MPI run on M1 Max

[ageorgios]

Prerequisites

Please answer the following questions for yourself before submitting an issue.

• [YES ] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
• [ YES] I carefully followed the README.md.
• [YES ] I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
• [ YES] I reviewed the Discussions, and have a new bug or useful enhancement to share.

Expected Behavior

run 2 chunks of the model on the same CPU-GPU

Current Behavior

ERRORS:
GGML_ASSERT: llama.cpp:8672: false && "not implemented"
GGML_ASSERT: llama.cpp:5443: false && "not implemented"

Environment and Context

Macbook M1 Max 32GB

Steps to Reproduce

cat hostfile 
127.0.0.1:2

1. mpirun -hostfile hostfile -n 2 ./main -m /Users/ageorgios/.ollama/models/blobs/sha256:29fdb92e57cf0827ded04ae6461b5931d01fa595843f55d36f5b275a52087dd2 -n 128 &> ../output.txt

Failure Logs

output.txt

[vvsotnikov]

Unfortunately, it's expected. You could try to use older versions of llama.cpp, before #3228 was merged.
MPI should be fixed in #3334, more context: #3752.

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.

[SVyatoslavG]

This is still an issue. Please reopen.