[llvm] [AMDGPU] Improve codegen for GFX10+ DPP reductions and scans (PR #107108)

[LLVM Continuous Integration via llvm-commits]

llvm-ci wrote:

LLVM Buildbot has detected a new failure on builder `ml-opt-devrel-x86-64` running on `ml-opt-devrel-x86-64-b2` while building `llvm` at step 6 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/175/builds/4559

<details>
<summary>Here is the relevant piece of the build log for the reference</summary>

```
Step 6 (test-build-unified-tree-check-all) failure: test (failure)
******************** TEST 'LLVM :: CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll' FAILED ********************
Exit Code: 1

Command Output (stderr):
--
RUN: at line 2: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX7LESS,GFX7LESS_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX7LESS,GFX7LESS_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
RUN: at line 3: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX8,GFX8_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX8,GFX8_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
RUN: at line 4: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx900 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX9,GFX9_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX9,GFX9_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx900 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
RUN: at line 5: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1064,GFX1064_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1064,GFX1064_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
RUN: at line 6: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1032,GFX1032_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1032,GFX1032_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
RUN: at line 7: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1164,GFX1164_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1164,GFX1164_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
RUN: at line 8: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1132,GFX1132_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1132,GFX1132_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
RUN: at line 9: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1200 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1264,GFX1264_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1264,GFX1264_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1200 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
RUN: at line 10: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1200 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1232,GFX1232_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1232,GFX1232_ITERATIVE /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1200 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=Iterative -verify-machineinstrs
RUN: at line 11: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX7LESS,GFX7LESS_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX7LESS,GFX7LESS_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs
RUN: at line 12: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX8,GFX8_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX8,GFX8_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs
RUN: at line 13: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx900 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX9,GFX9_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX9,GFX9_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx900 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs
RUN: at line 14: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1064,GFX1064_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1064,GFX1064_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs
RUN: at line 15: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1032,GFX1032_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1010 -mattr=+wavefrontsize32 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1032,GFX1032_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
RUN: at line 16: /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs < /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll | /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1164,GFX1164_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/FileCheck -enable-var-scope -check-prefixes=GFX1164,GFX1164_DPP /b/ml-opt-devrel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
+ /b/ml-opt-devrel-x86-64-b1/build/bin/llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=+wavefrontsize64 -mattr=-flat-for-global -amdgpu-atomic-optimizer-strategy=DPP -verify-machineinstrs
...

```

</details>

https://github.com/llvm/llvm-project/pull/107108