Skip to content

LLVM optimization on PR 164059 [bot] #80

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: pr-164059-before-1760775797
Choose a base branch
from
Open

LLVM optimization on PR 164059 [bot] #80

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
761 changes: 316 additions & 445 deletions results/MultiSource/Applications/oggenc/CMakeFiles/oggenc.dir/oggenc.s
View file
Open in desktop

Large diffs are not rendered by default.

369 changes: 170 additions & 199 deletions results/MultiSource/Applications/sgefa/CMakeFiles/sgefa.dir/driver.s
View file
Open in desktop

Large diffs are not rendered by default.

40 changes: 18 additions & 22 deletions results/MultiSource/Benchmarks/Bullet/CMakeFiles/bullet.dir/btBvhTriangleMeshShape.s
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1663,19 +1663,16 @@ _ZZNK22btBvhTriangleMeshShape19processAllTrianglesEP18btTriangleCallbackRK9btVec
ld.w $a3, $a3, 0
.LBB22_6: # %.split38.us
mul.w $a2, $a2, $a3
fldx.d $fa0, $a1, $a2
fld.s $fa1, $a0, 8
vldx $vr0, $a1, $a2
add.d $a1, $a1, $a2
fcvt.s.d $fa0, $fa0
fmul.s $fa0, $fa1, $fa0
fld.d $fa1, $a1, 8
fld.s $fa2, $a0, 12
fld.d $fa3, $a1, 16
fld.s $fa4, $a0, 16
ld.d $a2, $a0, 8
vfcvt.s.d $vr0, $vr0, $vr0
fld.d $fa1, $a1, 16
fld.s $fa2, $a0, 16
vinsgr2vr.d $vr3, $a2, 0
vfmul.s $vr0, $vr3, $vr0
fcvt.s.d $fa1, $fa1
fmul.s $fa1, $fa2, $fa1
fcvt.s.d $fa2, $fa3
fmul.s $fa2, $fa4, $fa2
b .LBB22_9
.LBB22_7: # %.split.us.split.preheader
ld.w $a4, $a3, 8
Expand Down Expand Up @@ -1714,20 +1711,19 @@ _ZZNK22btBvhTriangleMeshShape19processAllTrianglesEP18btTriangleCallbackRK9btVec
.LBB22_8: # %.split38.us
mul.w $a2, $a2, $a3
add.d $a3, $a1, $a2
fldx.s $fa0, $a1, $a2
fld.s $fa1, $a0, 8
fld.s $fa2, $a3, 4
fld.s $fa3, $a0, 12
fld.s $fa4, $a3, 8
fld.s $fa5, $a0, 16
fmul.s $fa0, $fa0, $fa1
fmul.s $fa1, $fa2, $fa3
fmul.s $fa2, $fa4, $fa5
ldx.d $a1, $a1, $a2
ld.d $a2, $a0, 8
fld.s $fa1, $a3, 8
fld.s $fa2, $a0, 16
vinsgr2vr.d $vr0, $a1, 0
vinsgr2vr.d $vr3, $a2, 0
vfmul.s $vr0, $vr0, $vr3
fmul.s $fa1, $fa1, $fa2
.LBB22_9: # %.split38.us
ld.d $a0, $fp, 16
fst.s $fa0, $fp, 24
fst.s $fa1, $fp, 28
fst.s $fa2, $fp, 32
vstelm.w $vr0, $fp, 24, 0
vstelm.w $vr0, $fp, 28, 1
fst.s $fa1, $fp, 32
ld.d $a1, $a0, 0
st.w $zero, $fp, 36
ld.d $a4, $a1, 16
Expand Down
62 changes: 26 additions & 36 deletions results/MultiSource/Benchmarks/DOE-ProxyApps-C/SimpleMOC/CMakeFiles/SimpleMOC.dir/tracks.s
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -682,9 +682,18 @@ free_tracks: # @free_tracks
.word 5 # 0x5
.word 6 # 0x6
.word 7 # 0x7
.LCPI6_1:
.word 0 # 0x0
.word 1 # 0x1
.word 4 # 0x4
.word 5 # 0x5
.word 2 # 0x2
.word 3 # 0x3
.word 6 # 0x6
.word 7 # 0x7
.section .rodata.cst8,"aM",@progbits,8
.p2align 3, 0x0
.LCPI6_1:
.LCPI6_2:
.dword 0x400921fb54442d18 # double 3.1415926535897931
.text
.globl generate_polar_angles
Expand Down Expand Up @@ -718,58 +727,39 @@ generate_polar_angles: # @generate_polar_angles
xvldi $xr3, -928
lu12i.w $a2, 345154
ori $a2, $a2, 3352
pcalau12i $a3, %pc_hi20(.LCPI6_1)
xvld $xr4, $a3, %pc_lo12(.LCPI6_1)
lu32i.d $a2, -450053
lu52i.d $a2, $a2, 1024
xvreplgr2vr.d $xr4, $a2
xvreplgr2vr.d $xr5, $a2
move $a2, $a0
move $a3, $a1
.p2align 4, , 16
.LBB6_4: # %vector.body
