]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: ff5a3f0)
cmake : make LLAMA_NATIVE flag actually use the instructions supported by the process...
authorEve <redacted>
Tue, 3 Oct 2023 16:53:15 +0000 (16:53 +0000)
committerGitHub <redacted>
Tue, 3 Oct 2023 16:53:15 +0000 (19:53 +0300)
commit017efe899d8fa76118aef88e963210d48da01172
tree6418cb5ff0723e7b746f924f3bcb7bbb021f85fdtree
parentff5a3f0c09dfa0a8e0bf76d1748df5c6dee0e8ffcommit | diff
cmake : make LLAMA_NATIVE flag actually use the instructions supported by the processor (#3273)

* fix LLAMA_NATIVE

* syntax

* alternate implementation

* my eyes must be getting bad...

* set cmake LLAMA_NATIVE=ON by default

* march=native doesn't work for ios/tvos, so disable for those targets. also see what happens if we use it on msvc

* revert 8283237 and only allow LLAMA_NATIVE on x86 like the Makefile

* remove -DLLAMA_MPI=ON

---------

Co-authored-by: netrunnereve <redacted>
.github/workflows/build.yml diff | blob | history
CMakeLists.txt diff | blob | history
flake.nix diff | blob | history
Packaging of ggml-org/llama.cpp