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 0];
+                B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
                 B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
                 B_TYPE_VEC2 b32 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16];
                 B_TYPE_VEC2 b48 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24];
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index ee439f5c4ca31..225f0ce70a6fc 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -6,8 +6,6 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
-shared FLOAT_TYPE sc[BLOCK_SIZE/16][2][8];
-
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
@@ -27,16 +25,6 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint v_im = il / 2;  // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224
     const uint v_in = il % 2;
 
-	// v_im may be incorrect
-	const uint itid8 = itid%8;
-	const uint a = (itid8 > 5) ? 4 : 0;
-	const uint b = (itid8 > 3) ? 0x0f : 0x3f;
-	const uint c = (itid8 > 3) ? 0xc0 : 0xFF;
-	const uint dd = (itid8 > 3) ? 2 : 0;
-	const uint gg = (itid8 > 1) ? 1 : 0;
-	const uint g = (itid8 > 3) ? 8 : gg;
-	const uint h = itid8%2;
-
     const uint l0 = n * (2 * ir + v_in);            // 0...15
     const uint q_offset = 32*v_im + l0;
     const uint y_offset = 64*v_im + l0;
@@ -62,23 +50,18 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
             const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
             const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-
-			uvec2 scale[3];
-            scale[0] = uvec2(unpack8(scale0_u32));
-            scale[1] = uvec2(unpack8(scale4_u32));
-            scale[2] = uvec2(unpack8(scale8_u32));
-
-			sc[ix][v_im][itid8] = FLOAT_TYPE(((scale[gg][h] >> a) & b) | ((scale[g][h] & c) >> dd));
-			barrier();
-
-            //const uint32_t sc0 = (  scale0.x       & 0x3f);
-            //const uint32_t sc1 = (  scale0.y       & 0x3f);
-            //const uint32_t sc2 = (  scale4.x       & 0x3f);
-            //const uint32_t sc3 = (  scale4.y       & 0x3f);
-            //const uint32_t sc4 = (( scale8.x       & 0x0f) | ((scale0.x & 0xc0) >> 2));
-            //const uint32_t sc5 = (( scale8.y       & 0x0f) | ((scale0.y & 0xc0) >> 2));
-            //const uint32_t sc6 = (((scale8.x >> 4) & 0x0f) | ((scale4.x & 0xc0) >> 2));
-            //const uint32_t sc7 = (((scale8.y >> 4) & 0x0f) | ((scale4.y & 0xc0) >> 2));
+            uvec2 scale0 = uvec2(unpack8(scale0_u32));
+            uvec2 scale4 = uvec2(unpack8(scale4_u32));
+            uvec2 scale8 = uvec2(unpack8(scale8_u32));
+
+            const uint32_t sc0 = (  scale0.x       & 0x3f);
+            const uint32_t sc1 = (  scale0.y       & 0x3f);
+            const uint32_t sc2 = (  scale4.x       & 0x3f);
+            const uint32_t sc3 = (  scale4.y       & 0x3f);
+            const uint32_t sc4 = (( scale8.x       & 0x0f) | ((scale0.x & 0xc0) >> 2));
+            const uint32_t sc5 = (( scale8.y       & 0x0f) | ((scale0.y & 0xc0) >> 2));
+            const uint32_t sc6 = (((scale8.x >> 4) & 0x0f) | ((scale4.x & 0xc0) >> 2));
+            const uint32_t sc7 = (((scale8.y >> 4) & 0x0f) | ((scale4.y & 0xc0) >> 2));
 
             uint32_t qs0_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4];
             uint32_t qs64_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4 + 16];
@@ -121,11 +104,11 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
                 const FLOAT_TYPE sz = fma(FLOAT_TYPE(by20.x),      q4_8,  fma(FLOAT_TYPE(by20.y),  q4_9,  fma(FLOAT_TYPE(by20.z),  q4_10, FLOAT_TYPE(by20.w) *  q4_11)));
                 const FLOAT_TYPE sw = fma(FLOAT_TYPE(by232.x),     q4_12, fma(FLOAT_TYPE(by232.y), q4_13, fma(FLOAT_TYPE(by232.z), q4_14, FLOAT_TYPE(by232.w) * q4_15)));
                 const FLOAT_TYPE smin =
-                    fma(FLOAT_TYPE(by10.x), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.x), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.x), sc[ix][v_im][6], fma(FLOAT_TYPE(by232.x), sc[ix][v_im][7],
-                    fma(FLOAT_TYPE(by10.y), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.y), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.y), sc[ix][v_im][6], fma(FLOAT_TYPE(by232.y), sc[ix][v_im][7],
-                    fma(FLOAT_TYPE(by10.z), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.z), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.z), sc[ix][v_im][6], fma(FLOAT_TYPE(by232.z), sc[ix][v_im][7],
-                    fma(FLOAT_TYPE(by10.w), sc[ix][v_im][2], fma(FLOAT_TYPE(by132.w), sc[ix][v_im][3], fma(FLOAT_TYPE(by20.w), sc[ix][v_im][6],     FLOAT_TYPE(by232.w) * sc[ix][v_im][7])))))))))))))));
-                temp[j][n] = fma(dall, fma(sx, sc[ix][v_im][0], fma(sy, sc[ix][v_im][1], fma(sz, sc[ix][v_im][4], sw * sc[ix][v_im][5]))), fma(-dmin, smin, temp[j][n]));
+                    fma(FLOAT_TYPE(by10.x), sc2, fma(FLOAT_TYPE(by132.x), sc3, fma(FLOAT_TYPE(by20.x), sc6, fma(FLOAT_TYPE(by232.x), sc7,
+                    fma(FLOAT_TYPE(by10.y), sc2, fma(FLOAT_TYPE(by132.y), sc3, fma(FLOAT_TYPE(by20.y), sc6, fma(FLOAT_TYPE(by232.y), sc7,
+                    fma(FLOAT_TYPE(by10.z), sc2, fma(FLOAT_TYPE(by132.z), sc3, fma(FLOAT_TYPE(by20.z), sc6, fma(FLOAT_TYPE(by232.z), sc7,
+                    fma(FLOAT_TYPE(by10.w), sc2, fma(FLOAT_TYPE(by132.w), sc3, fma(FLOAT_TYPE(by20.w), sc6,     FLOAT_TYPE(by232.w) * sc7)))))))))))))));
+                temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n]));
             }
         }
     }
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
index 0a68891c35a5f..bea1dc8608f6b 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
@@ -15,8 +15,8 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     // 16 threads are used to process each block
     const uint it_size = gl_WorkGroupSize.x/16;
     const uint tid = gl_LocalInvocationID.x;
-    const uint itid = tid%16;  // 0...16
-    const uint ix  = tid/16;
+    const uint itid = tid%16;  // 0...15
+    const uint ix = tid/16;
 
     const uint il = itid/4;                          // 0...3
     const uint ir = itid - 4*il;                     // 0...7 or 0...3
@@ -49,9 +49,9 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
             uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
             uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-            uvec4 scale0 = uvec4(unpack8(scale0_u32));
-            uvec4 scale4 = uvec4(unpack8(scale4_u32));
-            uvec4 scale8 = uvec4(unpack8(scale8_u32));
+            uvec2 scale0 = uvec2(unpack8(scale0_u32));
+            uvec2 scale4 = uvec2(unpack8(scale4_u32));
+            uvec2 scale8 = uvec2(unpack8(scale8_u32));
 
             const uint32_t sc0 = (  scale0.x       & 0x3f);
             const uint32_t sc1 = (  scale0.y       & 0x3f);
@@ -74,7 +74,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
             uint32_t qs0_16_lo4_offset16 = ((qh >> (2*v_im)) & 0x01010101) << 4;
             uint32_t qs0_16_hi4_offset16 = ((qh >> (2*v_im)) & 0x02020202) << 3;
-            uint32_t qs64_80_lo4_offset16 = ((qh >> (2*v_im)) & 0x10101010) << 0;
+            uint32_t qs64_80_lo4_offset16 = ((qh >> (2*v_im)) & 0x10101010);
             uint32_t qs64_80_hi4_offset16 = ((qh >> (2*v_im)) & 0x20202020) >> 1;
 
             qs0_16_u32_lo4 += qs0_16_lo4_offset16;

From 07d0d58bef57366233b52c421632cfb2e54c76d4 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 4 Jan 2025 16:24:29 -0500
Subject: [PATCH 05/23] q3_k

---
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      | 36 ++++++++-----------
 1 file changed, 14 insertions(+), 22 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index b9a87f13a99ed..0266417a88f72 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -5,6 +5,9 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
+shared uint32_t scu8[BLOCK_SIZE/16][12];
+shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][12];
+
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
@@ -16,6 +19,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint tid = gl_LocalInvocationID.x;
     const uint itid = tid%16;  // 0...15
     const uint ix = tid/16;
+    const uint itid8 = itid%8;
 
     const uint v_im = itid/8;                               // 0 or 1. 0 computes 0..., 1 computes 128...
     const uint v_in = itid - 8*v_im;                        // 0...15 or 0...7
@@ -34,8 +38,6 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
         }
     }
 
-    const uint s_shift = 4 * v_im;
-
     [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) {
         const uint y_idx = i * QUANT_K + y_offset;
 
@@ -43,18 +45,8 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
             const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
 
-            uint16_t s0_16 = data_a_packed16[ib0 + i].scales[0];
-            uint16_t s2_16 = data_a_packed16[ib0 + i].scales[1];
-            uint16_t s4_16 = data_a_packed16[ib0 + i].scales[2];
-            uint16_t s6_16 = data_a_packed16[ib0 + i].scales[3];
-            uint16_t s8_16 = data_a_packed16[ib0 + i].scales[4];
-            uint16_t s10_16 = data_a_packed16[ib0 + i].scales[5];
-            u8vec2 s0 = unpack8(s0_16);
-            u8vec2 s2 = unpack8(s2_16);
-            u8vec2 s4 = unpack8(s4_16);
-            u8vec2 s6 = unpack8(s6_16);
-            u8vec2 s8 = unpack8(s8_16);
-            u8vec2 s10 = unpack8(s10_16);
+            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> 4*v_im) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid%4+8]) >> (4*v_im + 2*(itid8/4)) & 0x3) << 4)) - 32);
+            barrier();
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
 
@@ -69,14 +61,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
                 FLOAT_TYPE sum = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum = fma(FLOAT_TYPE(b0[l])   * FLOAT_TYPE(int8_t(((s0[0] >> s_shift) & 0xF) | ((s8[0]  >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 0)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b32[l])  * FLOAT_TYPE(int8_t(((s2[0] >> s_shift) & 0xF) | ((s10[0] >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 1)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b64[l])  * FLOAT_TYPE(int8_t(((s4[0] >> s_shift) & 0xF) | ((s8[0]  >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 2)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b96[l])  * FLOAT_TYPE(int8_t(((s6[0] >> s_shift) & 0xF) | ((s10[0] >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 3)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b16[l])  * FLOAT_TYPE(int8_t(((s0[1] >> s_shift) & 0xF) | ((s8[1]  >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 0)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b48[l])  * FLOAT_TYPE(int8_t(((s2[1] >> s_shift) & 0xF) | ((s10[1] >> (s_shift + 0) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 1)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b80[l])  * FLOAT_TYPE(int8_t(((s4[1] >> s_shift) & 0xF) | ((s8[1]  >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 2)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b112[l]) * FLOAT_TYPE(int8_t(((s6[1] >> s_shift) & 0xF) | ((s10[1] >> (s_shift + 2) & 0x3) << 4)) - 32), FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 3)) != 0) ? 0 : 4)), sum))))))));
+                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m    )) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m    )) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 1)) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 1)) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 2)) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 2)) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 3)) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 3)) != 0) ? 0 : 4)), sum))))))));
                 }
                 temp[j][n] = fma(d, sum, temp[j][n]);
             }

From d70a731639d9acd0baa588a94bfaa5f928b26b9c Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 4 Jan 2025 20:48:27 -0500
Subject: [PATCH 06/23] q2_k

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      | 50 ++++++++-----------
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      |  5 +-
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      |  6 +--
 3 files changed, 27 insertions(+), 34 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 921759bc8bed2..098771493fbe1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -5,6 +5,8 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
+shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][16];
+
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
@@ -16,6 +18,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint tid = gl_LocalInvocationID.x;
     const uint itid = tid%16;  // 0...15
     const uint ix = tid/16;
+    const uint itid8 = itid%8;
 
     const uint v_im = itid/8;                                // 0 or 1. 0 computes 0..., 1 computes 128...
     const uint v_in = itid - 8*v_im;                         // 0...15 or 0...7
@@ -42,18 +45,9 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE dall = d.x;
             const FLOAT_TYPE dmin = d.y;
 
-            uint32_t s0_u32 = data_a_packed32[ib0 + i].scales[s_offset / 4];
-            uint32_t s4_u32 = data_a_packed32[ib0 + i].scales[s_offset / 4 + 1];
-
-            uint32_t s0_lo4_u32 = s0_u32 & 0x0F0F0F0F;
-            uint32_t s0_hi4_u32 = (s0_u32 >> 4) & 0x0F0F0F0F;
-            uint32_t s4_lo4_u32 = s4_u32 & 0x0F0F0F0F;
-            uint32_t s4_hi4_u32 = (s4_u32 >> 4) & 0x0F0F0F0F;
-
-            uvec4 s0_lo4 = uvec4(unpack8(s0_lo4_u32));
-            uvec4 s4_lo4 = uvec4(unpack8(s4_lo4_u32));
-            uvec4 s0_hi4 = uvec4(unpack8(s0_hi4_u32));
-            uvec4 s4_hi4 = uvec4(unpack8(s4_hi4_u32));
+            sccache[ix][0][itid] = FLOAT_TYPE(bitfieldExtract(uint(data_a[ib0 + i].scales[itid8]), int(v_im*4), 4)); // lower 8 bytes
+            sccache[ix][1][itid] = FLOAT_TYPE(bitfieldExtract(uint(data_a[ib0 + i].scales[itid8+8]), int(v_im*4), 4)); // upper 8 bytes
+            barrier();
 
