-
Notifications
You must be signed in to change notification settings - Fork 829
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support CUDA without cuBLAS #82
Merged
Merged
Commits on Dec 10, 2023
-
wip naive cublasGemmStridedBatchedEx
Numbers match cublas, but using this code leads to LLaVA outputting nothing but white squares.
Configuration menu - View commit details
-
Copy full SHA for 7118b15 - Browse repository at this point
Copy the full SHA 7118b15View commit details
Commits on Dec 11, 2023
-
Configuration menu - View commit details
-
Copy full SHA for 3d47d97 - Browse repository at this point
Copy the full SHA 3d47d97View commit details -
Configuration menu - View commit details
-
Copy full SHA for c2a7a5e - Browse repository at this point
Copy the full SHA c2a7a5eView commit details -
Configuration menu - View commit details
-
Copy full SHA for 4fb0813 - Browse repository at this point
Copy the full SHA 4fb0813View commit details
Commits on Dec 12, 2023
-
Configuration menu - View commit details
-
Copy full SHA for ffb039d - Browse repository at this point
Copy the full SHA ffb039dView commit details -
Configuration menu - View commit details
-
Copy full SHA for 43132ff - Browse repository at this point
Copy the full SHA 43132ffView commit details -
Remove remaining cublas library calls in naive mode
Uses some fairly disgusting preprocessor macros to get the job done while preserving behavior when `-DGGML_USE_CUBLAS`. With a bit of investigation into `ggml_cuda_mul_mat_mat_batched_cublas`, these can probably be removed or simplified.
Configuration menu - View commit details
-
Copy full SHA for 877c736 - Browse repository at this point
Copy the full SHA 877c736View commit details -
add header file, remove cublas_v2.h from naive
N.B. we include the source file rather than the header file in `ggml-cuda.cu` because `llamafile/cuda.c` assumes that everything lives in a single compilation unit.
Configuration menu - View commit details
-
Copy full SHA for 89f721d - Browse repository at this point
Copy the full SHA 89f721dView commit details -
Configuration menu - View commit details
-
Copy full SHA for 16c9276 - Browse repository at this point
Copy the full SHA 16c9276View commit details -
Configuration menu - View commit details
-
Copy full SHA for 211be30 - Browse repository at this point
Copy the full SHA 211be30View commit details -
Configuration menu - View commit details
-
Copy full SHA for 862bce5 - Browse repository at this point
Copy the full SHA 862bce5View commit details -
Configuration menu - View commit details
-
Copy full SHA for 881ebfd - Browse repository at this point
Copy the full SHA 881ebfdView commit details
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.