]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 5106764)
CUDA/HIP: Share the same unified memory allocation logic. (#12934)
authorDavid Huang <redacted>
Tue, 15 Apr 2025 09:20:38 +0000 (17:20 +0800)
committerGitHub <redacted>
Tue, 15 Apr 2025 09:20:38 +0000 (11:20 +0200)
commit84778e97703740d8ac5fb64e14d83b80eafa0f3c
treeca6786fe359efc9cd5de538ed2546e0ed51d8a41tree
parent510676475f885ec064ff147af9f20ee7a9b12a50commit | diff
CUDA/HIP: Share the same unified memory allocation logic. (#12934)

Replace compile-time `GGML_HIP_UMA` with environment variable `GGML_CUDA_ENABLE_UNIFIED_MEMORY`. This unifies the usage on NVIDIA and AMD GPUs, and allows a single binary to be shared between integrated and dedicated GPUs.
Makefile diff | blob | history
docs/build.md diff | blob | history
ggml/CMakeLists.txt diff | blob | history
ggml/src/ggml-cuda/ggml-cuda.cu diff | blob | history
ggml/src/ggml-cuda/vendors/hip.h diff | blob | history
ggml/src/ggml-hip/CMakeLists.txt diff | blob | history
Packaging of ggml-org/llama.cpp