             uint16_t qs0_u16 = data_a_packed16[ib0 + i].qs[q_offset / 2];
             uint16_t qs16_u16 = data_a_packed16[ib0 + i].qs[q_offset / 2 + 8];
@@ -73,22 +67,22 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
                 FLOAT_TYPE sum1 = FLOAT_TYPE(0.0);
                 FLOAT_TYPE sum2 = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum1 = fma(FLOAT_TYPE(b0[l]),   FLOAT_TYPE(s0_lo4[0]) * FLOAT_TYPE((qs0[l]  >> 0) & 3),
-                           fma(FLOAT_TYPE(b16[l]),  FLOAT_TYPE(s0_lo4[1]) * FLOAT_TYPE((qs16[l] >> 0) & 3),
-                           fma(FLOAT_TYPE(b32[l]),  FLOAT_TYPE(s0_lo4[2]) * FLOAT_TYPE((qs0[l]  >> 2) & 3),
-                           fma(FLOAT_TYPE(b48[l]),  FLOAT_TYPE(s0_lo4[3]) * FLOAT_TYPE((qs16[l] >> 2) & 3),
-                           fma(FLOAT_TYPE(b64[l]),  FLOAT_TYPE(s4_lo4[0]) * FLOAT_TYPE((qs0[l]  >> 4) & 3),
-                           fma(FLOAT_TYPE(b80[l]),  FLOAT_TYPE(s4_lo4[1]) * FLOAT_TYPE((qs16[l] >> 4) & 3),
-                           fma(FLOAT_TYPE(b96[l]),  FLOAT_TYPE(s4_lo4[2]) * FLOAT_TYPE((qs0[l]  >> 6) & 3),
-                           fma(FLOAT_TYPE(b112[l]), FLOAT_TYPE(s4_lo4[3]) * FLOAT_TYPE((qs16[l] >> 6) & 3), sum1))))))));
-                    sum2 = fma(FLOAT_TYPE(b0[l]),   FLOAT_TYPE(s0_hi4[0]),
-                           fma(FLOAT_TYPE(b16[l]),  FLOAT_TYPE(s0_hi4[1]),
-                           fma(FLOAT_TYPE(b32[l]),  FLOAT_TYPE(s0_hi4[2]),
-                           fma(FLOAT_TYPE(b48[l]),  FLOAT_TYPE(s0_hi4[3]),
-                           fma(FLOAT_TYPE(b64[l]),  FLOAT_TYPE(s4_hi4[0]),
-                           fma(FLOAT_TYPE(b80[l]),  FLOAT_TYPE(s4_hi4[1]),
-                           fma(FLOAT_TYPE(b96[l]),  FLOAT_TYPE(s4_hi4[2]),
-                           fma(FLOAT_TYPE(b112[l]), FLOAT_TYPE(s4_hi4[3]), sum2))))))));
+                    sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * FLOAT_TYPE((qs0[l]      ) & 3),
+                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * FLOAT_TYPE((qs16[l]     ) & 3),
+                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * FLOAT_TYPE((qs0[l]  >> 2) & 3),
+                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * FLOAT_TYPE((qs16[l] >> 2) & 3),
+                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * FLOAT_TYPE((qs0[l]  >> 4) & 3),
+                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * FLOAT_TYPE((qs16[l] >> 4) & 3),
+                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * FLOAT_TYPE((qs0[l]  >> 6) & 3),
+                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * FLOAT_TYPE((qs16[l] >> 6) & 3), sum1))))))));
+                    sum2 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][ 8],
+                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][ 9],
+                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][10],
+                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][11],
+                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][12],
+                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][13],
+                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][14],
+                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][15], sum2))))))));
                 }
                 temp[j][n] = fma(dall, sum1, fma(-dmin, sum2, temp[j][n]));
             }
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index 0266417a88f72..0658f46bd7688 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -5,7 +5,6 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
-shared uint32_t scu8[BLOCK_SIZE/16][12];
 shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][12];
 
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
@@ -61,8 +60,8 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
                 FLOAT_TYPE sum = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m    )) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m    )) != 0) ? 0 : 4)),
+                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m     )) != 0) ? 0 : 4)),
+                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m     )) != 0) ? 0 : 4)),
                           fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 1)) != 0) ? 0 : 4)),
                           fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 1)) != 0) ? 0 : 4)),
                           fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 2)) != 0) ? 0 : 4)),
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index 225f0ce70a6fc..8a3644e968c9c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -47,9 +47,9 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
             const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
 
-            const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
-            const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
+            uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
+            uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
+            uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
             uvec2 scale0 = uvec2(unpack8(scale0_u32));
             uvec2 scale4 = uvec2(unpack8(scale4_u32));
             uvec2 scale8 = uvec2(unpack8(scale8_u32));

From c01ccf8288f1e375ed5be535bab1efc43c213406 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sun, 5 Jan 2025 02:31:28 +0000
Subject: [PATCH 07/23] little stuff

---
 ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index 0658f46bd7688..54085f90e46a7 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -5,7 +5,7 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
-shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][12];
+shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][8];
 
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
@@ -44,7 +44,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
             const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
 
-            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> 4*v_im) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid%4+8]) >> (4*v_im + 2*(itid8/4)) & 0x3) << 4)) - 32);
+            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> 4*v_im) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid8%4+8]) >> (4*v_im + 2*(itid8/4)) & 0x3) << 4)) - 32);
             barrier();
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {

From bdd98c74e24b38a820aacecd5d0cef1149b66879 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sun, 5 Jan 2025 14:59:56 -0500
Subject: [PATCH 08/23] try precalculating products of a and q2_k scales

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      | 28 +++++++++++++------
 1 file changed, 20 insertions(+), 8 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 098771493fbe1..0a7ee58b39cee 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -54,6 +54,18 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             uvec2 qs0 =  uvec2(unpack8(qs0_u16));
             uvec2 qs16 = uvec2(unpack8(qs16_u16));
 
+            FLOAT_TYPE sc_q[2][8];
+            [[unroll]] for (int l = 0; l < 2; ++l) {
+                sc_q[l][0] = sccache[ix][v_im][0] * FLOAT_TYPE((qs0[l]      ) & 3);
+                sc_q[l][1] = sccache[ix][v_im][1] * FLOAT_TYPE((qs16[l]     ) & 3);
+                sc_q[l][2] = sccache[ix][v_im][2] * FLOAT_TYPE((qs0[l]  >> 2) & 3);
+                sc_q[l][3] = sccache[ix][v_im][3] * FLOAT_TYPE((qs16[l] >> 2) & 3);
+                sc_q[l][4] = sccache[ix][v_im][4] * FLOAT_TYPE((qs0[l]  >> 4) & 3);
+                sc_q[l][5] = sccache[ix][v_im][5] * FLOAT_TYPE((qs16[l] >> 4) & 3);
+                sc_q[l][6] = sccache[ix][v_im][6] * FLOAT_TYPE((qs0[l]  >> 6) & 3);
+                sc_q[l][7] = sccache[ix][v_im][7] * FLOAT_TYPE((qs16[l] >> 6) & 3);
+            }
+
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
                 B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
                 B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
@@ -67,14 +79,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
                 FLOAT_TYPE sum1 = FLOAT_TYPE(0.0);
                 FLOAT_TYPE sum2 = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * FLOAT_TYPE((qs0[l]      ) & 3),
-                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * FLOAT_TYPE((qs16[l]     ) & 3),
-                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * FLOAT_TYPE((qs0[l]  >> 2) & 3),
-                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * FLOAT_TYPE((qs16[l] >> 2) & 3),
-                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * FLOAT_TYPE((qs0[l]  >> 4) & 3),
-                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * FLOAT_TYPE((qs16[l] >> 4) & 3),
-                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * FLOAT_TYPE((qs0[l]  >> 6) & 3),
-                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * FLOAT_TYPE((qs16[l] >> 6) & 3), sum1))))))));
+                    sum1 = fma(FLOAT_TYPE(b0[l]),   sc_q[l][0],
+                           fma(FLOAT_TYPE(b16[l]),  sc_q[l][1],
+                           fma(FLOAT_TYPE(b32[l]),  sc_q[l][2],
+                           fma(FLOAT_TYPE(b48[l]),  sc_q[l][3],
+                           fma(FLOAT_TYPE(b64[l]),  sc_q[l][4],
+                           fma(FLOAT_TYPE(b80[l]),  sc_q[l][5],
+                           fma(FLOAT_TYPE(b96[l]),  sc_q[l][6],
+                           fma(FLOAT_TYPE(b112[l]), sc_q[l][7], sum1))))))));
                     sum2 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][ 8],
                            fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][ 9],
                            fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][10],

From 173077180ff5e63ffeda96a7b451303a9af69543 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sun, 5 Jan 2025 17:01:34 -0500
Subject: [PATCH 09/23] Revert "try precalculating products of a and q2_k
 scales"

This reverts commit 65110b81f23f66331a50c6e889a7c1ab9470a86b.
---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      | 28 ++++++-------------
 1 file changed, 8 insertions(+), 20 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 0a7ee58b39cee..098771493fbe1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -54,18 +54,6 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             uvec2 qs0 =  uvec2(unpack8(qs0_u16));
             uvec2 qs16 = uvec2(unpack8(qs16_u16));
 
-            FLOAT_TYPE sc_q[2][8];
-            [[unroll]] for (int l = 0; l < 2; ++l) {
-                sc_q[l][0] = sccache[ix][v_im][0] * FLOAT_TYPE((qs0[l]      ) & 3);
-                sc_q[l][1] = sccache[ix][v_im][1] * FLOAT_TYPE((qs16[l]     ) & 3);
-                sc_q[l][2] = sccache[ix][v_im][2] * FLOAT_TYPE((qs0[l]  >> 2) & 3);
-                sc_q[l][3] = sccache[ix][v_im][3] * FLOAT_TYPE((qs16[l] >> 2) & 3);
-                sc_q[l][4] = sccache[ix][v_im][4] * FLOAT_TYPE((qs0[l]  >> 4) & 3);
-                sc_q[l][5] = sccache[ix][v_im][5] * FLOAT_TYPE((qs16[l] >> 4) & 3);
-                sc_q[l][6] = sccache[ix][v_im][6] * FLOAT_TYPE((qs0[l]  >> 6) & 3);
-                sc_q[l][7] = sccache[ix][v_im][7] * FLOAT_TYPE((qs16[l] >> 6) & 3);
-            }
-
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
                 B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
                 B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
@@ -79,14 +67,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
                 FLOAT_TYPE sum1 = FLOAT_TYPE(0.0);
                 FLOAT_TYPE sum2 = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum1 = fma(FLOAT_TYPE(b0[l]),   sc_q[l][0],
-                           fma(FLOAT_TYPE(b16[l]),  sc_q[l][1],
-                           fma(FLOAT_TYPE(b32[l]),  sc_q[l][2],
-                           fma(FLOAT_TYPE(b48[l]),  sc_q[l][3],
-                           fma(FLOAT_TYPE(b64[l]),  sc_q[l][4],
-                           fma(FLOAT_TYPE(b80[l]),  sc_q[l][5],
-                           fma(FLOAT_TYPE(b96[l]),  sc_q[l][6],
-                           fma(FLOAT_TYPE(b112[l]), sc_q[l][7], sum1))))))));
+                    sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * FLOAT_TYPE((qs0[l]      ) & 3),
+                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * FLOAT_TYPE((qs16[l]     ) & 3),
+                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * FLOAT_TYPE((qs0[l]  >> 2) & 3),
+                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * FLOAT_TYPE((qs16[l] >> 2) & 3),
+                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * FLOAT_TYPE((qs0[l]  >> 4) & 3),
+                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * FLOAT_TYPE((qs16[l] >> 4) & 3),
+                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * FLOAT_TYPE((qs0[l]  >> 6) & 3),
+                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * FLOAT_TYPE((qs16[l] >> 6) & 3), sum1))))))));
                     sum2 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][ 8],
                            fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][ 9],
                            fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][10],

From b4ae7005e66cb03d8de601dd271b10a1127970c4 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sun, 5 Jan 2025 21:59:43 -0500
Subject: [PATCH 10/23] unpack should be u16, add vim swap to gitignore (about
 time)

---
 .gitignore                                           |  1 +
 .../ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp | 12 ++++++------
 .../ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp | 12 ++++++------
 3 files changed, 13 insertions(+), 12 deletions(-)

diff --git a/.gitignore b/.gitignore
index 1df7cf4a15ecd..694f36e042fb5 100644
--- a/.gitignore
+++ b/.gitignore
@@ -18,6 +18,7 @@
 *.metallib
 *.o
 *.so
+*.swp
 *.tmp
 
 # IDE / OS
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index 8a3644e968c9c..940ba3f51f299 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -47,12 +47,12 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
             const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
 
-            uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
-            uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-            uvec2 scale0 = uvec2(unpack8(scale0_u32));
-            uvec2 scale4 = uvec2(unpack8(scale4_u32));
-            uvec2 scale8 = uvec2(unpack8(scale8_u32));
+            uint16_t scale0_u16 = data_a_packed16[ib0 + i].scales[v_im    ];
+            uint16_t scale4_u16 = data_a_packed16[ib0 + i].scales[v_im + 2];
+            uint16_t scale8_u16 = data_a_packed16[ib0 + i].scales[v_im + 4];
+            uvec2 scale0 = uvec2(unpack8(scale0_u16));
+            uvec2 scale4 = uvec2(unpack8(scale4_u16));
+            uvec2 scale8 = uvec2(unpack8(scale8_u16));
 
             const uint32_t sc0 = (  scale0.x       & 0x3f);
             const uint32_t sc1 = (  scale0.y       & 0x3f);
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
index bea1dc8608f6b..c4aaf9fea525e 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
@@ -46,12 +46,12 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
             const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
 
-            uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
-            uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-            uvec2 scale0 = uvec2(unpack8(scale0_u32));
-            uvec2 scale4 = uvec2(unpack8(scale4_u32));
-            uvec2 scale8 = uvec2(unpack8(scale8_u32));
+            uint16_t scale0_u16 = data_a_packed16[ib0 + i].scales[v_im    ];
+            uint16_t scale4_u16 = data_a_packed16[ib0 + i].scales[v_im + 2];
+            uint16_t scale8_u16 = data_a_packed16[ib0 + i].scales[v_im + 4];
+            uvec2 scale0 = uvec2(unpack8(scale0_u16));
+            uvec2 scale4 = uvec2(unpack8(scale4_u16));
+            uvec2 scale8 = uvec2(unpack8(scale8_u16));
 
             const uint32_t sc0 = (  scale0.x       & 0x3f);
             const uint32_t sc1 = (  scale0.y       & 0x3f);

From cdf70cf27fb9c8abb0caa2f3104b66616c207f60 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sun, 5 Jan 2025 22:43:12 -0500
Subject: [PATCH 11/23] better q4_k scales

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      |  1 -
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      | 94 ++++++++++---------
 2 files changed, 48 insertions(+), 47 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 098771493fbe1..1f00e442defbb 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -25,7 +25,6 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
     const uint l0 = 2*v_in;                                  // 0...15
     const uint q_offset = 32*v_im + l0;
-    const uint s_offset = 8*v_im;
     const uint y_offset = 128*v_im + l0;
 
     FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index 940ba3f51f299..6761f3d32b3d9 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -43,55 +43,57 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
         [[unroll]] for (uint n = 0; n < num_rows; ++n) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-            f16vec2 d = data_a[ib0 + i].d;
+            const f16vec2 d = data_a[ib0 + i].d;
             const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
             const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
 
