diff --git a/vllm/model_executor/layers/batch_invariant.py b/vllm/model_executor/layers/batch_invariant.py index 3831f7aa96..bcdd305003 100644 --- a/vllm/model_executor/layers/batch_invariant.py +++ b/vllm/model_executor/layers/batch_invariant.py @@ -930,22 +930,17 @@ def enable_batch_invariant_mode(): _batch_invariant_MODE = True _batch_invariant_LIB = torch.library.Library("aten", "IMPL") - if current_platform.is_device_capability_family( - 100 - ) or current_platform.is_device_capability_family(80): - # For PyTorch 2.9, B200 uses GEMV for bs=1 - # Requires https://github.com/pytorch/pytorch/pull/166735 + if current_platform.is_device_capability_family(80): + # SM80 (Ampere) cannot rely on cuBLASLt-only determinism; install the + # triton persistent matmul overrides for mm/addmm/matmul/linear. _batch_invariant_LIB.impl("aten::mm", mm_batch_invariant, "CUDA") _batch_invariant_LIB.impl("aten::addmm", addmm_batch_invariant, "CUDA") _batch_invariant_LIB.impl("aten::matmul", matmul_batch_invariant, "CUDA") _batch_invariant_LIB.impl("aten::linear", linear_batch_invariant, "CUDA") - - # Query the shared memory size and set block size - # accordingly to avoid triton OutOfResources - _fp16_block_size_n = 256 if get_max_shared_memory_bytes() > 106496 else 128 else: - # Only source of batch invariance for Hopper is split-k, can disable through - # cuBLAS workspace config + # Hopper (SM90) and Blackwell (SM100): the only source of batch + # variance is split-k, which we disable via the cuBLAS workspace + # config. _original_cublas_workspace_cfg = os.environ.get("CUBLAS_WORKSPACE_CONFIG", None) _original_cublaslt_workspace_size = os.environ.get( "CUBLASLT_WORKSPACE_SIZE", None @@ -953,6 +948,11 @@ def enable_batch_invariant_mode(): os.environ["CUBLAS_WORKSPACE_CONFIG"] = ":16:8" os.environ["CUBLASLT_WORKSPACE_SIZE"] = "1" + # Triton bmm/persistent-matmul kernels read this for the FP16 N-tile size; + # set unconditionally because bmm is overridden on all CUDA platforms. + if current_platform.is_cuda(): + _fp16_block_size_n = 256 if get_max_shared_memory_bytes() > 106496 else 128 + _batch_invariant_LIB.impl( "aten::_log_softmax", _log_softmax_batch_invariant, "CUDA" )