[llvm] [llvm-calc-occupancy] Add an AMDGPU occupancy calculator tool (PR #208727)

Thu Jul 30 06:25:06 PDT 2026

llvm-ci wrote:

LLVM Buildbot has detected a new failure on builder `openmp-offload-amdgpu-runtime-2` running on `rocm-worker-hw-02` while building `llvm` at step 3 "annotate".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/10/builds/32926

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

```
Step 3 (annotate) failure: 'python ../llvm.src/offload/ci/openmp-offload-amdgpu-libc-runtime.py ...' (failure)
...
-- Performing Test COMPILER_RT_HAS_amdgpu_FLOAT16
-- Performing Test COMPILER_RT_HAS_amdgpu_FLOAT16 - Success
-- Builtin supported architectures: amdgpu
-- Performing additional configure checks with target flags: -nostdinc++ -Wno-c2y-extensions -fPIC -fno-builtin -fvisibility=hidden -fomit-frame-pointer -ffreestanding -nogpulib -flto SHELL:-Xclang -mcode-object-version=none
-- Performing Test COMPILER_RT_HAS_amdgpu_BFLOAT16
-- Performing Test COMPILER_RT_HAS_amdgpu_BFLOAT16 - Failed
-- Configuring done
-- Generating done
-- Build files have been written to: /home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/runtimes/builtins-amdgpu-amd-amdhsa-bins
[ 99%/36.656s :: 25->2->5074 (of 5100)] Building CXX object tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o
FAILED: tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o 
ccache /usr/bin/c++ -DLLVM_BUILD_STATIC -D_DEBUG -D_GLIBCXX_ASSERTIONS -D_GLIBCXX_USE_CXX11_ABI=1 -D_GNU_SOURCE -D_LIBCPP_HARDENING_MODE=_LIBCPP_HARDENING_MODE_EXTENSIVE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/tools/llvm-calc-occupancy -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/include -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/include -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/lib/Target/AMDGPU -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/lib/Target/AMDGPU -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wno-missing-field-initializers -pedantic -Wno-long-long -Wimplicit-fallthrough -Wno-uninitialized -Wno-nonnull -Wno-class-memaccess -Wno-redundant-move -Wno-pessimizing-move -Wno-array-bounds -Wno-stringop-overread -Wno-noexcept-type -Wdelete-non-virtual-dtor -Wsuggest-override -Wno-comment -Wno-misleading-indentation -Wctad-maybe-unsupported -fdiagnostics-color -ffunction-sections -fdata-sections -O3 -DNDEBUG -UNDEBUG -fno-exceptions -funwind-tables -fno-rtti -std=c++17 -MD -MT tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o -MF tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o.d -o tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o -c /home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy/llvm-calc-occupancy.cpp
/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy/llvm-calc-occupancy.cpp: In function ‘int main(int, char**)’:
/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy/llvm-calc-occupancy.cpp:229: error: ‘getAddressableNumSGPRs’ is not a member of ‘llvm::AMDGPU::IsaInfo’; did you mean ‘getAddressableNumVGPRs’?
  229 |   unsigned AddrSGPRs = AMDGPU::IsaInfo::getAddressableNumSGPRs(STI);
      | 
ninja: build stopped: subcommand failed.
Running: ccache -sv
Summary:
  Cache directory:            /home/botworker/.cache/ccache
  Primary config:             /home/botworker/.config/ccache/ccache.conf
  Secondary config:           /etc/ccache.conf
  Stats updated:              Thu Jul 30 13:24:49 2026
  Hits:                        3891 / 3897 (99.85 %)
    Direct:                    3348 / 3900 (85.85 %)
    Preprocessed:               543 /  550 (98.73 %)
  Misses:                         6
    Direct:                     552
    Preprocessed:                 7
  Uncacheable:                    3
Primary storage:
  Hits:                        7295 / 7796 (93.57 %)
  Misses:                       501
  Cache size (GB):             4.44 / 5.00 (88.70 %)
  Files:                      61489
Uncacheable:
  Compilation failed:             1
  Unsupported code directive:     2
ninja -C llvm.build -j64 exited with return code 1.
@@@STEP_FAILURE@@@
Traceback (most recent call last):
  File "/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/../llvm.src/offload/ci/openmp-offload-amdgpu-libc-runtime.py", line 37, in <module>
    w.run_ninja(builddir=llvmbuilddir, ccache_stats=True)
  File "/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/.ci/buildbot/worker.py", line 269, in run_ninja
    run_command(cmd, **kwargs)
  File "/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/.ci/buildbot/worker.py", line 132, in run_command
    subprocess.check_call(cmd, shell=shell, env=env, **kwargs)