-            uint16_t scale0_u16 = data_a_packed16[ib0 + i].scales[v_im    ];
-            uint16_t scale4_u16 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            uint16_t scale8_u16 = data_a_packed16[ib0 + i].scales[v_im + 4];
-            uvec2 scale0 = uvec2(unpack8(scale0_u16));
-            uvec2 scale4 = uvec2(unpack8(scale4_u16));
-            uvec2 scale8 = uvec2(unpack8(scale8_u16));
-
-            const uint32_t sc0 = (  scale0.x       & 0x3f);
-            const uint32_t sc1 = (  scale0.y       & 0x3f);
-            const uint32_t sc2 = (  scale4.x       & 0x3f);
-            const uint32_t sc3 = (  scale4.y       & 0x3f);
-            const uint32_t sc4 = (( scale8.x       & 0x0f) | ((scale0.x & 0xc0) >> 2));
-            const uint32_t sc5 = (( scale8.y       & 0x0f) | ((scale0.y & 0xc0) >> 2));
-            const uint32_t sc6 = (((scale8.x >> 4) & 0x0f) | ((scale4.x & 0xc0) >> 2));
-            const uint32_t sc7 = (((scale8.y >> 4) & 0x0f) | ((scale4.y & 0xc0) >> 2));
-
-            uint32_t qs0_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4];
-            uint32_t qs64_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4 + 16];
-
-            uint32_t qs0_u32_lo4 = qs0_u32 & 0x0F0F0F0F;
-            uint32_t qs0_u32_hi4 = (qs0_u32 >> 4) & 0x0F0F0F0F;
-            uint32_t qs64_u32_lo4 = qs64_u32 & 0x0F0F0F0F;
-            uint32_t qs64_u32_hi4 = (qs64_u32 >> 4) & 0x0F0F0F0F;
-
-            uvec4 qs0_lo4 = uvec4(unpack8(qs0_u32_lo4));
-            uvec4 qs64_lo4 = uvec4(unpack8(qs64_u32_lo4));
-            uvec4 qs0_hi4 = uvec4(unpack8(qs0_u32_hi4));
-            uvec4 qs64_hi4 = uvec4(unpack8(qs64_u32_hi4));
-
-            const uint32_t q4_0  = qs0_lo4.x;
-            const uint32_t q4_1  = qs0_lo4.y;
-            const uint32_t q4_2  = qs0_lo4.z;
-            const uint32_t q4_3  = qs0_lo4.w;
-            const uint32_t q4_4  = qs0_hi4.x;
-            const uint32_t q4_5  = qs0_hi4.y;
-            const uint32_t q4_6  = qs0_hi4.z;
-            const uint32_t q4_7  = qs0_hi4.w;
-            const uint32_t q4_8  = qs64_lo4.x;
-            const uint32_t q4_9  = qs64_lo4.y;
-            const uint32_t q4_10 = qs64_lo4.z;
-            const uint32_t q4_11 = qs64_lo4.w;
-            const uint32_t q4_12 = qs64_hi4.x;
-            const uint32_t q4_13 = qs64_hi4.y;
-            const uint32_t q4_14 = qs64_hi4.z;
-            const uint32_t q4_15 = qs64_hi4.w;
+            const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
+            const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
+            const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
+
+            const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32;
+            const uint32_t scale_0_4_h = (scale_0_4_l & 0xc0c0c0c0) >> 2;
+            const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3f3f3f3f));
+            const vec4 scale8_f = vec4(unpack8(((((scale8_u32 >> 4) << 16) | scale8_u32) & 0x0f0f0f0f) | scale_0_4_h));
+
+            const FLOAT_TYPE sc0 = scale_0_4_l_f.x;
+            const FLOAT_TYPE sc1 = scale_0_4_l_f.y;
+            const FLOAT_TYPE sc2 = scale_0_4_l_f.z;
+            const FLOAT_TYPE sc3 = scale_0_4_l_f.w;
+            const FLOAT_TYPE sc4 = scale8_f.x;
+            const FLOAT_TYPE sc5 = scale8_f.y;
+            const FLOAT_TYPE sc6 = scale8_f.z;
+            const FLOAT_TYPE sc7 = scale8_f.w;
+
+            const uint32_t qs0_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4];
+            const uint32_t qs64_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4 + 16];
+
+            const uint32_t qs0_u32_lo4 = qs0_u32 & 0x0F0F0F0F;
+            const uint32_t qs0_u32_hi4 = (qs0_u32 >> 4) & 0x0F0F0F0F;
+            const uint32_t qs64_u32_lo4 = qs64_u32 & 0x0F0F0F0F;
+            const uint32_t qs64_u32_hi4 = (qs64_u32 >> 4) & 0x0F0F0F0F;
+
+            const vec4 qs0_lo4 = vec4(unpack8(qs0_u32_lo4));
+            const vec4 qs64_lo4 = vec4(unpack8(qs64_u32_lo4));
+            const vec4 qs0_hi4 = vec4(unpack8(qs0_u32_hi4));
+            const vec4 qs64_hi4 = vec4(unpack8(qs64_u32_hi4));
+
+            const FLOAT_TYPE q4_0  = qs0_lo4.x;
+            const FLOAT_TYPE q4_1  = qs0_lo4.y;
+            const FLOAT_TYPE q4_2  = qs0_lo4.z;
+            const FLOAT_TYPE q4_3  = qs0_lo4.w;
+            const FLOAT_TYPE q4_4  = qs0_hi4.x;
+            const FLOAT_TYPE q4_5  = qs0_hi4.y;
+            const FLOAT_TYPE q4_6  = qs0_hi4.z;
+            const FLOAT_TYPE q4_7  = qs0_hi4.w;
+            const FLOAT_TYPE q4_8  = qs64_lo4.x;
+            const FLOAT_TYPE q4_9  = qs64_lo4.y;
+            const FLOAT_TYPE q4_10 = qs64_lo4.z;
+            const FLOAT_TYPE q4_11 = qs64_lo4.w;
+            const FLOAT_TYPE q4_12 = qs64_hi4.x;
+            const FLOAT_TYPE q4_13 = qs64_hi4.y;
+            const FLOAT_TYPE q4_14 = qs64_hi4.z;
+            const FLOAT_TYPE q4_15 = qs64_hi4.w;
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
                 B_TYPE_VEC4 by10 =  data_b_v4[(j*p.batch_stride_b + b_offset + y1_idx) / 4];

From 6f5d62b098a45d9c4a0d833d03ec68848b0c06b5 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Mon, 6 Jan 2025 17:13:23 -0500
Subject: [PATCH 12/23] q5_k

---
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      |  6 ++--
 .../vulkan-shaders/mul_mat_vec_q5_k.comp      | 32 ++++++++++---------
 2 files changed, 20 insertions(+), 18 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index 6761f3d32b3d9..28cde16db7c33 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -52,9 +52,9 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
 
             const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32;
-            const uint32_t scale_0_4_h = (scale_0_4_l & 0xc0c0c0c0) >> 2;
-            const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3f3f3f3f));
-            const vec4 scale8_f = vec4(unpack8(((((scale8_u32 >> 4) << 16) | scale8_u32) & 0x0f0f0f0f) | scale_0_4_h));
+            const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2;
+            const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F));
+            const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h));
 
             const FLOAT_TYPE sc0 = scale_0_4_l_f.x;
             const FLOAT_TYPE sc1 = scale_0_4_l_f.y;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
index c4aaf9fea525e..b5e8399fe5f8d 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
@@ -46,21 +46,23 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
             const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
 
-            uint16_t scale0_u16 = data_a_packed16[ib0 + i].scales[v_im    ];
-            uint16_t scale4_u16 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            uint16_t scale8_u16 = data_a_packed16[ib0 + i].scales[v_im + 4];
-            uvec2 scale0 = uvec2(unpack8(scale0_u16));
-            uvec2 scale4 = uvec2(unpack8(scale4_u16));
-            uvec2 scale8 = uvec2(unpack8(scale8_u16));
-
-            const uint32_t sc0 = (  scale0.x       & 0x3f);
-            const uint32_t sc1 = (  scale0.y       & 0x3f);
-            const uint32_t sc2 = (  scale4.x       & 0x3f);
-            const uint32_t sc3 = (  scale4.y       & 0x3f);
-            const uint32_t sc4 = (( scale8.x       & 0x0f) | ((scale0.x & 0xc0) >> 2));
-            const uint32_t sc5 = (( scale8.y       & 0x0f) | ((scale0.y & 0xc0) >> 2));
-            const uint32_t sc6 = (((scale8.x >> 4) & 0x0f) | ((scale4.x & 0xc0) >> 2));
-            const uint32_t sc7 = (((scale8.y >> 4) & 0x0f) | ((scale4.y & 0xc0) >> 2));
+            const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
+            const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
+            const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
+
+            const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32;
+            const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2;
+            const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F));
+            const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h));
+
+            const FLOAT_TYPE sc0 = scale_0_4_l_f.x;
+            const FLOAT_TYPE sc1 = scale_0_4_l_f.y;
+            const FLOAT_TYPE sc2 = scale_0_4_l_f.z;
+            const FLOAT_TYPE sc3 = scale_0_4_l_f.w;
+            const FLOAT_TYPE sc4 = scale8_f.x;
+            const FLOAT_TYPE sc5 = scale8_f.y;
+            const FLOAT_TYPE sc6 = scale8_f.z;
+            const FLOAT_TYPE sc7 = scale8_f.w;
 
             uint32_t qs0_16_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
             uint32_t qs64_80_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 32]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 40]) << 16);

From 91f1d9ce991f068a0bc39befecadf0fa52800e24 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Tue, 7 Jan 2025 19:57:55 -0500
Subject: [PATCH 13/23] better q6_k with separate paths for all threads and
 partial threads in use, plus some more optimizations

---
 .../vulkan-shaders/mul_mat_vec_q6_k.comp      | 126 ++++++++++--------
 1 file changed, 73 insertions(+), 53 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
index e1afd55e0f540..231bfc5da04a1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
@@ -8,7 +8,73 @@ layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
 shared FLOAT_TYPE sccache[BLOCK_SIZE/16][16];
 
-void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
+FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
+
+void calc_superblock(const uint a_offset, const uint b_offset, const uint itid, const uint ix, const uint ql_offset, const uint qh_offset, const uint s_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) {
+    const uint y_idx = i * QUANT_K + y_offset;
+
+    [[unroll]] for (uint n = 0; n < num_rows; ++n) {
+        const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
+
+        if (!all_threads) { // when we don't have enough blocks to use all threads
+            if (i < num_blocks_per_row)
+                sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]);
+            barrier();
+
+            if (i >= num_blocks_per_row)
+                continue;
+        }
+
+        const uint32_t ql0_u32 =  uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 1]) << 16);
+        const uint32_t ql32_u32 = uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 16]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 17]) << 16);
+
+        const uint32_t ql0_u32_lo4 = ql0_u32 & 0x0F0F0F0F;
+        const uint32_t ql0_u32_hi4 = (ql0_u32 >> 4) & 0x0F0F0F0F;
+        const uint32_t ql32_u32_lo4 = ql32_u32 & 0x0F0F0F0F;
+        const uint32_t ql32_u32_hi4 = (ql32_u32 >> 4) & 0x0F0F0F0F;
+
+        const uint32_t qh_u32 = uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2 + 1]) << 16);
+        const uint32_t qh0_u32 = (qh_u32 & 0x03030303) << 4;
+        const uint32_t qh2_u32 = (qh_u32 & 0x0C0C0C0C) << 2;
+        const uint32_t qh4_u32 = (qh_u32 & 0x30303030);
+        const uint32_t qh6_u32 = (qh_u32 & 0xC0C0C0C0) >> 2;
+
+        const uint32_t q0_u32 = ql0_u32_lo4  | qh0_u32;
+        const uint32_t q1_u32 = ql32_u32_lo4 | qh2_u32;
+        const uint32_t q2_u32 = ql0_u32_hi4  | qh4_u32;
+        const uint32_t q3_u32 = ql32_u32_hi4 | qh6_u32;
+
+        const vec4 q0 = vec4(unpack8(q0_u32)) - 32;
+        const vec4 q1 = vec4(unpack8(q1_u32)) - 32;
+        const vec4 q2 = vec4(unpack8(q2_u32)) - 32;
+        const vec4 q3 = vec4(unpack8(q3_u32)) - 32;
+
+        if (all_threads) {
+            sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]);
+            barrier();
+        }
+
+        const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
+
+        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
+            B_TYPE_VEC4 by0  = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4];
+            B_TYPE_VEC4 by32 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 8];
+            B_TYPE_VEC4 by64 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 16];
+            B_TYPE_VEC4 by96 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 24];
+
+            FLOAT_TYPE sum[4] = {0, 0, 0, 0};
+            [[unroll]] for (uint l = 0; l < 4; ++l) {
+                sum[0] = fma(FLOAT_TYPE(by0[l]), q0[l], sum[0]);
+                sum[1] = fma(FLOAT_TYPE(by32[l]), q1[l], sum[1]);
+                sum[2] = fma(FLOAT_TYPE(by64[l]), q2[l], sum[2]);
+                sum[3] = fma(FLOAT_TYPE(by96[l]), q3[l], sum[3]);
+            }
+            temp[j][n] = fma(fma(sum[0], sccache[ix][s_offset], fma(sum[1], sccache[ix][s_offset + 2], fma(sum[2], sccache[ix][s_offset + 4], sum[3] * sccache[ix][s_offset + 6]))), d, temp[j][n]);
+        }
+    }
+}
+
+void compute_outputs(const uint first_row, const uint num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
 
@@ -31,65 +97,19 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint s_offset  =  8*v_im + is;
     const uint y_offset = 128*v_im + l0;
 
-    FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
-
     [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
         [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) {
             temp[j][i] = FLOAT_TYPE(0);
         }
     }
 
-    [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) {
-        const uint y_idx = i * QUANT_K + y_offset;
-
-        [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-            const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-            const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
-
-            sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]);
-            barrier();
-
-            uint32_t ql0_u32 =  uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 1]) << 16);
-            uint32_t ql32_u32 = uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 16]) | (uint32_t(data_a_packed16[ib0 + i].ql[ql_offset / 2 + 17]) << 16);
-
-            uint32_t ql0_u32_lo4 = ql0_u32 & 0x0F0F0F0F;
-            uint32_t ql0_u32_hi4 = (ql0_u32 >> 4) & 0x0F0F0F0F;
-            uint32_t ql32_u32_lo4 = ql32_u32 & 0x0F0F0F0F;
-            uint32_t ql32_u32_hi4 = (ql32_u32 >> 4) & 0x0F0F0F0F;
-
-            uint32_t qh_u32 = uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qh[qh_offset / 2 + 1]) << 16);
-            uint32_t qh0_u32 = (qh_u32 & 0x03030303) << 4;
-            uint32_t qh2_u32 = (qh_u32 & 0x0C0C0C0C) << 2;
-            uint32_t qh4_u32 = (qh_u32 & 0x30303030);
-            uint32_t qh6_u32 = (qh_u32 & 0xC0C0C0C0) >> 2;
-
-            uint32_t q0_u32 = ql0_u32_lo4  | qh0_u32;
-            uint32_t q1_u32 = ql32_u32_lo4 | qh2_u32;
-            uint32_t q2_u32 = ql0_u32_hi4  | qh4_u32;
-            uint32_t q3_u32 = ql32_u32_hi4 | qh6_u32;
-
-            uvec4 q0 = uvec4(unpack8(q0_u32));
-            uvec4 q1 = uvec4(unpack8(q1_u32));
-            uvec4 q2 = uvec4(unpack8(q2_u32));
-            uvec4 q3 = uvec4(unpack8(q3_u32));
-
-            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-                B_TYPE_VEC4 by0  = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4];
-                B_TYPE_VEC4 by32 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 8];
-                B_TYPE_VEC4 by64 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 16];
-                B_TYPE_VEC4 by96 = data_b_v4[(j*p.batch_stride_b + b_offset + y_idx) / 4 + 24];
-
-                FLOAT_TYPE sum[4] = {0, 0, 0, 0};
-                [[unroll]] for (uint l = 0; l < 4; ++l) {
-                    sum[0] = fma(FLOAT_TYPE(by0[l]), FLOAT_TYPE(int8_t(q0[l]) - 32), sum[0]);
-                    sum[1] = fma(FLOAT_TYPE(by32[l]), FLOAT_TYPE(int8_t(q1[l]) - 32), sum[1]);
-                    sum[2] = fma(FLOAT_TYPE(by64[l]), FLOAT_TYPE(int8_t(q2[l]) - 32), sum[2]);
-                    sum[3] = fma(FLOAT_TYPE(by96[l]), FLOAT_TYPE(int8_t(q3[l]) - 32), sum[3]);
-                }
-                temp[j][n] = fma(fma(sum[0], sccache[ix][s_offset], fma(sum[1], sccache[ix][s_offset + 2], fma(sum[2], sccache[ix][s_offset + 4], sum[3] * sccache[ix][s_offset + 6]))), d, temp[j][n]);
-            }
-        }
+    const uint nbr_par_th = num_blocks_per_row%it_size;
+    const uint nbr_all_th = num_blocks_per_row - nbr_par_th;
+    uint i0 = 0;
+    [[unroll]] for (; i0 < nbr_all_th; i0 += it_size) {
+        calc_superblock(a_offset, b_offset, itid, ix, ql_offset, qh_offset, s_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, true);
     }
+    calc_superblock(a_offset, b_offset, itid, ix, ql_offset, qh_offset, s_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, false);
 
     reduce_result(temp, d_offset, first_row, num_rows, tid);
 }

From cc28742ca39b12efea5f9b8d87d44860a3430ccb Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Tue, 7 Jan 2025 21:20:33 -0500
Subject: [PATCH 14/23] q2_k better dequant

