]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: b0cefea)
vulkan: Throttle the number of shader compiles during the build step. (#10222)
authorJeff Bolz <redacted>
Mon, 11 Nov 2024 17:13:51 +0000 (11:13 -0600)
committerGitHub <redacted>
Mon, 11 Nov 2024 17:13:51 +0000 (18:13 +0100)
commit54ef9cfc726a799e6f454ac22c4815d037716eda
tree0acfd44cddfb054630970270961f0f2ae9398fcdtree
parentb0cefea58a20020754b7431e3c725625274372a5commit | diff
vulkan: Throttle the number of shader compiles during the build step. (#10222)

Fixes #9582

Spawning too many concurrent copies of glslc leads to "Failed to create pipes"
errors on Linux. This change applies the same throttling we use for
multithreaded pipeline creation.
ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp diff | blob | history
Packaging of ggml-org/llama.cpp