git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit

]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit

overview / pkg / ggml / sources / llama.cpp / commit

author	Frosty40 <redacted>
	Tue, 16 Jun 2026 05:35:00 +0000 (00:35 -0500)
committer	GitHub <redacted>
	Tue, 16 Jun 2026 05:35:00 +0000 (08:35 +0300)
commit	ac79caa7ce61dbd800451ce43adb3dd9bf2f633b
tree	ca38c08b2cecee5109dccc00b687f80144bfef2e	tree
parent	fdd109883df94789729cf3c411b709e0ecb129cc	commit \| diff

sycl: support reordered Q4_K/Q5_K/Q6_K MoE MUL_MAT_ID (#24452)

* sycl: support reordered Q4_K and Q5_K MoE MUL_MAT_ID

Extend reordered-weight handling to fused MoE MUL_MAT_ID for Q4_K and Q5_K expert tensors and add Q5_K reordered DMMV coverage. Unsupported 3D reorder cases now fall back instead of aborting.

* sycl: extend MoE reorder to Q6_K mul_mat_id

ggml/src/ggml-sycl/dmmv.cpp		diff \| blob \| history
ggml/src/ggml-sycl/ggml-sycl.cpp		diff \| blob \| history
ggml/src/ggml-sycl/mmvq.cpp		diff \| blob \| history
ggml/src/ggml-sycl/mmvq.hpp		diff \| blob \| history

Packaging of ggml-org/llama.cpp

RSS Atom