]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 249a790)
vulkan: Handle GPUs with less shared memory (#10468)
authorJeff Bolz <redacted>
Wed, 27 Nov 2024 07:30:27 +0000 (01:30 -0600)
committerGitHub <redacted>
Wed, 27 Nov 2024 07:30:27 +0000 (08:30 +0100)
commit5b3466bedfa84aa29c6871c7254467550186ecc6
tree7d8537160ffdc5dfb8c2590b39e17a2e8b34775ftree
parent249a7902ec710c8d027b9cc0ed10219d2b4184f8commit | diff
vulkan: Handle GPUs with less shared memory (#10468)

There have been reports of failure to compile on systems with <= 32KB
of shared memory (e.g. #10037). This change makes the large tile size
fall back to a smaller size if necessary, and makes mul_mat_id fall
back to CPU if there's only 16KB of shared memory.
ggml/src/ggml-vulkan/ggml-vulkan.cpp diff | blob | history
Packaging of ggml-org/llama.cpp