Test suite - Add QA pass/fail tests for F32 bit depth

[r-abishek]

The PR adds internal QA testing support for F32 bit depth for next set of functions.

• Add range checks rpp_pixel_check_0to1 to ensure F32 bit depth QA test pass.
• Add necessary bin files for reference output.

[Copilot]

Pull Request Overview

This PR introduces additional QA tests for F32 bit depth processing along with range‐check validations for tensor functions. Key changes include:

• Adding new validations for data type and layout consistency before processing in several host and GPU kernels.
• Inserting boundary checks for F32 values (using rpp_pixel_check_0to1) across multiple tensor processing functions.
• Updating API definitions and simd load/store helper functions to support the new checks.

Reviewed Changes

Copilot reviewed 23 out of 23 changed files in this pull request and generated 1 comment.

File	Description
src/modules/tensor/rppt_tensor_geometric_augmentations.cpp	Added validations for matching data types and valid layouts prior to crop/patch operations.
src/include/common/cpu/rpp_cpu_simd_load_store.hpp	Updated rpp_pixel_check_0to1 functions to modify vectors in place rather than returning clamped values.
src/modules/tensor/cpu/kernel/contrast.cpp	Inserted boundary checks for F32 values to ensure pixel intensity remains in the expected [0, 255] range.

Comments suppressed due to low confidence (2)

src/include/common/cpu/rpp_cpu_simd_load_store.hpp:275

• The function signature of rpp_pixel_check_0to1 has been changed to modify the vector in place rather than returning a value; please ensure that all call sites have been updated accordingly.

inline __m256 rpp_pixel_check_0to1_avx(__m256 p)

src/modules/tensor/cpu/kernel/contrast.cpp:310

• [nitpick] The addition of boundary checks for F32 pixels ensures values remain within acceptable limits; verify that similar checks in other kernels are consistent and do not introduce performance regressions.

rpp_pixel_check_0to1(p, 3);

[Copilot]

Pull Request Overview

This pull request adds F32 bit depth QA test support by introducing range validations and error checking both at the API level and inside the host/GPU kernels. Key changes include adding explicit error returns for mismatched dataType and unsupported layouts, inserting boundary check calls (rpp_pixel_check_0to1) after pixel processing, and updating SIMD load/store functions to work with the revised boundary enforcement.

Reviewed Changes

Copilot reviewed 23 out of 23 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
src/modules/tensor/rppt_tensor_geometric_augmentations.cpp	Added validation checks on data types and layouts in crop functions.
src/modules/tensor/rppt_tensor_color_augmentations.cpp	Introduced similar dataType/layout constraints and boundary checks.
src/modules/tensor/cpu/kernel/threshold.cpp	Replaced legacy rpp_pixel_check_0to1_avx calls with the updated rpp_pixel_check_0to1 function for F32.
src/include/common/cpu/rpp_cpu_simd_load_store.hpp	Updated the rpp_pixel_check_0to1 functions’ signatures to operate on vector arrays and return void.
...	Several kernel files (phase.cpp, brightness.cpp, etc.) have analogous boundary check insertions and API updates.

Comments suppressed due to low confidence (1)

src/modules/tensor/cpu/kernel/threshold.cpp:42

• The legacy rpp_pixel_check_0to1_avx call has been replaced by a new boundary check call; please verify that the new rpp_pixel_check_0to1 function correctly enforces the [0,1] range across all SIMD vector lengths.

p[0] = _mm256_blendv_ps(avx_p0, avx_p1, _mm256_and_ps(_mm256_and_ps(pChannelCheck[0], pChannelCheck[1]), pChannelCheck[2]));

[rrawther]

Added few comments

[codecov]

Codecov Report

Attention: Patch coverage is 80.12422% with 32 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/include/common/cpu/rpp_cpu_simd_load_store.hpp	11.11%	32 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##           develop     #549   +/-   ##
========================================
  Coverage    85.78%   85.78%           
========================================
  Files          185      185           
  Lines        78985    79060   +75     
========================================
+ Hits         67751    67819   +68     
- Misses       11234    11241    +7     

Files with missing lines	Coverage Δ
src/modules/tensor/cpu/kernel/blend.cpp	100.00% <100.00%> (ø)
src/modules/tensor/cpu/kernel/brightness.cpp	94.31% <100.00%> (+0.05%)	⬆️
src/modules/tensor/cpu/kernel/color_cast.cpp	95.51% <100.00%> (+0.04%)	⬆️
src/modules/tensor/cpu/kernel/color_twist.cpp	89.18% <100.00%> (+0.07%)	⬆️
src/modules/tensor/cpu/kernel/contrast.cpp	100.00% <100.00%> (ø)
src/modules/tensor/cpu/kernel/exposure.cpp	100.00% <100.00%> (ø)
src/modules/tensor/cpu/kernel/magnitude.cpp	96.03% <100.00%> (+0.03%)	⬆️
src/modules/tensor/cpu/kernel/phase.cpp	96.95% <100.00%> (+0.02%)	⬆️
src/modules/tensor/cpu/kernel/threshold.cpp	95.62% <100.00%> (ø)
...modules/tensor/rppt_tensor_color_augmentations.cpp	98.36% <100.00%> (+0.05%)	⬆️
... and 2 more

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[rrawther]

As we have discussed, please add the pixel check only in the functions where it is needed.

[rrawther]

@r-abishek there was a comment which is not resolved

[r-abishek]

@rrawther All comments resolved now

[kiritigowda]

/AzurePipelines run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).