]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
metal: optimise `GGML_OP_SUM` (llama/16559)
authorSam/Samuel <redacted>
Wed, 15 Oct 2025 14:05:56 +0000 (23:05 +0900)
committerGeorgi Gerganov <redacted>
Tue, 21 Oct 2025 15:14:33 +0000 (18:14 +0300)
commit270da010c0a85a557fc35f796dbd0706fb18bb2f
tree651290a4ed31b78bd390dc936f43e2e25aa4a92f
parentca02fcb733f2a108c5663e1b30862cd98ae3f4ae
metal: optimise `GGML_OP_SUM` (llama/16559)

* optimise GGML_OP_SUM

* add non-contiguous tests by permuting the input

* change tests to require full contiguity of OP_SUM

* cuda : add check GGML_OP_SUM

---------

Co-authored-by: Georgi Gerganov <redacted>
src/ggml-cuda/ggml-cuda.cu
src/ggml-metal/ggml-metal-device.m
src/ggml-metal/ggml-metal-ops.cpp
src/ggml-metal/ggml-metal.metal
tests/test-backend-ops.cpp