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      | 27 ++++-----
 .../vulkan-shaders/mul_mat_vec_q5_k.comp      | 56 +++++++++----------
 2 files changed, 42 insertions(+), 41 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 1f00e442defbb..ed924a072a982 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -40,7 +40,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
         [[unroll]] for (uint n = 0; n < num_rows; ++n) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-            f16vec2 d = data_a[ib0 + i].d;
+            const f16vec2 d = data_a[ib0 + i].d;
             const FLOAT_TYPE dall = d.x;
             const FLOAT_TYPE dmin = d.y;
 
@@ -48,10 +48,11 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             sccache[ix][1][itid] = FLOAT_TYPE(bitfieldExtract(uint(data_a[ib0 + i].scales[itid8+8]), int(v_im*4), 4)); // upper 8 bytes
             barrier();
 
-            uint16_t qs0_u16 = data_a_packed16[ib0 + i].qs[q_offset / 2];
-            uint16_t qs16_u16 = data_a_packed16[ib0 + i].qs[q_offset / 2 + 8];
-            uvec2 qs0 =  uvec2(unpack8(qs0_u16));
-            uvec2 qs16 = uvec2(unpack8(qs16_u16));
+            const uint32_t qs_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
+            const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303));
+            const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303));
+            const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
+            const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
                 B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
@@ -66,14 +67,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
                 FLOAT_TYPE sum1 = FLOAT_TYPE(0.0);
                 FLOAT_TYPE sum2 = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * FLOAT_TYPE((qs0[l]      ) & 3),
-                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * FLOAT_TYPE((qs16[l]     ) & 3),
-                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * FLOAT_TYPE((qs0[l]  >> 2) & 3),
-                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * FLOAT_TYPE((qs16[l] >> 2) & 3),
-                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * FLOAT_TYPE((qs0[l]  >> 4) & 3),
-                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * FLOAT_TYPE((qs16[l] >> 4) & 3),
-                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * FLOAT_TYPE((qs0[l]  >> 6) & 3),
-                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * FLOAT_TYPE((qs16[l] >> 6) & 3), sum1))))))));
+                    sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * qs_u32_0[l  ],
+                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * qs_u32_0[l+2],
+                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * qs_u32_2[l  ],
+                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * qs_u32_2[l+2],
+                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * qs_u32_4[l  ],
+                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * qs_u32_4[l+2],
+                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * qs_u32_6[l  ],
+                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * qs_u32_6[l+2], sum1))))))));
                     sum2 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][ 8],
                            fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][ 9],
                            fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][10],
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
index b5e8399fe5f8d..9039ce1f34cbf 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
@@ -64,47 +64,47 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const FLOAT_TYPE sc6 = scale8_f.z;
             const FLOAT_TYPE sc7 = scale8_f.w;
 
-            uint32_t qs0_16_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
-            uint32_t qs64_80_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 32]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 40]) << 16);
+            const uint32_t qs0_16_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
+            const uint32_t qs64_80_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 32]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 40]) << 16);
 
             uint32_t qs0_16_u32_lo4 = qs0_16_u32 & 0x0F0F0F0F;
             uint32_t qs0_16_u32_hi4 = (qs0_16_u32 >> 4) & 0x0F0F0F0F;
             uint32_t qs64_80_u32_lo4 = qs64_80_u32 & 0x0F0F0F0F;
             uint32_t qs64_80_u32_hi4 = (qs64_80_u32 >> 4) & 0x0F0F0F0F;
 
-            uint32_t qh = pack32(u16vec2(data_a_packed16[ib0 + i].qh[l0 / 2], data_a_packed16[ib0 + i].qh[l0 / 2 + 8]));
+            const uint32_t qh = pack32(u16vec2(data_a_packed16[ib0 + i].qh[l0 / 2], data_a_packed16[ib0 + i].qh[l0 / 2 + 8]));
 
-            uint32_t qs0_16_lo4_offset16 = ((qh >> (2*v_im)) & 0x01010101) << 4;
-            uint32_t qs0_16_hi4_offset16 = ((qh >> (2*v_im)) & 0x02020202) << 3;
-            uint32_t qs64_80_lo4_offset16 = ((qh >> (2*v_im)) & 0x10101010);
-            uint32_t qs64_80_hi4_offset16 = ((qh >> (2*v_im)) & 0x20202020) >> 1;
+            const uint32_t qs0_16_lo4_offset16 = ((qh >> (2*v_im)) & 0x01010101) << 4;
+            const uint32_t qs0_16_hi4_offset16 = ((qh >> (2*v_im)) & 0x02020202) << 3;
+            const uint32_t qs64_80_lo4_offset16 = ((qh >> (2*v_im)) & 0x10101010);
+            const uint32_t qs64_80_hi4_offset16 = ((qh >> (2*v_im)) & 0x20202020) >> 1;
 
             qs0_16_u32_lo4 += qs0_16_lo4_offset16;
             qs0_16_u32_hi4 += qs0_16_hi4_offset16;
             qs64_80_u32_lo4 += qs64_80_lo4_offset16;
             qs64_80_u32_hi4 += qs64_80_hi4_offset16;
 
-            uvec4 qs0_16_lo4 = uvec4(unpack8(qs0_16_u32_lo4));
-            uvec4 qs64_80_lo4 = uvec4(unpack8(qs64_80_u32_lo4));
-            uvec4 qs0_16_hi4 = uvec4(unpack8(qs0_16_u32_hi4));
-            uvec4 qs64_80_hi4 = uvec4(unpack8(qs64_80_u32_hi4));
-
-            const uint32_t q4_0  = qs0_16_lo4.x;
-            const uint32_t q4_1  = qs0_16_lo4.y;
-            const uint32_t q4_2  = qs0_16_lo4.z;
-            const uint32_t q4_3  = qs0_16_lo4.w;
-            const uint32_t q4_4  = qs0_16_hi4.x;
-            const uint32_t q4_5  = qs0_16_hi4.y;
-            const uint32_t q4_6  = qs0_16_hi4.z;
-            const uint32_t q4_7  = qs0_16_hi4.w;
-            const uint32_t q4_8  = qs64_80_lo4.x;
-            const uint32_t q4_9  = qs64_80_lo4.y;
-            const uint32_t q4_10 = qs64_80_lo4.z;
-            const uint32_t q4_11 = qs64_80_lo4.w;
-            const uint32_t q4_12 = qs64_80_hi4.x;
-            const uint32_t q4_13 = qs64_80_hi4.y;
-            const uint32_t q4_14 = qs64_80_hi4.z;
-            const uint32_t q4_15 = qs64_80_hi4.w;
+            const vec4 qs0_16_lo4 = vec4(unpack8(qs0_16_u32_lo4));
+            const vec4 qs64_80_lo4 = vec4(unpack8(qs64_80_u32_lo4));
+            const vec4 qs0_16_hi4 = vec4(unpack8(qs0_16_u32_hi4));
+            const vec4 qs64_80_hi4 = vec4(unpack8(qs64_80_u32_hi4));
+
+            const FLOAT_TYPE q4_0  = qs0_16_lo4.x;
+            const FLOAT_TYPE q4_1  = qs0_16_lo4.y;
+            const FLOAT_TYPE q4_2  = qs0_16_lo4.z;
+            const FLOAT_TYPE q4_3  = qs0_16_lo4.w;
+            const FLOAT_TYPE q4_4  = qs0_16_hi4.x;
+            const FLOAT_TYPE q4_5  = qs0_16_hi4.y;
+            const FLOAT_TYPE q4_6  = qs0_16_hi4.z;
+            const FLOAT_TYPE q4_7  = qs0_16_hi4.w;
+            const FLOAT_TYPE q4_8  = qs64_80_lo4.x;
+            const FLOAT_TYPE q4_9  = qs64_80_lo4.y;
+            const FLOAT_TYPE q4_10 = qs64_80_lo4.z;
+            const FLOAT_TYPE q4_11 = qs64_80_lo4.w;
+            const FLOAT_TYPE q4_12 = qs64_80_hi4.x;
+            const FLOAT_TYPE q4_13 = qs64_80_hi4.y;
+            const FLOAT_TYPE q4_14 = qs64_80_hi4.z;
+            const FLOAT_TYPE q4_15 = qs64_80_hi4.w;
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
                 B_TYPE_VEC2 by10 =  data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2];

From fe71a8c4a12540f0ae666e2b6d518f54a057d833 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Tue, 7 Jan 2025 22:05:51 -0500
Subject: [PATCH 15/23] q3_k optimizations

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      |  2 +-
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      | 29 +++++++++++++------
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      |  2 +-
 .../vulkan-shaders/mul_mat_vec_q5_k.comp      |  2 +-
 .../vulkan-shaders/mul_mat_vec_q6_k.comp      |  2 +-
 5 files changed, 24 insertions(+), 13 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index ed924a072a982..ae78a2d14123b 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -21,7 +21,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint itid8 = itid%8;
 
     const uint v_im = itid/8;                                // 0 or 1. 0 computes 0..., 1 computes 128...
-    const uint v_in = itid - 8*v_im;                         // 0...15 or 0...7
+    const uint v_in = itid - 8*v_im;                         // 0...7
 
     const uint l0 = 2*v_in;                                  // 0...15
     const uint q_offset = 32*v_im + l0;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index 54085f90e46a7..cbe3269d62210 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -21,7 +21,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint itid8 = itid%8;
 
     const uint v_im = itid/8;                               // 0 or 1. 0 computes 0..., 1 computes 128...
-    const uint v_in = itid - 8*v_im;                        // 0...15 or 0...7
+    const uint v_in = itid - 8*v_im;                        // 0...7
 
     const uint8_t m = uint8_t(1 << (4 * v_im));
 
@@ -47,6 +47,17 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> 4*v_im) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid8%4+8]) >> (4*v_im + 2*(itid8/4)) & 0x3) << 4)) - 32);
             barrier();
 
+            // 0, 1, 16, 17
+            uint32_t qs_u32 = uint32_t(data_a[ib0 + i].qs[q_offset]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 1]) << 8);
+            qs_u32 |= (uint32_t(data_a[ib0 + i].qs[q_offset + 16]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 17]) << 8)) << 16;
+            const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303));
+            const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303));
+            const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
+            const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
+
+            const uvec2 hmk0 = uvec2(unpack8(data_a_packed16[ib0 + i].hmask[v_in]));
+            const uvec2 hmk16 = uvec2(unpack8(data_a_packed16[ib0 + i].hmask[v_in + 8]));
+
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
 
                 B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
@@ -60,14 +71,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
                 FLOAT_TYPE sum = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m     )) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16]     ) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m     )) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 1)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 2) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 1)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 2)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 4) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 2)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l   ] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l   ] & (m << 3)) != 0) ? 0 : 4)),
-                          fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], FLOAT_TYPE(((data_a[ib0 + i].qs[q_offset + l+16] >> 6) & 3) - (((data_a[ib0 + i].hmask[l0 + l+16] & (m << 3)) != 0) ? 0 : 4)), sum))))))));
+                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], qs_u32_0[l  ] - FLOAT_TYPE((( hmk0[l] & (m     )) != 0) ? 0 : 4),
+                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], qs_u32_0[l+2] - FLOAT_TYPE(((hmk16[l] & (m     )) != 0) ? 0 : 4),
+                          fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], qs_u32_2[l  ] - FLOAT_TYPE((( hmk0[l] & (m << 1)) != 0) ? 0 : 4),
+                          fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], qs_u32_2[l+2] - FLOAT_TYPE(((hmk16[l] & (m << 1)) != 0) ? 0 : 4),
+                          fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], qs_u32_4[l  ] - FLOAT_TYPE((( hmk0[l] & (m << 2)) != 0) ? 0 : 4),
+                          fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], qs_u32_4[l+2] - FLOAT_TYPE(((hmk16[l] & (m << 2)) != 0) ? 0 : 4),
+                          fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], qs_u32_6[l  ] - FLOAT_TYPE((( hmk0[l] & (m << 3)) != 0) ? 0 : 4),
+                          fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], qs_u32_6[l+2] - FLOAT_TYPE(((hmk16[l] & (m << 3)) != 0) ? 0 : 4), sum))))))));
                 }
                 temp[j][n] = fma(d, sum, temp[j][n]);
             }
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index 28cde16db7c33..f8bc885dccda8 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -19,7 +19,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint ix = tid/16;
 
     const uint il = itid/4;                         // 0...3
-    const uint ir = itid - 4*il;                    // 0...7 or 0...3
+    const uint ir = itid - 4*il;                    // 0...3
     const uint n =  4;
 
     const uint v_im = il / 2;  // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
index 9039ce1f34cbf..72b36d9af80c8 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
@@ -19,7 +19,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint ix = tid/16;
 
     const uint il = itid/4;                          // 0...3
-    const uint ir = itid - 4*il;                     // 0...7 or 0...3
+    const uint ir = itid - 4*il;                     // 0...3
 
     const uint v_im = il / 2;  // 0 or 1. 0 computes 0,32 + 128,160, 1 computes 64,96 + 192,224
     const uint v_in = il % 2;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
index 231bfc5da04a1..efbdfe939cd87 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
@@ -87,7 +87,7 @@ void compute_outputs(const uint first_row, const uint num_rows) {
     const uint ix = tid/16;
 
     const uint v_im = itid/8;                               // 0 or 1. 0 computes 0..., 1 computes 128...
-    const uint v_in = itid - 8*v_im;                        // 0...15 or 0...7
+    const uint v_in = itid - 8*v_im;                        // 0...7
 
     const uint l0 = 4 * v_in;                               // 0, 4, 8, ..., 28
     const uint is = v_in / 4;

From 923e9a8377dfc76a189c0e3f5e06aff4384453e3 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Wed, 8 Jan 2025 21:13:09 -0500
Subject: [PATCH 16/23] q3_k use hmask simd from cpu avx version

