]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
CUDA: use shared mem for ssm_conv (#20128)
authorAman Gupta <redacted>
Fri, 6 Mar 2026 15:09:59 +0000 (23:09 +0800)
committerGitHub <redacted>
Fri, 6 Mar 2026 15:09:59 +0000 (23:09 +0800)
commit1e38a7a6fa115de0a2731cb67ce554b7df5e8e2c
tree58d2fadd5ba949b3a4a0953a3aa98060cdce7be5
parent388baabc06be4cbcb64b546c4b67a1aa2f64858b
CUDA: use shared mem for ssm_conv (#20128)

* CUDA: use shared mem for ssm_conv

* fuse silu + ssm_conv

* fuse unary + mul

* enable for fp16

* formatting

Co-authored-by: Johannes Gäßler <redacted>
---------

Co-authored-by: Johannes Gäßler <redacted>
ggml/src/ggml-cuda/ggml-cuda.cu
ggml/src/ggml-cuda/ssm-conv.cu
ggml/src/ggml-cuda/ssm-conv.cuh
ggml/src/ggml-cuda/unary.cu
ggml/src/ggml-cuda/unary.cuh
tests/test-backend-ops.cpp