Ilya Markov 2d7779f888

[Perf] SM100 FP8 GEMM Optimizations after cutlass_profiler (#20071) ...

Signed-off-by: ilmarkov <imarkov@redhat.com>
Co-authored-by: ilmarkov <imarkov@redhat.com>

2025-06-26 20:50:09 -07:00

..

attention

[MISC] Remove unused variableds in C++ (#19609)

2025-06-15 20:05:28 -07:00

core

[Kernel] fp4 marlin kernel (#17687)

2025-05-10 19:58:49 -07:00

cpu

[CPU] Fix torch version in x86 CPU backend (#19258)

2025-06-26 03:34:47 -07:00

cutlass_extensions

[Misc] Add SPDX-FileCopyrightText (#19100)

2025-06-03 11:20:17 -07:00

mamba

feat(rocm-support): support mamba2 on rocm (#18565)

2025-05-27 00:07:53 -07:00

moe

[Perf] Optimize moe_align_block_size CUDA kernel (#19572)

2025-06-17 11:49:26 -07:00

prepare_inputs

[MISC] Remove unused variableds in C++ (#19609)

2025-06-15 20:05:28 -07:00

quantization

[Perf] SM100 FP8 GEMM Optimizations after cutlass_profiler (#20071)

2025-06-26 20:50:09 -07:00

rocm

[MISC] Remove unused variableds in C++ (#19609)

2025-06-15 20:05:28 -07:00

sparse/cutlass

[CI] change spell checker from codespell to typos (#18711)

2025-06-11 19:57:10 -07:00

activation_kernels.cu

Modularize fused experts and integrate PPLX kernels (#15956)

2025-05-14 13:11:54 -07:00

cache_kernels.cu

Allocate kv_cache with stride order (#16605)

2025-04-25 22:03:31 -07:00

cache.h

[Attention] MLA with chunked prefill (#12639)

2025-02-21 15:30:12 -08:00

cuda_compat.h

[Kernel][ROCm][AMD] enable fused topk_softmax kernel for moe layer (#4927)

2024-06-02 14:13:26 -07:00

cuda_utils_kernels.cu

[NVIDIA] Support nvfp4 quantization (#12784)

2025-02-12 19:51:51 -08:00

cuda_utils.h

[Attention] MLA with chunked prefill (#12639)

2025-02-21 15:30:12 -08:00

cuda_view.cu

[V1] Fully Transparent Implementation of CPU Offloading (#15354)

2025-03-31 20:22:34 +08:00

cumem_allocator.cpp

[core] improve error handling when wake up from sleep mode (#12981)

2025-02-10 09:38:57 +08:00

custom_all_reduce_test.cu

[Distributed] Add custom allreduce support for ROCM (#14125)

2025-03-31 22:49:12 -07:00

custom_all_reduce.cu

[Distributed] Add custom allreduce support for ROCM (#14125)

2025-03-31 22:49:12 -07:00

custom_all_reduce.cuh

fix: spelling (#16466)

2025-04-11 23:24:22 -07:00

dispatch_utils.h

Modularize fused experts and integrate PPLX kernels (#15956)

2025-05-14 13:11:54 -07:00

layernorm_kernels.cu

[Kernel] Use fused rmsnorm for some models like qwen3 series (#17735)

2025-05-06 23:10:02 -07:00

layernorm_quant_kernels.cu

dynamic distpatch of fp8 kernels (#14245)

2025-03-11 10:54:56 -04:00

ops.h

[Kernel] Integrate CUTLASS MoE kernel with PPLX (#18762)

2025-06-06 18:26:11 -07:00

permute_cols.cu

[Kernel] (2/N) Machete - Integrate into CompressedTensorsWNA16 and GPTQMarlin (#7701)

2024-09-23 13:46:26 -04:00

pos_encoding_kernels.cu

[Kernel] Have rotary embeddings support tensors (#18046)

2025-05-14 15:43:55 -07:00

sampler.cu

[KERNEL] Sampler. CUDA kernel for applying repetition penalty (#18437)

2025-06-03 21:13:01 -07:00

torch_bindings.cpp

[Kernel] Integrate CUTLASS MoE kernel with PPLX (#18762)

2025-06-06 18:26:11 -07:00

type_convert.cuh

[torch.compile] Fuse RMSNorm with quant (#9138)

2024-11-08 21:20:08 +00:00