---
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      | 31 ++++++++++++-------
 1 file changed, 19 insertions(+), 12 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index cbe3269d62210..b9bc0bfcdc389 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -21,9 +21,13 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint itid8 = itid%8;
 
     const uint v_im = itid/8;                               // 0 or 1. 0 computes 0..., 1 computes 128...
+    const uint v_im4 = v_im*4;
     const uint v_in = itid - 8*v_im;                        // 0...7
 
-    const uint8_t m = uint8_t(1 << (4 * v_im));
+    const uint32_t m = 0x01010101 << (4 * v_im);
+    uint32_t hm_m[4];
+    [[unroll]] for (uint j = 0; j < 4; ++j)
+        hm_m[j] = m << j;
 
     const uint l0 = 2*v_in;                                 // 0...15
     const uint q_offset = 32*v_im + l0;
@@ -44,7 +48,7 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
             const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
 
-            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> 4*v_im) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid8%4+8]) >> (4*v_im + 2*(itid8/4)) & 0x3) << 4)) - 32);
+            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> v_im4) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid8%4+8]) >> (v_im4 + 2*(itid8/4)) & 0x3) << 4)) - 32);
             barrier();
 
             // 0, 1, 16, 17
@@ -55,8 +59,11 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
             const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
 
-            const uvec2 hmk0 = uvec2(unpack8(data_a_packed16[ib0 + i].hmask[v_in]));
-            const uvec2 hmk16 = uvec2(unpack8(data_a_packed16[ib0 + i].hmask[v_in + 8]));
+            const uint32_t hmk = ~(uint32_t(data_a_packed16[ib0 + i].hmask[v_in]) | (uint32_t(data_a_packed16[ib0 + i].hmask[v_in + 8]) << 16));
+            const vec4 hmk_0 = vec4(unpack8(((hmk & hm_m[0]) >> (    v_im4)) << 2));
+            const vec4 hmk_1 = vec4(unpack8(((hmk & hm_m[1]) >> (1 + v_im4)) << 2));
+            const vec4 hmk_2 = vec4(unpack8(((hmk & hm_m[2]) >> (2 + v_im4)) << 2));
+            const vec4 hmk_3 = vec4(unpack8(((hmk & hm_m[3]) >> (3 + v_im4)) << 2));
 
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
 
@@ -71,14 +78,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
                 FLOAT_TYPE sum = FLOAT_TYPE(0.0);
                 [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], qs_u32_0[l  ] - FLOAT_TYPE((( hmk0[l] & (m     )) != 0) ? 0 : 4),
-                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], qs_u32_0[l+2] - FLOAT_TYPE(((hmk16[l] & (m     )) != 0) ? 0 : 4),
-                          fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], qs_u32_2[l  ] - FLOAT_TYPE((( hmk0[l] & (m << 1)) != 0) ? 0 : 4),
-                          fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], qs_u32_2[l+2] - FLOAT_TYPE(((hmk16[l] & (m << 1)) != 0) ? 0 : 4),
-                          fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], qs_u32_4[l  ] - FLOAT_TYPE((( hmk0[l] & (m << 2)) != 0) ? 0 : 4),
-                          fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], qs_u32_4[l+2] - FLOAT_TYPE(((hmk16[l] & (m << 2)) != 0) ? 0 : 4),
-                          fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], qs_u32_6[l  ] - FLOAT_TYPE((( hmk0[l] & (m << 3)) != 0) ? 0 : 4),
-                          fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], qs_u32_6[l+2] - FLOAT_TYPE(((hmk16[l] & (m << 3)) != 0) ? 0 : 4), sum))))))));
+                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], qs_u32_0[l  ] - hmk_0[l  ],
+                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], qs_u32_0[l+2] - hmk_0[l+2],
+                          fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], qs_u32_2[l  ] - hmk_1[l  ],
+                          fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], qs_u32_2[l+2] - hmk_1[l+2],
+                          fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], qs_u32_4[l  ] - hmk_2[l  ],
+                          fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], qs_u32_4[l+2] - hmk_2[l+2],
+                          fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], qs_u32_6[l  ] - hmk_3[l  ],
+                          fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], qs_u32_6[l+2] - hmk_3[l+2], sum))))))));
                 }
                 temp[j][n] = fma(d, sum, temp[j][n]);
             }

From 51b5ac507db6c7e4288c57cf2e7955c411ef8490 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Thu, 9 Jan 2025 17:06:54 -0500
Subject: [PATCH 17/23] make the caches happy

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp       |  7 ++++---
 .../vulkan-shaders/mul_mat_vec_q3_k.comp       | 18 +++++++++---------
 2 files changed, 13 insertions(+), 12 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index ae78a2d14123b..129442f7ec924 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -40,9 +40,6 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
         [[unroll]] for (uint n = 0; n < num_rows; ++n) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-            const f16vec2 d = data_a[ib0 + i].d;
-            const FLOAT_TYPE dall = d.x;
-            const FLOAT_TYPE dmin = d.y;
 
             sccache[ix][0][itid] = FLOAT_TYPE(bitfieldExtract(uint(data_a[ib0 + i].scales[itid8]), int(v_im*4), 4)); // lower 8 bytes
             sccache[ix][1][itid] = FLOAT_TYPE(bitfieldExtract(uint(data_a[ib0 + i].scales[itid8+8]), int(v_im*4), 4)); // upper 8 bytes
@@ -54,6 +51,10 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
             const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
 
+            const f16vec2 d = data_a[ib0 + i].d;
+            const FLOAT_TYPE dall = d.x;
+            const FLOAT_TYPE dmin = d.y;
+
             [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
                 B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
                 B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index b9bc0bfcdc389..0a48f84de307a 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -46,10 +46,12 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
 
         [[unroll]] for (uint n = 0; n < num_rows; ++n) {
             const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-            const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
 
-            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> v_im4) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid8%4+8]) >> (v_im4 + 2*(itid8/4)) & 0x3) << 4)) - 32);
-            barrier();
+            const uint32_t hmk = ~(uint32_t(data_a_packed16[ib0 + i].hmask[v_in]) | (uint32_t(data_a_packed16[ib0 + i].hmask[v_in + 8]) << 16));
+            const vec4 hmk_0 = vec4(unpack8(((hmk & hm_m[0]) >> (    v_im4)) << 2));
+            const vec4 hmk_1 = vec4(unpack8(((hmk & hm_m[1]) >> (1 + v_im4)) << 2));
+            const vec4 hmk_2 = vec4(unpack8(((hmk & hm_m[2]) >> (2 + v_im4)) << 2));
+            const vec4 hmk_3 = vec4(unpack8(((hmk & hm_m[3]) >> (3 + v_im4)) << 2));
 
             // 0, 1, 16, 17
             uint32_t qs_u32 = uint32_t(data_a[ib0 + i].qs[q_offset]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 1]) << 8);
@@ -59,14 +61,12 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
             const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
 
-            const uint32_t hmk = ~(uint32_t(data_a_packed16[ib0 + i].hmask[v_in]) | (uint32_t(data_a_packed16[ib0 + i].hmask[v_in + 8]) << 16));
-            const vec4 hmk_0 = vec4(unpack8(((hmk & hm_m[0]) >> (    v_im4)) << 2));
-            const vec4 hmk_1 = vec4(unpack8(((hmk & hm_m[1]) >> (1 + v_im4)) << 2));
-            const vec4 hmk_2 = vec4(unpack8(((hmk & hm_m[2]) >> (2 + v_im4)) << 2));
-            const vec4 hmk_3 = vec4(unpack8(((hmk & hm_m[3]) >> (3 + v_im4)) << 2));
+            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> v_im4) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid8%4+8]) >> (v_im4 + 2*(itid8/4)) & 0x3) << 4)) - 32);
+            barrier();
 
-            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
+            const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
 
+            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
                 B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
                 B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
                 B_TYPE_VEC2 b32 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16];

From 973bc4069f0d8da9e11c2612bc4b085e18d975af Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Thu, 9 Jan 2025 21:06:05 -0500
Subject: [PATCH 18/23] q3_k separate out calculation

---
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      | 121 ++++++++++--------
 .../vulkan-shaders/mul_mat_vec_q6_k.comp      |   3 +-
 2 files changed, 71 insertions(+), 53 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index 0a48f84de307a..e395b61438f80 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -6,6 +6,69 @@
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
 shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][8];
+FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
+
+void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, const uint itid8, const uint v_im, const uint v_im4, const uint v_in, const uint32_t hm_m[4], const uint q_offset, const uint y_offset, const uint s_shift, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) {
+    const uint y_idx = i * QUANT_K + y_offset;
+
+    [[unroll]] for (uint n = 0; n < num_rows; ++n) {
+        const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
+
+        if (!all_threads) { // when we don't have enough blocks to use all threads
+            if (i < num_blocks_per_row)
+                sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | ((data_a[ib0+i].scales[itid8%4+8] >> s_shift & 3) << 4)) - 32);
+            barrier();
+
+            if (i >= num_blocks_per_row)
+                continue;
+        }
+
+        const uint32_t hmk = ~(uint32_t(data_a_packed16[ib0 + i].hmask[v_in]) | (uint32_t(data_a_packed16[ib0 + i].hmask[v_in + 8]) << 16));
+        const vec4 hmk_0 = vec4(unpack8(((hmk & hm_m[0]) >> (    v_im4)) << 2));
+        const vec4 hmk_1 = vec4(unpack8(((hmk & hm_m[1]) >> (1 + v_im4)) << 2));
+        const vec4 hmk_2 = vec4(unpack8(((hmk & hm_m[2]) >> (2 + v_im4)) << 2));
+        const vec4 hmk_3 = vec4(unpack8(((hmk & hm_m[3]) >> (3 + v_im4)) << 2));
+
+        // 0, 1, 16, 17
+        uint32_t qs_u32 = uint32_t(data_a[ib0 + i].qs[q_offset]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 1]) << 8);
+        qs_u32 |= (uint32_t(data_a[ib0 + i].qs[q_offset + 16]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 17]) << 8)) << 16;
+        const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303));
+        const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303));
+        const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
+        const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
+
+        if (all_threads) {
+            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | ((data_a[ib0+i].scales[itid8%4+8] >> s_shift & 3) << 4)) - 32);
+            barrier();
+        }
+
+        const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
+
+        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
+            B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
+            B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
+            B_TYPE_VEC2 b32 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16];
+            B_TYPE_VEC2 b48 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24];
+            B_TYPE_VEC2 b64 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 32];
+            B_TYPE_VEC2 b80 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 40];
+            B_TYPE_VEC2 b96 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 48];
+            B_TYPE_VEC2 b112 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 56];
+
+            FLOAT_TYPE sum = FLOAT_TYPE(0.0);
+            [[unroll]] for (int l = 0; l < 2; ++l) {
+                sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], qs_u32_0[l  ] - hmk_0[l  ],
+                      fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], qs_u32_0[l+2] - hmk_0[l+2],
+                      fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], qs_u32_2[l  ] - hmk_1[l  ],
+                      fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], qs_u32_2[l+2] - hmk_1[l+2],
+                      fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], qs_u32_4[l  ] - hmk_2[l  ],
+                      fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], qs_u32_4[l+2] - hmk_2[l+2],
+                      fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], qs_u32_6[l  ] - hmk_3[l  ],
+                      fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], qs_u32_6[l+2] - hmk_3[l+2], sum))))))));
+            }
+            temp[j][n] = fma(d, sum, temp[j][n]);
+        }
+    }
+}
 
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
@@ -33,64 +96,20 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint q_offset = 32*v_im + l0;
     const uint y_offset = 128*v_im + l0;
 
-    FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
-
     [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
         [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) {
             temp[j][i] = FLOAT_TYPE(0);
         }
     }
 
-    [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) {
-        const uint y_idx = i * QUANT_K + y_offset;
-
-        [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-            const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-
-            const uint32_t hmk = ~(uint32_t(data_a_packed16[ib0 + i].hmask[v_in]) | (uint32_t(data_a_packed16[ib0 + i].hmask[v_in + 8]) << 16));
-            const vec4 hmk_0 = vec4(unpack8(((hmk & hm_m[0]) >> (    v_im4)) << 2));
-            const vec4 hmk_1 = vec4(unpack8(((hmk & hm_m[1]) >> (1 + v_im4)) << 2));
-            const vec4 hmk_2 = vec4(unpack8(((hmk & hm_m[2]) >> (2 + v_im4)) << 2));
-            const vec4 hmk_3 = vec4(unpack8(((hmk & hm_m[3]) >> (3 + v_im4)) << 2));
-
-            // 0, 1, 16, 17
-            uint32_t qs_u32 = uint32_t(data_a[ib0 + i].qs[q_offset]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 1]) << 8);
-            qs_u32 |= (uint32_t(data_a[ib0 + i].qs[q_offset + 16]) | (uint32_t(data_a[ib0 + i].qs[q_offset + 17]) << 8)) << 16;
-            const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303));
-            const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303));
-            const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
-            const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
+    const uint s_shift = v_im4 + 2*(itid8/4);
 
-            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((int8_t(data_a[ib0+i].scales[itid8]) >> v_im4) & 0xF) | ((int8_t(data_a[ib0+i].scales[itid8%4+8]) >> (v_im4 + 2*(itid8/4)) & 0x3) << 4)) - 32);
-            barrier();
-
-            const FLOAT_TYPE d = FLOAT_TYPE(data_a[ib0 + i].d);
-
-            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-                B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
-                B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
-                B_TYPE_VEC2 b32 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16];
-                B_TYPE_VEC2 b48 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24];
-                B_TYPE_VEC2 b64 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 32];
-                B_TYPE_VEC2 b80 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 40];
-                B_TYPE_VEC2 b96 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 48];
-                B_TYPE_VEC2 b112 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 56];
-
-                FLOAT_TYPE sum = FLOAT_TYPE(0.0);
-                [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum = fma(FLOAT_TYPE(  b0[l]) * sccache[ix][v_im][0], qs_u32_0[l  ] - hmk_0[l  ],
-                          fma(FLOAT_TYPE( b16[l]) * sccache[ix][v_im][1], qs_u32_0[l+2] - hmk_0[l+2],
-                          fma(FLOAT_TYPE( b32[l]) * sccache[ix][v_im][2], qs_u32_2[l  ] - hmk_1[l  ],
-                          fma(FLOAT_TYPE( b48[l]) * sccache[ix][v_im][3], qs_u32_2[l+2] - hmk_1[l+2],
-                          fma(FLOAT_TYPE( b64[l]) * sccache[ix][v_im][4], qs_u32_4[l  ] - hmk_2[l  ],
-                          fma(FLOAT_TYPE( b80[l]) * sccache[ix][v_im][5], qs_u32_4[l+2] - hmk_2[l+2],
-                          fma(FLOAT_TYPE( b96[l]) * sccache[ix][v_im][6], qs_u32_6[l  ] - hmk_3[l  ],
-                          fma(FLOAT_TYPE(b112[l]) * sccache[ix][v_im][7], qs_u32_6[l+2] - hmk_3[l+2], sum))))))));
-                }
-                temp[j][n] = fma(d, sum, temp[j][n]);
-            }
-        }
-    }
+    const uint nbr_par_th = num_blocks_per_row%it_size;
+    const uint nbr_all_th = num_blocks_per_row - nbr_par_th;
+    uint i0 = 0;
+    [[unroll]] for (; i0 < nbr_all_th; i0 += it_size)
+        calc_superblock(a_offset, b_offset, ix, itid8, v_im, v_im4, v_in, hm_m, q_offset, y_offset, s_shift, i0 + ix, num_blocks_per_row, first_row, num_rows, true);
+    calc_superblock(a_offset, b_offset, ix, itid8, v_im, v_im4, v_in, hm_m, q_offset, y_offset, s_shift, i0 + ix, num_blocks_per_row, first_row, num_rows, false);
 
     reduce_result(temp, d_offset, first_row, num_rows, tid);
 }
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
index efbdfe939cd87..55d928e6688e4 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
@@ -106,9 +106,8 @@ void compute_outputs(const uint first_row, const uint num_rows) {
     const uint nbr_par_th = num_blocks_per_row%it_size;
     const uint nbr_all_th = num_blocks_per_row - nbr_par_th;
     uint i0 = 0;
-    [[unroll]] for (; i0 < nbr_all_th; i0 += it_size) {
+    [[unroll]] for (; i0 < nbr_all_th; i0 += it_size)
         calc_superblock(a_offset, b_offset, itid, ix, ql_offset, qh_offset, s_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, true);
-    }
     calc_superblock(a_offset, b_offset, itid, ix, ql_offset, qh_offset, s_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, false);
 
     reduce_result(temp, d_offset, first_row, num_rows, tid);

