[CI Perf] Prune tests in tests/kernels/quantization/

[mgoin]

Purpose

pytest --collect-only -qq tests/kernels/quantization 

# main
tests/kernels/quantization/test_allspark_gemm.py: 32
tests/kernels/quantization/test_aqlm.py: 2
tests/kernels/quantization/test_awq.py: 2
tests/kernels/quantization/test_awq_triton.py: 360
tests/kernels/quantization/test_block_fp8.py: 324
tests/kernels/quantization/test_block_int8.py: 32
tests/kernels/quantization/test_cutlass_2of4_sparse.py: 267
tests/kernels/quantization/test_cutlass_scaled_mm.py: 689
tests/kernels/quantization/test_flashinfer_nvfp4_scaled_mm.py: 48
tests/kernels/quantization/test_fp8_quant.py: 757
tests/kernels/quantization/test_ggml.py: 1
tests/kernels/quantization/test_gguf.py: 160
tests/kernels/quantization/test_gptq.py: 2
tests/kernels/quantization/test_int8_kernel.py: 32
tests/kernels/quantization/test_int8_quant.py: 1346
tests/kernels/quantization/test_machete_mm.py: 316
tests/kernels/quantization/test_marlin_gemm.py: 6804
tests/kernels/quantization/test_nvfp4_quant.py: 18
tests/kernels/quantization/test_nvfp4_scaled_mm.py: 12
tests/kernels/quantization/test_per_token_group_quant.py: 30
tests/kernels/quantization/test_rocm_skinny_gemms.py: 3184
tests/kernels/quantization/test_triton_scaled_mm.py: 1153

# pr
tests/kernels/quantization/test_allspark_gemm.py: 32
tests/kernels/quantization/test_aqlm.py: 2
tests/kernels/quantization/test_awq.py: 2
tests/kernels/quantization/test_awq_triton.py: 360
tests/kernels/quantization/test_block_fp8.py: 324
tests/kernels/quantization/test_block_int8.py: 32
tests/kernels/quantization/test_cutlass_2of4_sparse.py: 267
tests/kernels/quantization/test_cutlass_scaled_mm.py: 689
tests/kernels/quantization/test_flashinfer_nvfp4_scaled_mm.py: 48
tests/kernels/quantization/test_fp8_quant.py: 109
tests/kernels/quantization/test_ggml.py: 1
tests/kernels/quantization/test_gguf.py: 160
tests/kernels/quantization/test_gptq.py: 2
tests/kernels/quantization/test_int8_kernel.py: 32
tests/kernels/quantization/test_int8_quant.py: 290
tests/kernels/quantization/test_machete_mm.py: 220
tests/kernels/quantization/test_marlin_gemm.py: 3404
tests/kernels/quantization/test_nvfp4_quant.py: 18
tests/kernels/quantization/test_nvfp4_scaled_mm.py: 12
tests/kernels/quantization/test_per_token_group_quant.py: 30
tests/kernels/quantization/test_rocm_skinny_gemms.py: 102
tests/kernels/quantization/test_triton_scaled_mm.py: 97

Test Plan

Test Result

(Optional) Documentation Update

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[gemini-code-assist]

Code Review

This pull request aims to improve CI performance by pruning the test suite for quantization kernels. While many of the changes, such as reducing the parameter space for tests and moving to curated test cases, are reasonable optimizations, there is a significant concern. The removal of torch.float16 from the data types being tested in test_fp8_quant.py, test_int8_quant.py, and test_triton_scaled_mm.py could mask potential regressions for this widely used data type. It is recommended to restore testing for torch.float16 to maintain robust test coverage.

[yewentao256]

LGTM, thanks for the work!