  File "/usr/lib/python3.10/subprocess.py", line 369, in check_call
    raise CalledProcessError(retcode, cmd)
Step 7 (compile llvm) failure: compile llvm (failure)
...
-- Performing Test COMPILER_RT_HAS_amdgpu_FLOAT16
-- Performing Test COMPILER_RT_HAS_amdgpu_FLOAT16 - Success
-- Builtin supported architectures: amdgpu
-- Performing additional configure checks with target flags: -nostdinc++ -Wno-c2y-extensions -fPIC -fno-builtin -fvisibility=hidden -fomit-frame-pointer -ffreestanding -nogpulib -flto SHELL:-Xclang -mcode-object-version=none
-- Performing Test COMPILER_RT_HAS_amdgpu_BFLOAT16
-- Performing Test COMPILER_RT_HAS_amdgpu_BFLOAT16 - Failed
-- Configuring done
-- Generating done
-- Build files have been written to: /home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/runtimes/builtins-amdgpu-amd-amdhsa-bins
[ 99%/36.656s :: 25->2->5074 (of 5100)] Building CXX object tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o
FAILED: tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o 
ccache /usr/bin/c++ -DLLVM_BUILD_STATIC -D_DEBUG -D_GLIBCXX_ASSERTIONS -D_GLIBCXX_USE_CXX11_ABI=1 -D_GNU_SOURCE -D_LIBCPP_HARDENING_MODE=_LIBCPP_HARDENING_MODE_EXTENSIVE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/tools/llvm-calc-occupancy -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/include -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/include -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/lib/Target/AMDGPU -I/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/llvm.build/lib/Target/AMDGPU -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wno-missing-field-initializers -pedantic -Wno-long-long -Wimplicit-fallthrough -Wno-uninitialized -Wno-nonnull -Wno-class-memaccess -Wno-redundant-move -Wno-pessimizing-move -Wno-array-bounds -Wno-stringop-overread -Wno-noexcept-type -Wdelete-non-virtual-dtor -Wsuggest-override -Wno-comment -Wno-misleading-indentation -Wctad-maybe-unsupported -fdiagnostics-color -ffunction-sections -fdata-sections -O3 -DNDEBUG -UNDEBUG -fno-exceptions -funwind-tables -fno-rtti -std=c++17 -MD -MT tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o -MF tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o.d -o tools/llvm-calc-occupancy/CMakeFiles/llvm-calc-occupancy.dir/llvm-calc-occupancy.cpp.o -c /home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy/llvm-calc-occupancy.cpp
/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy/llvm-calc-occupancy.cpp: In function ‘int main(int, char**)’:
/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/llvm/tools/llvm-calc-occupancy/llvm-calc-occupancy.cpp:229: error: ‘getAddressableNumSGPRs’ is not a member of ‘llvm::AMDGPU::IsaInfo’; did you mean ‘getAddressableNumVGPRs’?
  229 |   unsigned AddrSGPRs = AMDGPU::IsaInfo::getAddressableNumSGPRs(STI);
      | 
ninja: build stopped: subcommand failed.
Running: ccache -sv
Summary:
  Cache directory:            /home/botworker/.cache/ccache
  Primary config:             /home/botworker/.config/ccache/ccache.conf
  Secondary config:           /etc/ccache.conf
  Stats updated:              Thu Jul 30 13:24:49 2026
  Hits:                        3891 / 3897 (99.85 %)
    Direct:                    3348 / 3900 (85.85 %)
    Preprocessed:               543 /  550 (98.73 %)
  Misses:                         6
    Direct:                     552
    Preprocessed:                 7
  Uncacheable:                    3
Primary storage:
  Hits:                        7295 / 7796 (93.57 %)
  Misses:                       501
  Cache size (GB):             4.44 / 5.00 (88.70 %)
  Files:                      61489
Uncacheable:
  Compilation failed:             1
  Unsupported code directive:     2
ninja -C llvm.build -j64 exited with return code 1.
Traceback (most recent call last):
  File "/home/botworker/builds/openmp-offload-amdgpu-runtime-2/build/../llvm.src/offload/ci/openmp-offload-amdgpu-libc-runtime.py", line 37, in <module>
    w.run_ninja(builddir=llvmbuilddir, ccache_stats=True)
  File "/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/.ci/buildbot/worker.py", line 269, in run_ninja
    run_command(cmd, **kwargs)
  File "/home/botworker/builds/openmp-offload-amdgpu-runtime-2/llvm.src/.ci/buildbot/worker.py", line 132, in run_command
    subprocess.check_call(cmd, shell=shell, env=env, **kwargs)
  File "/usr/lib/python3.10/subprocess.py", line 369, in check_call
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['ninja', '-C', 'llvm.build', '-j64']' returned non-zero exit status 1.

```

</details>

https://github.com/llvm/llvm-project/pull/208727