From 6145fc79e5117959e49a667ea76f72649922e705 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Thu, 9 Jan 2025 21:41:50 -0500
Subject: [PATCH 19/23] q2_k separate out

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      | 130 ++++++++++--------
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      |   4 +-
 2 files changed, 76 insertions(+), 58 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 129442f7ec924..99db8f32ba20a 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -7,6 +7,74 @@ layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
 shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][16];
 
+FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
+
+void calc_superblock(const uint a_offset, const uint b_offset, const uint itid, const uint itid8, const uint v_im, const uint ix, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) {
+    const uint y_idx = i * QUANT_K + y_offset;
+
+    [[unroll]] for (uint n = 0; n < num_rows; ++n) {
+        const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
+
+        if (!all_threads) { // when we don't have enough blocks to use all threads
+            if (i < num_blocks_per_row) {
+                sccache[ix][0][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8] >> v_im*4) & 0xF); // lower 8 bytes
+                sccache[ix][1][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8+8] >> v_im*4) & 0xF); // upper 8 bytes
+                                                                                                                 }
+            barrier();
+
+            if (i >= num_blocks_per_row)
+                continue;
+        } else {
+            sccache[ix][0][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8] >> v_im*4) & 0xF);
+            sccache[ix][1][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8+8] >> v_im*4) & 0xF);
+            barrier();
+        }
+
+        const uint32_t qs_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
+        const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303));
+        const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303));
+        const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
+        const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
+
+        const f16vec2 d = data_a[ib0 + i].d;
+        const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
+        const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
+
+        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
+            B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
+            B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
+            B_TYPE_VEC2 b32 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16];
+            B_TYPE_VEC2 b48 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24];
+            B_TYPE_VEC2 b64 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 32];
+            B_TYPE_VEC2 b80 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 40];
+            B_TYPE_VEC2 b96 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 48];
+            B_TYPE_VEC2 b112 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 56];
+
+            FLOAT_TYPE sum1 = FLOAT_TYPE(0.0);
+            FLOAT_TYPE sum2 = FLOAT_TYPE(0.0);
+            [[unroll]] for (int l = 0; l < 2; ++l) {
+                sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * qs_u32_0[l  ],
+                       fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * qs_u32_0[l+2],
+                       fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * qs_u32_2[l  ],
+                       fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * qs_u32_2[l+2],
+                       fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * qs_u32_4[l  ],
+                       fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * qs_u32_4[l+2],
+                       fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * qs_u32_6[l  ],
+                       fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * qs_u32_6[l+2], sum1))))))));
+                sum2 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][ 8],
+                       fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][ 9],
+                       fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][10],
+                       fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][11],
+                       fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][12],
+                       fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][13],
+                       fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][14],
+                       fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][15], sum2))))))));
+            }
+            temp[j][n] = fma(dall, sum1, fma(-dmin, sum2, temp[j][n]));
+        }
+    }
+}
+
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
@@ -27,68 +95,18 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint q_offset = 32*v_im + l0;
     const uint y_offset = 128*v_im + l0;
 
-    FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
-
     [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
         [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) {
             temp[j][i] = FLOAT_TYPE(0);
         }
     }
 
-    [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) {
-        const uint y_idx = i * QUANT_K + y_offset;
-
-        [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-            const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-
-            sccache[ix][0][itid] = FLOAT_TYPE(bitfieldExtract(uint(data_a[ib0 + i].scales[itid8]), int(v_im*4), 4)); // lower 8 bytes
-            sccache[ix][1][itid] = FLOAT_TYPE(bitfieldExtract(uint(data_a[ib0 + i].scales[itid8+8]), int(v_im*4), 4)); // upper 8 bytes
-            barrier();
-
-            const uint32_t qs_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
-            const vec4 qs_u32_0 = vec4(unpack8(qs_u32 & 0x03030303));
-            const vec4 qs_u32_2 = vec4(unpack8((qs_u32 >> 2) & 0x03030303));
-            const vec4 qs_u32_4 = vec4(unpack8((qs_u32 >> 4) & 0x03030303));
-            const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
-
-            const f16vec2 d = data_a[ib0 + i].d;
-            const FLOAT_TYPE dall = d.x;
-            const FLOAT_TYPE dmin = d.y;
-
-            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-                B_TYPE_VEC2 b0 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2];
-                B_TYPE_VEC2 b16 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 8];
-                B_TYPE_VEC2 b32 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 16];
-                B_TYPE_VEC2 b48 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 24];
-                B_TYPE_VEC2 b64 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 32];
-                B_TYPE_VEC2 b80 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 40];
-                B_TYPE_VEC2 b96 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 48];
-                B_TYPE_VEC2 b112 = data_b_v2[(j*p.batch_stride_b + b_offset + y_idx) / 2 + 56];
-
-                FLOAT_TYPE sum1 = FLOAT_TYPE(0.0);
-                FLOAT_TYPE sum2 = FLOAT_TYPE(0.0);
-                [[unroll]] for (int l = 0; l < 2; ++l) {
-                    sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * qs_u32_0[l  ],
-                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * qs_u32_0[l+2],
-                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * qs_u32_2[l  ],
-                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * qs_u32_2[l+2],
-                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * qs_u32_4[l  ],
-                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * qs_u32_4[l+2],
-                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * qs_u32_6[l  ],
-                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * qs_u32_6[l+2], sum1))))))));
-                    sum2 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][ 8],
-                           fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][ 9],
-                           fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][10],
-                           fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][11],
-                           fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][12],
-                           fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][13],
-                           fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][14],
-                           fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][15], sum2))))))));
-                }
-                temp[j][n] = fma(dall, sum1, fma(-dmin, sum2, temp[j][n]));
-            }
-        }
-    }
+    const uint nbr_par_th = num_blocks_per_row%it_size;
+    const uint nbr_all_th = num_blocks_per_row - nbr_par_th;
+    uint i0 = 0;
+    [[unroll]] for (; i0 < nbr_all_th; i0 += it_size)
+        calc_superblock(a_offset, b_offset, itid, itid8, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, true);
+    calc_superblock(a_offset, b_offset, itid, itid8, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, false);
 
     reduce_result(temp, d_offset, first_row, num_rows, tid);
 }
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index e395b61438f80..6d655205d454e 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -16,7 +16,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, co
 
         if (!all_threads) { // when we don't have enough blocks to use all threads
             if (i < num_blocks_per_row)
-                sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | ((data_a[ib0+i].scales[itid8%4+8] >> s_shift & 3) << 4)) - 32);
+                sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | (((data_a[ib0+i].scales[itid8%4+8] >> s_shift) & 3) << 4)) - 32);
             barrier();
 
             if (i >= num_blocks_per_row)
@@ -38,7 +38,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, co
         const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
 
         if (all_threads) {
-            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | ((data_a[ib0+i].scales[itid8%4+8] >> s_shift & 3) << 4)) - 32);
+            sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | (((data_a[ib0+i].scales[itid8%4+8] >> s_shift) & 3) << 4)) - 32);
             barrier();
         }
 

From 845d572b877a94c91b756e8532787f7b9507458f Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Thu, 9 Jan 2025 21:58:26 -0500
Subject: [PATCH 20/23] little stuff

---
 ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp | 2 +-
 ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 99db8f32ba20a..9c41c55e54bb4 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -19,7 +19,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint itid,
             if (i < num_blocks_per_row) {
                 sccache[ix][0][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8] >> v_im*4) & 0xF); // lower 8 bytes
                 sccache[ix][1][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8+8] >> v_im*4) & 0xF); // upper 8 bytes
-                                                                                                                 }
+            }
             barrier();
 
             if (i >= num_blocks_per_row)
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index 6d655205d454e..a80764f9ce63e 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -6,6 +6,7 @@
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
 shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][8];
+
 FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
 
 void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, const uint itid8, const uint v_im, const uint v_im4, const uint v_in, const uint32_t hm_m[4], const uint q_offset, const uint y_offset, const uint s_shift, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) {

From 30eacad2905e375540fa800dccb4d62b053d3df5 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 11 Jan 2025 17:14:22 -0500
Subject: [PATCH 21/23] use calc_superblock everywhere

---
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      | 161 +++++++------
 .../vulkan-shaders/mul_mat_vec_q5_k.comp      | 225 +++++++++---------
 2 files changed, 196 insertions(+), 190 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index dc1540a5b7029..f9cde064887a8 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -6,6 +6,86 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
+FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
+
+void calc_superblock(const uint a_offset, const uint b_offset, const uint v_im, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows) {
+    const uint y1_idx = i * QUANT_K + y_offset;
+    const uint y2_idx = y1_idx + 128;
+
+    [[unroll]] for (uint n = 0; n < num_rows; ++n) {
+        const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
+        vec2 d = vec2(data_a[ib0 + i].d);
+        const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
+        const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
+
+        const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
+        const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
+        const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
+
+        const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32;
+        const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2;
+        const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F));
+        const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h));
+
+        const FLOAT_TYPE sc0 = scale_0_4_l_f.x;
+        const FLOAT_TYPE sc1 = scale_0_4_l_f.y;
+        const FLOAT_TYPE sc2 = scale_0_4_l_f.z;
+        const FLOAT_TYPE sc3 = scale_0_4_l_f.w;
+        const FLOAT_TYPE sc4 = scale8_f.x;
+        const FLOAT_TYPE sc5 = scale8_f.y;
+        const FLOAT_TYPE sc6 = scale8_f.z;
+        const FLOAT_TYPE sc7 = scale8_f.w;
+
+        const uint32_t qs0_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4];
+        const uint32_t qs64_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4 + 16];
+
+        const uint32_t qs0_u32_lo4 = qs0_u32 & 0x0F0F0F0F;
+        const uint32_t qs0_u32_hi4 = (qs0_u32 >> 4) & 0x0F0F0F0F;
+        const uint32_t qs64_u32_lo4 = qs64_u32 & 0x0F0F0F0F;
+        const uint32_t qs64_u32_hi4 = (qs64_u32 >> 4) & 0x0F0F0F0F;
+
+        const vec4 qs0_lo4 = vec4(unpack8(qs0_u32_lo4));
+        const vec4 qs64_lo4 = vec4(unpack8(qs64_u32_lo4));
+        const vec4 qs0_hi4 = vec4(unpack8(qs0_u32_hi4));
+        const vec4 qs64_hi4 = vec4(unpack8(qs64_u32_hi4));
+
+        const FLOAT_TYPE q4_0  = qs0_lo4.x;
+        const FLOAT_TYPE q4_1  = qs0_lo4.y;
+        const FLOAT_TYPE q4_2  = qs0_lo4.z;
+        const FLOAT_TYPE q4_3  = qs0_lo4.w;
+        const FLOAT_TYPE q4_4  = qs0_hi4.x;
+        const FLOAT_TYPE q4_5  = qs0_hi4.y;
+        const FLOAT_TYPE q4_6  = qs0_hi4.z;
+        const FLOAT_TYPE q4_7  = qs0_hi4.w;
+        const FLOAT_TYPE q4_8  = qs64_lo4.x;
+        const FLOAT_TYPE q4_9  = qs64_lo4.y;
+        const FLOAT_TYPE q4_10 = qs64_lo4.z;
+        const FLOAT_TYPE q4_11 = qs64_lo4.w;
+        const FLOAT_TYPE q4_12 = qs64_hi4.x;
+        const FLOAT_TYPE q4_13 = qs64_hi4.y;
+        const FLOAT_TYPE q4_14 = qs64_hi4.z;
+        const FLOAT_TYPE q4_15 = qs64_hi4.w;
+
+        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
+            vec4 by10 =  vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y1_idx) / 4    ]);
+            vec4 by132 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y1_idx) / 4 + 8]);
+            vec4 by20 =  vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y2_idx) / 4    ]);
+            vec4 by232 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y2_idx) / 4 + 8]);
+
+            const FLOAT_TYPE sx = fma(FLOAT_TYPE(by10.x),      q4_0,  fma(FLOAT_TYPE(by10.y),  q4_1,  fma(FLOAT_TYPE(by10.z),  q4_2,  FLOAT_TYPE(by10.w) *  q4_3)));
+            const FLOAT_TYPE sy = fma(FLOAT_TYPE(by132.x),     q4_4,  fma(FLOAT_TYPE(by132.y), q4_5,  fma(FLOAT_TYPE(by132.z), q4_6,  FLOAT_TYPE(by132.w) * q4_7)));
+            const FLOAT_TYPE sz = fma(FLOAT_TYPE(by20.x),      q4_8,  fma(FLOAT_TYPE(by20.y),  q4_9,  fma(FLOAT_TYPE(by20.z),  q4_10, FLOAT_TYPE(by20.w) *  q4_11)));
+            const FLOAT_TYPE sw = fma(FLOAT_TYPE(by232.x),     q4_12, fma(FLOAT_TYPE(by232.y), q4_13, fma(FLOAT_TYPE(by232.z), q4_14, FLOAT_TYPE(by232.w) * q4_15)));
+            const FLOAT_TYPE smin =
+                fma(FLOAT_TYPE(by10.x), sc2, fma(FLOAT_TYPE(by132.x), sc3, fma(FLOAT_TYPE(by20.x), sc6, fma(FLOAT_TYPE(by232.x), sc7,
+                fma(FLOAT_TYPE(by10.y), sc2, fma(FLOAT_TYPE(by132.y), sc3, fma(FLOAT_TYPE(by20.y), sc6, fma(FLOAT_TYPE(by232.y), sc7,
+                fma(FLOAT_TYPE(by10.z), sc2, fma(FLOAT_TYPE(by132.z), sc3, fma(FLOAT_TYPE(by20.z), sc6, fma(FLOAT_TYPE(by232.z), sc7,
+                fma(FLOAT_TYPE(by10.w), sc2, fma(FLOAT_TYPE(by132.w), sc3, fma(FLOAT_TYPE(by20.w), sc6,     FLOAT_TYPE(by232.w) * sc7)))))))))))))));
+            temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n]));
+        }
+    }
+}
+
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
@@ -29,91 +109,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint q_offset = 32*v_im + l0;
     const uint y_offset = 64*v_im + l0;
 
-    FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
-
     [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
         [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) {
             temp[j][i] = FLOAT_TYPE(0);
         }
     }
 
