llama.cpp
a15ef8f8 - CUDA: fix partial offloading for ne0 % 256 != 0 (#8572)

Commit

328 days ago

CUDA: fix partial offloading for ne0 % 256 != 0 (#8572)

References

#8572 - CUDA: fix partial offloading for ne0 % 256 != 0

Author

JohannesGaessler

JohannesGaessler

Parents

Files4

ggml
- include
  - ggml-backend.h
- src
  - ggml-alloc.c
  - ggml-backend.c
  - ggml-cuda.cu