vllm/kernels at 3a557ea3da2c248983a8e7bf2e3abc7218797807 - vllm

mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-18 11:07:11 +08:00

History

c0de128 719ccfd773 Add unit tests for get_fp8_min_max() helper function

Add test_fp8_min_max_helper.py with mocked unit tests that verify:
- Standard FP8 dtype uses PyTorch's finfo values
- fnuz dtype on fnuz platform (MI300) returns 224.0, not 240.0
- Standard dtype on fnuz platform uses finfo values
- fnuz dtype on non-fnuz platform uses finfo values

These tests use mocking to verify the logic without requiring
actual ROCm hardware.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Signed-off-by: c0de128 <kevin.mckay@outlook.com>

2025-12-24 13:20:26 -06:00

attention

[Misc] Remove unused custom ops copy_blocks and copy_blocks_mla (#30967 )

2025-12-23 18:22:35 -08:00

core

[Kernel] Enable fused_qknorm_rope_kernel supports partial rope (#30821 )

2025-12-21 18:39:22 -08:00

mamba

Add SpecDec support to selective_state_update (#29488 )

2025-12-08 16:45:18 -05:00

moe

[MoE Refactor][9/N] Use modular kernel for unquantized Triton MoE (#31052 )