Fix warning when building kernel libraries

[larryliu0820]

Getting warnings like the following:

  /home/larryliu/executorch/kernels/portable/cpu/op_maximum.cpp: In lambda function:
  /home/larryliu/executorch/kernels/portable/cpu/op_maximum.cpp:52:9: note: the ABI for passing parameters with 32-byte alignment has changed in GCC 4.6
     52 |         [](const auto val_a, const auto val_b) {
        |         ^
  /home/larryliu/executorch/../executorch/runtime/core/exec_aten/util/scalar_type_util.h:919:7: note: in definition of macro 'ET_INTERNAL_SWITCH'
    919 |       __VA_ARGS__                                              \
        |       ^~~~~~~~~~~
  /home/larryliu/executorch/../executorch/runtime/core/exec_aten/util/scalar_type_util.h:931:3: note: in expansion of macro 'ET_INTERNAL_SWITCH_CASE'
    931 |   ET_INTERNAL_SWITCH_CASE(                                             \
        |   ^~~~~~~~~~~~~~~~~~~~~~~
  /home/larryliu/executorch/../executorch/runtime/core/exec_aten/util/scalar_type_util.h:957:3: note: in expansion of macro 'ET_INTERNAL_SWITCH_CASE_INT_TYPES'
    957 |   ET_INTERNAL_SWITCH_CASE_INT_TYPES(CTYPE_ALIAS, __VA_ARGS__) \
        |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  /home/larryliu/executorch/../executorch/runtime/core/exec_aten/util/scalar_type_util.h:1008:3: note: in expansion of macro 'ET_INTERNAL_SWITCH_CASE_REAL_TYPES'
   1008 |   ET_INTERNAL_SWITCH_CASE_REAL_TYPES(CTYPE_ALIAS, __VA_ARGS__)               \
        |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  /home/larryliu/executorch/../executorch/runtime/core/exec_aten/util/scalar_type_util.h:1136:7: note: in expansion of macro 'ET_INTERNAL_SWITCH_CASE_REAL_TYPES_AND'
   1136 |       ET_INTERNAL_SWITCH_CASE_REAL_TYPES_AND(          \
        |       ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  /home/larryliu/executorch/../executorch/runtime/core/exec_aten/util/scalar_type_util.h:1172:3: note: in expansion of macro 'ET_SWITCH_REAL_TYPES_AND'
   1172 |   ET_SWITCH_REAL_TYPES_AND(Bool, TYPE, CONTEXT, NAME, CTYPE_ALIAS, __VA_ARGS__)
        |   ^~~~~~~~~~~~~~~~~~~~~~~~
  /home/larryliu/executorch/kernels/portable/cpu/op_maximum.cpp:47:3: note: in expansion of macro 'ET_SWITCH_REALB_TYPES'
     47 |   ET_SWITCH_REALB_TYPES(compute_type, ctx, op_name, CTYPE_COMPUTE, [&]() {
        |   ^~~~~~~~~~~~~~~~~~~~~

Fixing them in this PR

Overview

---

This branch addresses GCC compiler warnings related to 32-byte alignment (ABI) when passing SIMD vector types to lambda functions in both portable and optimized kernel implementations.

Key Changes:

• CMakeLists.txt: Added -Wno-psabi compiler flag to suppress ABI warnings for GNU compiler: $<$<CXX_COMPILER_ID:GNU>:-Wno-psabi>
• Removed explicit template parameter <CTYPE> and changed lambda parameters from reference & to value (pass by value) from at::vec::map() and at::vec::map3() calls

// Before: 
at::vec::map<CTYPE>([alpha_val, b_val](Vec& x) { return x + Vec(alpha_val * b_val); }, ...);  

// After: 
at::vec::map([alpha_val, b_val](Vec x) { return x + Vec(alpha_val * b_val); }, ...);

• Changed lambda parameters from value to reference & for lambdas being passed to apply_bitensor_elementwise_fn() and similar functions.

// Before:
[val_alpha](const auto val_a, const auto val_b) {
  return val_a + val_alpha * val_b;
}

// After:
[val_alpha](const auto& val_a, const auto& val_b) {
  return val_a + val_alpha * val_b;
}

Problem Solved

---

Before: GCC compiler produced numerous warnings like:

note: the ABI for passing parameters with 32-byte alignment has changed in GCC 4.6

These warnings appeared when SIMD vector types (e.g., __m256 from AVX) were passed to lambda functions, cluttering build output and making it difficult to spot real issues.

After: Clean builds with no ABI warnings, while maintaining identical functionality and performance characteristics.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15405

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❗ 2 Active SEVs

There are 2 currently active SEVs. If your PR is affected, please view them below:

• Long queue for ROCM runners, also B200 and XPU queueing is observed
• B200 queuing was observed on Oct 29

⏳ No Failures, 7 Pending

As of commit c23e44f with merge base 72b1fa1 ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[meta-codesync]

@larryliu0820 has imported this pull request. If you are a Meta employee, you can view this in D85778660.