[clang][Driver][HIP] Change OffloadingActionBuilder to respect the --no-gpu-bundle-output flag

[MixedMatched]

Currently, the command clang -c -emit-llvm --no-gpu-bundle-output --offload-arch=gfx900,gfx1030 -O3 -x hip square.hip will lead to a bundled output:

❯ ../bin/clang -c -emit-llvm --no-gpu-bundle-output --offload-arch=gfx900,gfx1030 -O3 -x hip square.hip
❯ ls
square.hip
square.bc

This doesn't match my expectation of the behavior of --no-gpu-bundle-output, so this adds a check into OffloadingActionBuilder for the flag when replacing the host compile action for a bundling action.

[github-actions]

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

[llvmbot]

@llvm/pr-subscribers-clang-driver

@llvm/pr-subscribers-clang

Author: alessandra simmons (MixedMatched)

Changes

Currently, the command clang -c -emit-llvm --no-gpu-bundle-output --offload-arch=gfx900,gfx1030 -O3 -x hip square.hip will lead to a bundled output:

❯ ../bin/clang -c -emit-llvm --no-gpu-bundle-output --offload-arch=gfx900,gfx1030 -O3 -x hip square.hip
❯ ls
square.hip
square.bc

This doesn't match my expectation of the behavior of --no-gpu-bundle-output, so this adds a check into OffloadingActionBuilder for the flag when replacing the host compile action for a bundling action.

---

Full diff: https://github.com/llvm/llvm-project/pull/163834.diff

2 Files Affected:

• (modified) clang/lib/Driver/Driver.cpp (+11-5)
• (added) clang/test/Driver/no-gpu-bundle-respected.hip (+18)

diff --git a/clang/lib/Driver/Driver.cpp b/clang/lib/Driver/Driver.cpp
index 40ea513e85427..02cfc4314bbc4 100644
--- a/clang/lib/Driver/Driver.cpp
+++ b/clang/lib/Driver/Driver.cpp
@@ -3844,6 +3844,9 @@ class OffloadingActionBuilder final {
   /// Flag set to true if all valid builders allow file bundling/unbundling.
   bool CanUseBundler;
 
+  /// Flag set to false if an argument turns off bundling.
+  bool ShouldUseBundler;
+
 public:
   OffloadingActionBuilder(Compilation &C, DerivedArgList &Args,
                           const Driver::InputList &Inputs)
@@ -3878,6 +3881,9 @@ class OffloadingActionBuilder final {
     }
     CanUseBundler =
         ValidBuilders && ValidBuilders == ValidBuildersSupportingBundling;
+
+    ShouldUseBundler = Args.hasFlag(options::OPT_gpu_bundle_output,
+                                    options::OPT_no_gpu_bundle_output, true);
   }
 
   ~OffloadingActionBuilder() {
@@ -4029,11 +4035,11 @@ class OffloadingActionBuilder final {
       SB->appendTopLevelActions(OffloadAL);
     }
 
-    // If we can use the bundler, replace the host action by the bundling one in
-    // the resulting list. Otherwise, just append the device actions. For
-    // device only compilation, HostAction is a null pointer, therefore only do
-    // this when HostAction is not a null pointer.
-    if (CanUseBundler && HostAction &&
+    // If we can and should use the bundler, replace the host action by the
+    // bundling one in the resulting list. Otherwise, just append the device
+    // actions. For device only compilation, HostAction is a null pointer,
+    // therefore only do this when HostAction is not a null pointer.
+    if (CanUseBundler && ShouldUseBundler && HostAction &&
         HostAction->getType() != types::TY_Nothing && !OffloadAL.empty()) {
       // Add the host action to the list in order to create the bundling action.
       OffloadAL.push_back(HostAction);
diff --git a/clang/test/Driver/no-gpu-bundle-respected.hip b/clang/test/Driver/no-gpu-bundle-respected.hip
new file mode 100644
index 0000000000000..1587551f0322d
--- /dev/null
+++ b/clang/test/Driver/no-gpu-bundle-respected.hip
@@ -0,0 +1,18 @@
+// RUN: %clang -ccc-print-phases -c -emit-llvm \
+// RUN:   --offload-arch=gfx900,gfx1030 -O3 -x hip %s \
+// RUN:   2>&1 | FileCheck %s --check-prefix=OFFLOAD
+
+// RUN: %clang -ccc-print-phases -c -emit-llvm \
+// RUN:   --gpu-bundle-output --offload-arch=gfx900,gfx1030 -O3 -x hip %s \
+// RUN:   2>&1 | FileCheck %s --check-prefix=OFFLOAD
+
+// RUN: %clang -ccc-print-phases -c -emit-llvm \
+// RUN:   --no-gpu-bundle-output --offload-arch=gfx900,gfx1030 -O3 -x hip %s \
+// RUN:   2>&1 | FileCheck %s --check-prefix=OFFLOAD2
+
+// OFFLOAD: clang-offload-bundler
+// OFFLOAD2-NOT: clang-offload-bundler
+
+int square(int num) {
+    return num * num;
+}

[MixedMatched]

@lamb-j @yxsamliu

[github-actions]

⚠️ We detected that you are using a GitHub private e-mail address to contribute to the repo.
Please turn off Keep my email addresses private setting in your account.
See LLVM Developer Policy and LLVM Discourse for more information.

[github-actions]

⚠️ We detected that you are using a GitHub private e-mail address to contribute to the repo.
Please turn off Keep my email addresses private setting in your account.
See LLVM Developer Policy and LLVM Discourse for more information.

[lamb-j]

lgtm, seems reasonable to not unbundle when --no-gpu-bundle-output is specified

Can you wait a day or two before landing to see if any other reviewer has feedback? Thanks!

[yxsamliu]

use a meaningful check line, e.g. BUNDLE, NOBUNDLE. also check the action entries for the NOBUNDLE case.

[MixedMatched]

use a meaningful check line

ah! so sorry, I should have caught those

also check the action entries for the NOBUNDLE case.

I tried looking through the -ccc-print-phases and the verbose outputs and I couldn't find a difference besides the order (which I thought might be flaky?) and the absence of the clang-offload-bundler command (which I'm already testing for). is there another way to test that which I'm missing?

[github-actions]

@MixedMatched Congratulations on having your first Pull Request (PR) merged into the LLVM Project!

Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR.

Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues.

How to do this, and the rest of the post-merge process, is covered in detail here.

If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again.

If you don't get any reports, no action is required from you. Your changes are working as expected, well done!