# =>This Inner Loop Header: Depth=1
xvpermi.q $xr5, $xr1, 1
vext2xv.du.wu $xr5, $xr5
xvffint.d.lu $xr5, $xr5
vext2xv.du.wu $xr6, $xr1
xvpermi.q $xr6, $xr1, 1
vext2xv.du.wu $xr6, $xr6
xvffint.d.lu $xr6, $xr6
xvfadd.d $xr5, $xr5, $xr3
vext2xv.du.wu $xr7, $xr1
xvffint.d.lu $xr7, $xr7
xvfadd.d $xr7, $xr7, $xr3
xvfadd.d $xr6, $xr6, $xr3
xvfmul.d $xr6, $xr6, $xr4
xvfmul.d $xr5, $xr5, $xr4
xvfdiv.d $xr5, $xr5, $xr2
xvfmul.d $xr6, $xr6, $xr5
xvfmul.d $xr7, $xr7, $xr5
xvfdiv.d $xr7, $xr7, $xr2
xvfdiv.d $xr6, $xr6, $xr2
xvpickve.d $xr7, $xr6, 1
fcvt.s.d $fa7, $fa7
xvpickve.d $xr8, $xr6, 0
fcvt.s.d $ft0, $ft0
vextrins.w $vr8, $vr7, 16
xvpickve.d $xr7, $xr6, 2
fcvt.s.d $fa7, $fa7
vextrins.w $vr8, $vr7, 32
xvpickve.d $xr6, $xr6, 3
fcvt.s.d $fa6, $fa6
vextrins.w $vr8, $vr6, 48
xvpickve.d $xr6, $xr5, 1
fcvt.s.d $fa6, $fa6
xvpickve.d $xr7, $xr5, 0
fcvt.s.d $fa7, $fa7
vextrins.w $vr7, $vr6, 16
xvpickve.d $xr6, $xr5, 2
fcvt.s.d $fa6, $fa6
vextrins.w $vr7, $vr6, 32
xvpickve.d $xr5, $xr5, 3
fcvt.s.d $fa5, $fa5
vextrins.w $vr7, $vr5, 48
xvpermi.q $xr8, $xr7, 2
xvst $xr8, $a2, 0
xvfcvt.s.d $xr6, $xr6, $xr7
xvperm.w $xr6, $xr6, $xr4
xvst $xr6, $a2, 0
xvaddi.wu $xr1, $xr1, 8
addi.d $a3, $a3, -8
addi.d $a2, $a2, 32
bnez $a3, .LBB6_4
# %bb.5: # %middle.block
beq $a1, $fp, .LBB6_8
.LBB6_6: # %scalar.ph.preheader
pcalau12i $a2, %pc_hi20(.LCPI6_1)
fld.d $fa1, $a2, %pc_lo12(.LCPI6_1)
pcalau12i $a2, %pc_hi20(.LCPI6_2)
fld.d $fa1, $a2, %pc_lo12(.LCPI6_2)
alsl.d $a2, $a1, $a0, 2
sub.d $a3, $fp, $a1
vldi $vr2, -928
Expand Down
36 changes: 9 additions & 27 deletions results/MultiSource/Benchmarks/MiBench/consumer-jpeg/CMakeFiles/consumer-jpeg.dir/jcdctmgr.s
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -812,33 +812,24 @@ start_pass_fdctmgr: # @start_pass_fdctmgr
ffint.d.w $fa5, $fa1
vextrins.d $vr5, $vr4, 16
vreplvei.d $vr1, $vr0, 0
vfmul.d $vr4, $vr1, $vr5
xvori.b $xr5, $xr0, 0
fst.d $fs0, $sp, 184
fst.d $fs1, $sp, 176
fst.d $fs2, $sp, 168
fst.d $fa3, $sp, 160
xvld $xr3, $sp, 160
vfmul.d $vr4, $vr1, $vr5
xvori.b $xr5, $xr0, 0
xvinsve0.d $xr5, $xr2, 1
xvpermi.q $xr5, $xr4, 2
ld.hu $a3, $a2, 0
xvfmul.d $xr2, $xr5, $xr3
xvfmul.d $xr2, $xr2, $xr6
xvfrecip.d $xr2, $xr2
xvpickve.d $xr3, $xr2, 1
fcvt.s.d $fa3, $fa3
xvpickve.d $xr4, $xr2, 0
fcvt.s.d $fa4, $fa4
vextrins.w $vr4, $vr3, 16
xvpickve.d $xr3, $xr2, 2
ld.hu $a3, $a2, 0
fcvt.s.d $fa3, $fa3
vextrins.w $vr4, $vr3, 32
xvpickve.d $xr2, $xr2, 3
movgr2fr.w $fa3, $a3
ld.hu $a3, $a2, 2
fcvt.s.d $fa2, $fa2
vextrins.w $vr4, $vr2, 48
vst $vr4, $a0, -16
xvpermi.q $xr4, $xr2, 1
vfcvt.s.d $vr2, $vr4, $vr2
vst $vr2, $a0, -16
movgr2fr.w $fa2, $a3
ld.w $a3, $a2, 4
ffint.d.w $fa3, $fa3
Expand All @@ -865,18 +856,9 @@ start_pass_fdctmgr: # @start_pass_fdctmgr
xvfmul.d $xr0, $xr0, $xr3
xvfmul.d $xr0, $xr0, $xr6
xvfrecip.d $xr0, $xr0
xvpickve.d $xr1, $xr0, 1
fcvt.s.d $fa1, $fa1
xvpickve.d $xr2, $xr0, 0
fcvt.s.d $fa2, $fa2
vextrins.w $vr2, $vr1, 16
xvpickve.d $xr1, $xr0, 2
fcvt.s.d $fa1, $fa1
vextrins.w $vr2, $vr1, 32
xvpickve.d $xr0, $xr0, 3
fcvt.s.d $fa0, $fa0
vextrins.w $vr2, $vr0, 48
vst $vr2, $a0, 0
xvpermi.q $xr1, $xr0, 1
vfcvt.s.d $vr0, $vr1, $vr0
vst $vr0, $a0, 0
addi.d $a1, $a1, 8
addi.d $a2, $a2, 16
addi.d $a0, $a0, 32
Expand Down
Loading