-    [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) {
-        const uint y1_idx = i * QUANT_K + y_offset;
-        const uint y2_idx = y1_idx + 128;
-
-        [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-            const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-            vec2 d = vec2(data_a[ib0 + i].d);
-            const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
-            const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
-
-            const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
-            const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-
-            const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32;
-            const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2;
-            const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F));
-            const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h));
-
-            const FLOAT_TYPE sc0 = scale_0_4_l_f.x;
-            const FLOAT_TYPE sc1 = scale_0_4_l_f.y;
-            const FLOAT_TYPE sc2 = scale_0_4_l_f.z;
-            const FLOAT_TYPE sc3 = scale_0_4_l_f.w;
-            const FLOAT_TYPE sc4 = scale8_f.x;
-            const FLOAT_TYPE sc5 = scale8_f.y;
-            const FLOAT_TYPE sc6 = scale8_f.z;
-            const FLOAT_TYPE sc7 = scale8_f.w;
-
-            const uint32_t qs0_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4];
-            const uint32_t qs64_u32 = data_a_packed32[ib0 + i].qs[q_offset / 4 + 16];
-
-            const uint32_t qs0_u32_lo4 = qs0_u32 & 0x0F0F0F0F;
-            const uint32_t qs0_u32_hi4 = (qs0_u32 >> 4) & 0x0F0F0F0F;
-            const uint32_t qs64_u32_lo4 = qs64_u32 & 0x0F0F0F0F;
-            const uint32_t qs64_u32_hi4 = (qs64_u32 >> 4) & 0x0F0F0F0F;
-
-            const vec4 qs0_lo4 = vec4(unpack8(qs0_u32_lo4));
-            const vec4 qs64_lo4 = vec4(unpack8(qs64_u32_lo4));
-            const vec4 qs0_hi4 = vec4(unpack8(qs0_u32_hi4));
-            const vec4 qs64_hi4 = vec4(unpack8(qs64_u32_hi4));
-
-            const FLOAT_TYPE q4_0  = qs0_lo4.x;
-            const FLOAT_TYPE q4_1  = qs0_lo4.y;
-            const FLOAT_TYPE q4_2  = qs0_lo4.z;
-            const FLOAT_TYPE q4_3  = qs0_lo4.w;
-            const FLOAT_TYPE q4_4  = qs0_hi4.x;
-            const FLOAT_TYPE q4_5  = qs0_hi4.y;
-            const FLOAT_TYPE q4_6  = qs0_hi4.z;
-            const FLOAT_TYPE q4_7  = qs0_hi4.w;
-            const FLOAT_TYPE q4_8  = qs64_lo4.x;
-            const FLOAT_TYPE q4_9  = qs64_lo4.y;
-            const FLOAT_TYPE q4_10 = qs64_lo4.z;
-            const FLOAT_TYPE q4_11 = qs64_lo4.w;
-            const FLOAT_TYPE q4_12 = qs64_hi4.x;
-            const FLOAT_TYPE q4_13 = qs64_hi4.y;
-            const FLOAT_TYPE q4_14 = qs64_hi4.z;
-            const FLOAT_TYPE q4_15 = qs64_hi4.w;
-
-            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-                vec4 by10 =  vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y1_idx) / 4    ]);
-                vec4 by132 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y1_idx) / 4 + 8]);
-                vec4 by20 =  vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y2_idx) / 4    ]);
-                vec4 by232 = vec4(data_b_v4[(j*p.batch_stride_b + b_offset + y2_idx) / 4 + 8]);
-
-                const FLOAT_TYPE sx = fma(FLOAT_TYPE(by10.x),      q4_0,  fma(FLOAT_TYPE(by10.y),  q4_1,  fma(FLOAT_TYPE(by10.z),  q4_2,  FLOAT_TYPE(by10.w) *  q4_3)));
-                const FLOAT_TYPE sy = fma(FLOAT_TYPE(by132.x),     q4_4,  fma(FLOAT_TYPE(by132.y), q4_5,  fma(FLOAT_TYPE(by132.z), q4_6,  FLOAT_TYPE(by132.w) * q4_7)));
-                const FLOAT_TYPE sz = fma(FLOAT_TYPE(by20.x),      q4_8,  fma(FLOAT_TYPE(by20.y),  q4_9,  fma(FLOAT_TYPE(by20.z),  q4_10, FLOAT_TYPE(by20.w) *  q4_11)));
-                const FLOAT_TYPE sw = fma(FLOAT_TYPE(by232.x),     q4_12, fma(FLOAT_TYPE(by232.y), q4_13, fma(FLOAT_TYPE(by232.z), q4_14, FLOAT_TYPE(by232.w) * q4_15)));
-                const FLOAT_TYPE smin =
-                    fma(FLOAT_TYPE(by10.x), sc2, fma(FLOAT_TYPE(by132.x), sc3, fma(FLOAT_TYPE(by20.x), sc6, fma(FLOAT_TYPE(by232.x), sc7,
-                    fma(FLOAT_TYPE(by10.y), sc2, fma(FLOAT_TYPE(by132.y), sc3, fma(FLOAT_TYPE(by20.y), sc6, fma(FLOAT_TYPE(by232.y), sc7,
-                    fma(FLOAT_TYPE(by10.z), sc2, fma(FLOAT_TYPE(by132.z), sc3, fma(FLOAT_TYPE(by20.z), sc6, fma(FLOAT_TYPE(by232.z), sc7,
-                    fma(FLOAT_TYPE(by10.w), sc2, fma(FLOAT_TYPE(by132.w), sc3, fma(FLOAT_TYPE(by20.w), sc6,     FLOAT_TYPE(by232.w) * sc7)))))))))))))));
-                temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n]));
-            }
-        }
-    }
+    [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size)
+        calc_superblock(a_offset, b_offset, v_im, q_offset, y_offset, i, num_blocks_per_row, first_row, num_rows);
 
     reduce_result(temp, d_offset, first_row, num_rows, tid);
 }
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
index df2a8bd20913b..6c84ef3cde3ff 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
@@ -6,6 +6,118 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
+FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
+
+void calc_superblock(const uint a_offset, const uint b_offset, const uint v_im, const uint l0, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows) {
+    const uint y1_idx = i * QUANT_K + y_offset;
+    const uint y2_idx = y1_idx + 128;
+
+    [[unroll]] for (uint n = 0; n < num_rows; ++n) {
+        const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
+        vec2 d = vec2(data_a[ib0 + i].d);
+        const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
+        const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
+
+        const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
+        const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
+        const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
+
+        const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32;
+        const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2;
+        const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F));
+        const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h));
+
+        const FLOAT_TYPE sc0 = scale_0_4_l_f.x;
+        const FLOAT_TYPE sc1 = scale_0_4_l_f.y;
+        const FLOAT_TYPE sc2 = scale_0_4_l_f.z;
+        const FLOAT_TYPE sc3 = scale_0_4_l_f.w;
+        const FLOAT_TYPE sc4 = scale8_f.x;
+        const FLOAT_TYPE sc5 = scale8_f.y;
+        const FLOAT_TYPE sc6 = scale8_f.z;
+        const FLOAT_TYPE sc7 = scale8_f.w;
+
+        const uint32_t qs0_16_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
+        const uint32_t qs64_80_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 32]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 40]) << 16);
+
+        uint32_t qs0_16_u32_lo4 = qs0_16_u32 & 0x0F0F0F0F;
+        uint32_t qs0_16_u32_hi4 = (qs0_16_u32 >> 4) & 0x0F0F0F0F;
+        uint32_t qs64_80_u32_lo4 = qs64_80_u32 & 0x0F0F0F0F;
+        uint32_t qs64_80_u32_hi4 = (qs64_80_u32 >> 4) & 0x0F0F0F0F;
+
+        const uint32_t qh = pack32(u16vec2(data_a_packed16[ib0 + i].qh[l0 / 2], data_a_packed16[ib0 + i].qh[l0 / 2 + 8]));
+
+        const uint32_t qs0_16_lo4_offset16 = ((qh >> (2*v_im)) & 0x01010101) << 4;
+        const uint32_t qs0_16_hi4_offset16 = ((qh >> (2*v_im)) & 0x02020202) << 3;
+        const uint32_t qs64_80_lo4_offset16 = ((qh >> (2*v_im)) & 0x10101010);
+        const uint32_t qs64_80_hi4_offset16 = ((qh >> (2*v_im)) & 0x20202020) >> 1;
+
+        qs0_16_u32_lo4 += qs0_16_lo4_offset16;
+        qs0_16_u32_hi4 += qs0_16_hi4_offset16;
+        qs64_80_u32_lo4 += qs64_80_lo4_offset16;
+        qs64_80_u32_hi4 += qs64_80_hi4_offset16;
+
+        const vec4 qs0_16_lo4 = vec4(unpack8(qs0_16_u32_lo4));
+        const vec4 qs64_80_lo4 = vec4(unpack8(qs64_80_u32_lo4));
+        const vec4 qs0_16_hi4 = vec4(unpack8(qs0_16_u32_hi4));
+        const vec4 qs64_80_hi4 = vec4(unpack8(qs64_80_u32_hi4));
+
+        const FLOAT_TYPE q4_0  = qs0_16_lo4.x;
+        const FLOAT_TYPE q4_1  = qs0_16_lo4.y;
+        const FLOAT_TYPE q4_2  = qs0_16_lo4.z;
+        const FLOAT_TYPE q4_3  = qs0_16_lo4.w;
+        const FLOAT_TYPE q4_4  = qs0_16_hi4.x;
+        const FLOAT_TYPE q4_5  = qs0_16_hi4.y;
+        const FLOAT_TYPE q4_6  = qs0_16_hi4.z;
+        const FLOAT_TYPE q4_7  = qs0_16_hi4.w;
+        const FLOAT_TYPE q4_8  = qs64_80_lo4.x;
+        const FLOAT_TYPE q4_9  = qs64_80_lo4.y;
+        const FLOAT_TYPE q4_10 = qs64_80_lo4.z;
+        const FLOAT_TYPE q4_11 = qs64_80_lo4.w;
+        const FLOAT_TYPE q4_12 = qs64_80_hi4.x;
+        const FLOAT_TYPE q4_13 = qs64_80_hi4.y;
+        const FLOAT_TYPE q4_14 = qs64_80_hi4.z;
+        const FLOAT_TYPE q4_15 = qs64_80_hi4.w;
+
+        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
+            vec2 by10 =  vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2     ]);
+            vec2 by116 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 +  8]);
+            vec2 by132 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 + 16]);
+            vec2 by148 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 + 24]);
+            vec2 by20 =  vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2     ]);
+            vec2 by216 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 +  8]);
+            vec2 by232 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 + 16]);
+            vec2 by248 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 + 24]);
+
+            const FLOAT_TYPE sx =
+              fma(FLOAT_TYPE(by10.x), q4_0,
+              fma(FLOAT_TYPE(by10.y), q4_1,
+              fma(FLOAT_TYPE(by116.x), q4_2,
+                 FLOAT_TYPE(by116.y) * q4_3)));
+            const FLOAT_TYPE sy =
+              fma(FLOAT_TYPE(by132.x), q4_4,
+              fma(FLOAT_TYPE(by132.y), q4_5,
+              fma(FLOAT_TYPE(by148.x), q4_6,
+                 FLOAT_TYPE(by148.y) * q4_7)));
+            const FLOAT_TYPE sz =
+              fma(FLOAT_TYPE(by20.x), q4_8,
+              fma(FLOAT_TYPE(by20.y), q4_9,
+              fma(FLOAT_TYPE(by216.x), q4_10,
+                 FLOAT_TYPE(by216.y) * q4_11)));
+            const FLOAT_TYPE sw =
+              fma(FLOAT_TYPE(by232.x), q4_12,
+              fma(FLOAT_TYPE(by232.y), q4_13,
+              fma(FLOAT_TYPE(by248.x), q4_14,
+                 FLOAT_TYPE(by248.y) * q4_15)));
+            const FLOAT_TYPE smin =
+              fma(FLOAT_TYPE(by10.x) + FLOAT_TYPE(by10.y) + FLOAT_TYPE(by116.x) + FLOAT_TYPE(by116.y), sc2,
+              fma(FLOAT_TYPE(by132.x) + FLOAT_TYPE(by132.y) + FLOAT_TYPE(by148.x) + FLOAT_TYPE(by148.y), sc3,
+              fma(FLOAT_TYPE(by20.x) + FLOAT_TYPE(by20.y) + FLOAT_TYPE(by216.x) + FLOAT_TYPE(by216.y), sc6,
+                  (FLOAT_TYPE(by232.x) + FLOAT_TYPE(by232.y) + FLOAT_TYPE(by248.x) + FLOAT_TYPE(by248.y)) * sc7)));
+            temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n]));
+        }
+    }
+}
+
 void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     uint a_offset, b_offset, d_offset;
     get_offsets(a_offset, b_offset, d_offset);
@@ -28,123 +140,14 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint q_offset = 32*v_im + l0;
     const uint y_offset = 64*v_im + l0;
 
-    FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
-
     [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
         [[unroll]] for (uint i = 0; i < NUM_ROWS; ++i) {
             temp[j][i] = FLOAT_TYPE(0);
         }
     }
 
