]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 6152129)
vulkan: fix diag_mask_inf (#11323)
authorJeff Bolz <redacted>
Thu, 23 Jan 2025 07:01:17 +0000 (01:01 -0600)
committerGitHub <redacted>
Thu, 23 Jan 2025 07:01:17 +0000 (08:01 +0100)
commit5245729e3317064959faefc5e5cbc63f4e9cfbea
tree9fc8451641eeb4ed97c03602eef3a20876ce2282tree
parent6152129d05870cb38162c422c6ba80434e021e9fcommit | diff
vulkan: fix diag_mask_inf (#11323)

With robustbufferaccess disabled, this shader was showing OOB stores. There
is a bounds check in the code, but the workgrouop dimensions were reversed vs
CUDA and it was running the wrong number of threads. So fix the workgroup
dimensions and disable robustness for this pipeline.
ggml/src/ggml-vulkan/ggml-vulkan.cpp diff | blob | history
ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp diff | blob | history
Packaging of ggml-org/llama.cpp