]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
vulkan: clamp matmul and FA results to the max finite value (#15652)
authorJeff Bolz <redacted>
Sun, 31 Aug 2025 06:27:57 +0000 (01:27 -0500)
committerGitHub <redacted>
Sun, 31 Aug 2025 06:27:57 +0000 (08:27 +0200)
commit94e82c7eadeb8fff0db4bfd1ab6d8cf65fa6f2e0
tree887d0e477d54dc48d34ef489ea3d82342716a982
parent4d74393bcc956ccd7df68a6a06d1a0575cfa712c
vulkan: clamp matmul and FA results to the max finite value (#15652)

* vulkan: clamp matmul and FA results to the max finite value

* only clamp for fp16
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_split_k_reduce.comp
ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp
ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp
ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp