llama.cpp
ggml-cuda: fix ROCm multi-GPU illegal memory access in recurrent state restore
#21170
Open
Open
ggml-cuda: fix ROCm multi-GPU illegal memory access in recurrent state restore #21170
uaruss
wants to merge
1 commit into
ggml-org:master
from
uaruss:fix/rocm-multigpu-set-device-recurrent
ggml-cuda: fix ROCm multi-GPU illegal memory access in recurrent stat…
5d9f64c5
Reviewers
No reviews
Assignees
No one assigned
Labels
Nvidia GPU
ggml
Milestone
No milestone
Login to write a write a comment.
Login via GitHub