-    [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size) {
-        const uint y1_idx = i * QUANT_K + y_offset;
-        const uint y2_idx = y1_idx + 128;
-
-        [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-            const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
-            vec2 d = vec2(data_a[ib0 + i].d);
-            const FLOAT_TYPE dall = FLOAT_TYPE(d.x);
-            const FLOAT_TYPE dmin = FLOAT_TYPE(d.y);
-
-            const uint32_t scale0_u32 = data_a_packed16[ib0 + i].scales[v_im    ];
-            const uint32_t scale4_u32 = data_a_packed16[ib0 + i].scales[v_im + 2];
-            const uint32_t scale8_u32 = data_a_packed16[ib0 + i].scales[v_im + 4];
-
-            const uint32_t scale_0_4_l = (scale4_u32 << 16) | scale0_u32;
-            const uint32_t scale_0_4_h = (scale_0_4_l & 0xC0C0C0C0) >> 2;
-            const vec4 scale_0_4_l_f = vec4(unpack8(scale_0_4_l & 0x3F3F3F3F));
-            const vec4 scale8_f = vec4(unpack8((((scale8_u32 << 12) | scale8_u32) & 0x0F0F0F0F) | scale_0_4_h));
-
-            const FLOAT_TYPE sc0 = scale_0_4_l_f.x;
-            const FLOAT_TYPE sc1 = scale_0_4_l_f.y;
-            const FLOAT_TYPE sc2 = scale_0_4_l_f.z;
-            const FLOAT_TYPE sc3 = scale_0_4_l_f.w;
-            const FLOAT_TYPE sc4 = scale8_f.x;
-            const FLOAT_TYPE sc5 = scale8_f.y;
-            const FLOAT_TYPE sc6 = scale8_f.z;
-            const FLOAT_TYPE sc7 = scale8_f.w;
-
-            const uint32_t qs0_16_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 8]) << 16);
-            const uint32_t qs64_80_u32 = uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 32]) | (uint32_t(data_a_packed16[ib0 + i].qs[q_offset / 2 + 40]) << 16);
-
-            uint32_t qs0_16_u32_lo4 = qs0_16_u32 & 0x0F0F0F0F;
-            uint32_t qs0_16_u32_hi4 = (qs0_16_u32 >> 4) & 0x0F0F0F0F;
-            uint32_t qs64_80_u32_lo4 = qs64_80_u32 & 0x0F0F0F0F;
-            uint32_t qs64_80_u32_hi4 = (qs64_80_u32 >> 4) & 0x0F0F0F0F;
-
-            const uint32_t qh = pack32(u16vec2(data_a_packed16[ib0 + i].qh[l0 / 2], data_a_packed16[ib0 + i].qh[l0 / 2 + 8]));
-
-            const uint32_t qs0_16_lo4_offset16 = ((qh >> (2*v_im)) & 0x01010101) << 4;
-            const uint32_t qs0_16_hi4_offset16 = ((qh >> (2*v_im)) & 0x02020202) << 3;
-            const uint32_t qs64_80_lo4_offset16 = ((qh >> (2*v_im)) & 0x10101010);
-            const uint32_t qs64_80_hi4_offset16 = ((qh >> (2*v_im)) & 0x20202020) >> 1;
-
-            qs0_16_u32_lo4 += qs0_16_lo4_offset16;
-            qs0_16_u32_hi4 += qs0_16_hi4_offset16;
-            qs64_80_u32_lo4 += qs64_80_lo4_offset16;
-            qs64_80_u32_hi4 += qs64_80_hi4_offset16;
-
-            const vec4 qs0_16_lo4 = vec4(unpack8(qs0_16_u32_lo4));
-            const vec4 qs64_80_lo4 = vec4(unpack8(qs64_80_u32_lo4));
-            const vec4 qs0_16_hi4 = vec4(unpack8(qs0_16_u32_hi4));
-            const vec4 qs64_80_hi4 = vec4(unpack8(qs64_80_u32_hi4));
-
-            const FLOAT_TYPE q4_0  = qs0_16_lo4.x;
-            const FLOAT_TYPE q4_1  = qs0_16_lo4.y;
-            const FLOAT_TYPE q4_2  = qs0_16_lo4.z;
-            const FLOAT_TYPE q4_3  = qs0_16_lo4.w;
-            const FLOAT_TYPE q4_4  = qs0_16_hi4.x;
-            const FLOAT_TYPE q4_5  = qs0_16_hi4.y;
-            const FLOAT_TYPE q4_6  = qs0_16_hi4.z;
-            const FLOAT_TYPE q4_7  = qs0_16_hi4.w;
-            const FLOAT_TYPE q4_8  = qs64_80_lo4.x;
-            const FLOAT_TYPE q4_9  = qs64_80_lo4.y;
-            const FLOAT_TYPE q4_10 = qs64_80_lo4.z;
-            const FLOAT_TYPE q4_11 = qs64_80_lo4.w;
-            const FLOAT_TYPE q4_12 = qs64_80_hi4.x;
-            const FLOAT_TYPE q4_13 = qs64_80_hi4.y;
-            const FLOAT_TYPE q4_14 = qs64_80_hi4.z;
-            const FLOAT_TYPE q4_15 = qs64_80_hi4.w;
-
-            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-                vec2 by10 =  vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2     ]);
-                vec2 by116 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 +  8]);
-                vec2 by132 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 + 16]);
-                vec2 by148 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y1_idx) / 2 + 24]);
-                vec2 by20 =  vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2     ]);
-                vec2 by216 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 +  8]);
-                vec2 by232 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 + 16]);
-                vec2 by248 = vec2(data_b_v2[(j*p.batch_stride_b + b_offset + y2_idx) / 2 + 24]);
-
-                const FLOAT_TYPE sx =
-                  fma(FLOAT_TYPE(by10.x), q4_0,
-                  fma(FLOAT_TYPE(by10.y), q4_1,
-                  fma(FLOAT_TYPE(by116.x), q4_2,
-                     FLOAT_TYPE(by116.y) * q4_3)));
-                const FLOAT_TYPE sy =
-                  fma(FLOAT_TYPE(by132.x), q4_4,
-                  fma(FLOAT_TYPE(by132.y), q4_5,
-                  fma(FLOAT_TYPE(by148.x), q4_6,
-                     FLOAT_TYPE(by148.y) * q4_7)));
-                const FLOAT_TYPE sz =
-                  fma(FLOAT_TYPE(by20.x), q4_8,
-                  fma(FLOAT_TYPE(by20.y), q4_9,
-                  fma(FLOAT_TYPE(by216.x), q4_10,
-                     FLOAT_TYPE(by216.y) * q4_11)));
-                const FLOAT_TYPE sw =
-                  fma(FLOAT_TYPE(by232.x), q4_12,
-                  fma(FLOAT_TYPE(by232.y), q4_13,
-                  fma(FLOAT_TYPE(by248.x), q4_14,
-                     FLOAT_TYPE(by248.y) * q4_15)));
-                const FLOAT_TYPE smin =
-                  fma(FLOAT_TYPE(by10.x) + FLOAT_TYPE(by10.y) + FLOAT_TYPE(by116.x) + FLOAT_TYPE(by116.y), sc2,
-                  fma(FLOAT_TYPE(by132.x) + FLOAT_TYPE(by132.y) + FLOAT_TYPE(by148.x) + FLOAT_TYPE(by148.y), sc3,
-                  fma(FLOAT_TYPE(by20.x) + FLOAT_TYPE(by20.y) + FLOAT_TYPE(by216.x) + FLOAT_TYPE(by216.y), sc6,
-                      (FLOAT_TYPE(by232.x) + FLOAT_TYPE(by232.y) + FLOAT_TYPE(by248.x) + FLOAT_TYPE(by248.y)) * sc7)));
-                temp[j][n] = fma(dall, fma(sx, sc0, fma(sy, sc1, fma(sz, sc4, sw * sc5))), fma(-dmin, smin, temp[j][n]));
-            }
-        }
-    }
+    [[unroll]] for (uint i = ix; i < num_blocks_per_row; i += it_size)
+        calc_superblock(a_offset, b_offset, v_im, l0, q_offset, y_offset, i, num_blocks_per_row, first_row, num_rows);
 
     reduce_result(temp, d_offset, first_row, num_rows, tid);
 }

From ed1ad94c8411213d7a22eedfa47fd7e205783e90 Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 11 Jan 2025 21:46:53 -0500
Subject: [PATCH 22/23] q2_k optimize scale calculation

---
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      | 52 ++++++++++---------
 1 file changed, 27 insertions(+), 25 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 45d6a830c28cc..afa109262d9b3 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -5,11 +5,12 @@
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
-shared FLOAT_TYPE sccache[BLOCK_SIZE/16][2][16];
+shared FLOAT_TYPE sccache1[BLOCK_SIZE/16][16];
+shared FLOAT_TYPE sccache2[BLOCK_SIZE/16][16];
 
 FLOAT_TYPE temp[NUM_COLS][NUM_ROWS];
 
-void calc_superblock(const uint a_offset, const uint b_offset, const uint itid, const uint itid8, const uint v_im, const uint ix, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) {
+void calc_superblock(const uint a_offset, const uint b_offset, const uint itid, const uint v_im, const uint ix, const uint q_offset, const uint y_offset, const uint i, const uint num_blocks_per_row, const uint first_row, const uint num_rows, const bool all_threads) {
     const uint y_idx = i * QUANT_K + y_offset;
 
     [[unroll]] for (uint n = 0; n < num_rows; ++n) {
@@ -17,16 +18,18 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint itid,
 
         if (!all_threads) { // when we don't have enough blocks to use all threads
             if (i < num_blocks_per_row) {
-                sccache[ix][0][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8] >> v_im*4) & 0xF); // lower 8 bytes
-                sccache[ix][1][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8+8] >> v_im*4) & 0xF); // upper 8 bytes
+                const uint32_t scale = uint32_t(data_a[ib0 + i].scales[itid]);
+                sccache1[ix][itid] = FLOAT_TYPE(scale & 0xF);
+                sccache2[ix][itid] = FLOAT_TYPE((scale >> 4) & 0xF);
             }
             barrier();
 
             if (i >= num_blocks_per_row)
                 continue;
         } else {
-            sccache[ix][0][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8] >> v_im*4) & 0xF);
-            sccache[ix][1][itid] = FLOAT_TYPE((data_a[ib0 + i].scales[itid8+8] >> v_im*4) & 0xF);
+            const uint32_t scale = uint32_t(data_a[ib0 + i].scales[itid]);
+            sccache1[ix][itid] = FLOAT_TYPE(scale & 0xF);
+            sccache2[ix][itid] = FLOAT_TYPE((scale >> 4) & 0xF);
             barrier();
         }
 
@@ -53,22 +56,22 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint itid,
             FLOAT_TYPE sum1 = FLOAT_TYPE(0.0);
             FLOAT_TYPE sum2 = FLOAT_TYPE(0.0);
             [[unroll]] for (int l = 0; l < 2; ++l) {
-                sum1 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][0] * qs_u32_0[l  ],
-                       fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][1] * qs_u32_0[l+2],
-                       fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][2] * qs_u32_2[l  ],
-                       fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][3] * qs_u32_2[l+2],
-                       fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][4] * qs_u32_4[l  ],
-                       fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][5] * qs_u32_4[l+2],
-                       fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][6] * qs_u32_6[l  ],
-                       fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][7] * qs_u32_6[l+2], sum1))))))));
-                sum2 = fma(FLOAT_TYPE(b0[l]),   sccache[ix][v_im][ 8],
-                       fma(FLOAT_TYPE(b16[l]),  sccache[ix][v_im][ 9],
-                       fma(FLOAT_TYPE(b32[l]),  sccache[ix][v_im][10],
-                       fma(FLOAT_TYPE(b48[l]),  sccache[ix][v_im][11],
-                       fma(FLOAT_TYPE(b64[l]),  sccache[ix][v_im][12],
-                       fma(FLOAT_TYPE(b80[l]),  sccache[ix][v_im][13],
-                       fma(FLOAT_TYPE(b96[l]),  sccache[ix][v_im][14],
-                       fma(FLOAT_TYPE(b112[l]), sccache[ix][v_im][15], sum2))))))));
+                sum1 = fma(FLOAT_TYPE(b0[l]),   sccache1[ix][    8*v_im] * qs_u32_0[l  ],
+                       fma(FLOAT_TYPE(b16[l]),  sccache1[ix][1 + 8*v_im] * qs_u32_0[l+2],
+                       fma(FLOAT_TYPE(b32[l]),  sccache1[ix][2 + 8*v_im] * qs_u32_2[l  ],
+                       fma(FLOAT_TYPE(b48[l]),  sccache1[ix][3 + 8*v_im] * qs_u32_2[l+2],
+                       fma(FLOAT_TYPE(b64[l]),  sccache1[ix][4 + 8*v_im] * qs_u32_4[l  ],
+                       fma(FLOAT_TYPE(b80[l]),  sccache1[ix][5 + 8*v_im] * qs_u32_4[l+2],
+                       fma(FLOAT_TYPE(b96[l]),  sccache1[ix][6 + 8*v_im] * qs_u32_6[l  ],
+                       fma(FLOAT_TYPE(b112[l]), sccache1[ix][7 + 8*v_im] * qs_u32_6[l+2], sum1))))))));
+                sum2 = fma(FLOAT_TYPE(b0[l]),   sccache2[ix][    8*v_im],
+                       fma(FLOAT_TYPE(b16[l]),  sccache2[ix][1 + 8*v_im],
+                       fma(FLOAT_TYPE(b32[l]),  sccache2[ix][2 + 8*v_im],
+                       fma(FLOAT_TYPE(b48[l]),  sccache2[ix][3 + 8*v_im],
+                       fma(FLOAT_TYPE(b64[l]),  sccache2[ix][4 + 8*v_im],
+                       fma(FLOAT_TYPE(b80[l]),  sccache2[ix][5 + 8*v_im],
+                       fma(FLOAT_TYPE(b96[l]),  sccache2[ix][6 + 8*v_im],
+                       fma(FLOAT_TYPE(b112[l]), sccache2[ix][7 + 8*v_im], sum2))))))));
             }
             temp[j][n] = fma(dall, sum1, fma(-dmin, sum2, temp[j][n]));
         }
@@ -86,7 +89,6 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint tid = gl_LocalInvocationID.x;
     const uint itid = tid%16;  // 0...15
     const uint ix = tid/16;
-    const uint itid8 = itid%8;
 
     const uint v_im = itid/8;                                // 0 or 1. 0 computes 0..., 1 computes 128...
     const uint v_in = itid - 8*v_im;                         // 0...7
@@ -105,8 +107,8 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     const uint nbr_all_th = num_blocks_per_row - nbr_par_th;
     uint i0 = 0;
     [[unroll]] for (; i0 < nbr_all_th; i0 += it_size)
-        calc_superblock(a_offset, b_offset, itid, itid8, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, true);
-    calc_superblock(a_offset, b_offset, itid, itid8, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, false);
+        calc_superblock(a_offset, b_offset, itid, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, true);
+    calc_superblock(a_offset, b_offset, itid, v_im, ix, q_offset, y_offset, i0 + ix, num_blocks_per_row, first_row, num_rows, false);
 
     reduce_result(temp, d_offset, first_row, num_rows, tid);
 }

From 4ae3fc01552cb3a3a2f210fafb7f277c41cb906f Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sun, 12 Jan 2025 17:21:57 -0500
Subject: [PATCH 23/23] more barriers

---
 ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp | 1 +
 ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp | 2 ++
 ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp | 2 ++
 3 files changed, 5 insertions(+)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index afa109262d9b3..8cdc640e80e31 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -16,6 +16,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint itid,
     [[unroll]] for (uint n = 0; n < num_rows; ++n) {
         const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
 
+        barrier();
         if (!all_threads) { // when we don't have enough blocks to use all threads
             if (i < num_blocks_per_row) {
                 const uint32_t scale = uint32_t(data_a[ib0 + i].scales[itid]);
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index 58be53bb8cd05..3116fad165be0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -16,6 +16,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, co
         const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
 
         if (!all_threads) { // when we don't have enough blocks to use all threads
+            barrier();
             if (i < num_blocks_per_row)
                 sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | (((data_a[ib0+i].scales[itid8%4+8] >> s_shift) & 3) << 4)) - 32);
             barrier();
@@ -39,6 +40,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint ix, co
         const vec4 qs_u32_6 = vec4(unpack8((qs_u32 >> 6) & 0x03030303));
 
         if (all_threads) {
+            barrier();
             sccache[ix][v_im][itid8] = FLOAT_TYPE(int8_t(((data_a[ib0+i].scales[itid8] >> v_im4) & 0xF) | (((data_a[ib0+i].scales[itid8%4+8] >> s_shift) & 3) << 4)) - 32);
             barrier();
         }
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
index b362c46bb07c5..f05f96b5efb9d 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
@@ -17,6 +17,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint itid,
         const uint ib0 = a_offset / QUANT_K + (first_row+n)*num_blocks_per_row;
 
         if (!all_threads) { // when we don't have enough blocks to use all threads
+            barrier();
             if (i < num_blocks_per_row)
                 sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]);
             barrier();
@@ -50,6 +51,7 @@ void calc_superblock(const uint a_offset, const uint b_offset, const uint itid,
         const vec4 q3 = vec4(unpack8(q3_u32)) - 32;
 
         if (all_threads) {
+            barrier();
             sccache[ix][itid] = FLOAT_TYPE(data_a[ib0 + i].scales[itid]);
             barrier();
         }