llama.cpp
ggml-rpc: chunk send()/recv() to avoid EINVAL for very large tensors over RPC (macOS & others)
#15188
Merged
Merged
ggml-rpc: chunk send()/recv() to avoid EINVAL for very large tensors over RPC (macOS & others) #15188
rgerganov
merged
4 commits into
ggml-org:master
from
Tak-RS:fix/rpc-chunked-io
ggml-rpc: chunk send()/recv() to avoid EINVAL for very large tensors …
0e7aa4ee
ggml-rpc: rename RPC_IO_CHUNK->MAX_CHUNK_SIZE, use std::min() for cap…
829d6b64
rgerganov
commented
on 2025-08-11
rpc: drop n==0 special case in send_data(); retry in loop per review
b44560e0
rgerganov
approved these changes
on 2025-08-12
rgerganov
commented
on 2025-08-12
rpc: remove trailing whitespace in send_data()
7c7c3d9f
slaren
approved these changes
on 2025-08-12
Login to write a write a comment.
Login via GitHub