Skip to content

Conversation

ggerganov
Copy link
Member

@ggerganov ggerganov commented Sep 13, 2024

Using set(GGML_XXX ON) before option(GGML_XXX OFF) will hide the options from configuration tools like ccmake. To fix this, we define GGML_XXX_DEFAULT and use it in option(GGML_XXX ${GGML_XXX_DEFAULT}). This allows ggml-based projects such as llama.cpp override the default option without hiding it from the config.

Also rename GGML_CUDA_USE_GRAPHS CMake option to GGML_CUDA_GRAPHS for consistency.

@github-actions github-actions bot added the build Compilation issues label Sep 13, 2024
@ggerganov ggerganov changed the title cmake : do not hide GGML options cmake : do not hide GGML options + rename option Sep 16, 2024
@ggerganov ggerganov merged commit 19514d6 into master Sep 16, 2024
59 of 60 checks passed
@ggerganov ggerganov deleted the gg/cmake-defaults branch September 16, 2024 07:27
dsx1986 pushed a commit to dsx1986/llama.cpp that referenced this pull request Oct 29, 2024
* cmake : do not hide GGML options

ggml-ci

* build : rename flag GGML_CUDA_USE_GRAPHS -> GGML_CUDA_GRAPHS

for consistency

ggml-ci
arthw pushed a commit to arthw/llama.cpp that referenced this pull request Nov 15, 2024
* cmake : do not hide GGML options

ggml-ci

* build : rename flag GGML_CUDA_USE_GRAPHS -> GGML_CUDA_GRAPHS

for consistency

ggml-ci
arthw pushed a commit to arthw/llama.cpp that referenced this pull request Nov 18, 2024
* cmake : do not hide GGML options

ggml-ci

* build : rename flag GGML_CUDA_USE_GRAPHS -> GGML_CUDA_GRAPHS

for consistency

ggml-ci
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

build Compilation issues

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant