Jee Jee Li ecd1ea1363 [Kernel] Porting the TRTLLM minimax_allreduce_rms kernels (#37045) ...

Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>

2026-04-11 00:20:20 +08:00

..

attention

[Bugfix] Fix broken explicit unquantized kv cache dtype support (#38922)

2026-04-09 22:27:53 -07:00

core

…

cpu

feat(cpu): add CPU support for draft model speculative decoding (#32662)

2026-04-10 11:49:52 +08:00

cutlass_extensions

[BugFix] Fix OOB read in CUTLASS grouped GEMM with epilogue (#38571)

2026-04-09 23:52:52 -04:00

libtorch_stable

[Kernel] Add swapAB support for SM120 CUTLASS blockwise FP8 GEMM (#38325)

2026-04-03 15:49:59 +02:00

mamba/mamba_ssm

[Bugfix] Use null block (0) for padded block table entries (#35431)

2026-03-30 14:02:51 -07:00

moe

Revert "[Kernel] Add gpt-oss Router GEMM kernel (#37205)" (#38778)

2026-04-01 22:02:32 -07:00

quantization

[Bugfix] Fix broken explicit unquantized kv cache dtype support (#38922)

2026-04-09 22:27:53 -07:00

quickreduce

…

rocm

…

activation_kernels.cu

…

cache_kernels_fused.cu

…

cache_kernels.cu

[Bug] Fix compile error for swap_blocks_batch in CUDA 13 (#38915)

2026-04-03 16:56:38 -07:00

cache.h

[Perf] Batch KV cache swap copies via cuMemcpyBatchAsync (#38460)

2026-04-03 03:13:23 +00:00

concat_mla_q.cuh

…

cub_helpers.h

…

cuda_compat.h

…

cuda_utils_kernels.cu

…

cuda_utils.h

…

cuda_vec_utils.cuh

[4/n] Migrate FP4/W4A8 CUTLASS kernels to torch stable ABI (#37503)

2026-03-31 10:21:13 -07:00

cuda_view.cu

…

cumem_allocator_compat.h

…

cumem_allocator.cpp

…

custom_all_reduce_test.cu

…

custom_all_reduce.cu

…

custom_all_reduce.cuh

…

custom_quickreduce.cu

…

dispatch_utils.h

…

dsv3_fused_a_gemm.cu

…

fused_qknorm_rope_kernel.cu

…

launch_bounds_utils.h

…

layernorm_kernels.cu

[2/n] Migrate per_token_group_quant to torch stable ABI (#36058)

2026-03-25 10:15:13 -07:00

layernorm_quant_kernels.cu

[2/n] Migrate per_token_group_quant to torch stable ABI (#36058)

2026-03-25 10:15:13 -07:00

minimax_reduce_rms_kernel.cu

[Kernel] Porting the TRTLLM minimax_allreduce_rms kernels (#37045)

2026-04-11 00:20:20 +08:00

minimax_reduce_rms_kernel.h

[Kernel] Porting the TRTLLM minimax_allreduce_rms kernels (#37045)

2026-04-11 00:20:20 +08:00

ops.h

[Kernel] Porting the TRTLLM minimax_allreduce_rms kernels (#37045)

2026-04-11 00:20:20 +08:00

persistent_topk.cuh

[Perf][Kernel] Persistent TopK scheduler: unified CUDAGraph-safe kernel with dynamic per-row dispatch - DeepSeek-V3.2 DSA decode (#37421)

2026-04-08 13:35:57 -04:00

pos_encoding_kernels.cu

…

sampler.cu

[Refactor] Improve indexer decode path metadata preparation (#38865)

2026-04-08 20:49:15 -07:00

topk.cu

[Refactor] Improve indexer decode path metadata preparation (#38865)

2026-04-08 20:49:15 -07:00

torch_bindings.cpp

[Kernel] Porting the TRTLLM minimax_allreduce_rms kernels (#37045)

2026-04-11 00:20:20 +08:00

type_convert.cuh

…