CUDA GEMM and GEMV for IQ4_KS_R4 and IQ5_KS_R4 #462
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
This PR is a follow up to PR #461 and adds CUDA implementation for
IQ4_KS_R4andIQ5_KS_R4Note: because GEMM is implemented via dequantize+cuBLAS, if you want to use a IQX_K_R4 DeepSeek-V3/R1 model on the GPU, you may need to build with -DGGML_CUDA_IQK_FORCE_BF16=1 to force bf16 arithmetic with cuBLAS as fp16 has been noted to lead to numerical instabilities and garbled output. I did not enable GGML_CUDA_IQK_FORCE_BF16 by default as it reduces prompt processing performance while, as far as I can tell, bf16 is only required for DeepSeek.