Skip to content

Conversation

@EikanWang
Copy link
Contributor

No description provided.

@EikanWang EikanWang merged commit 04aa449 into intel:master Jun 22, 2020
zhuhaozhe pushed a commit to zhuhaozhe/intel-extension-for-pytorch that referenced this pull request Jun 24, 2020
EikanWang pushed a commit that referenced this pull request Oct 4, 2021
… recipe on Xeon CPU (#88)

* Add launcher to auto enable best performanc recipe for DL on Xeon CPU

1) Intel OpenMP and Jemalloc/TCMalloc memory allocator optimization
2) Multi-instance launcher
3) NUMA aware multi-instance
4) Distributed launcher for MPI

Usage:
1. single instance
   >>> python -m intel_pytorch_extension.launch python_script args

2. multi-instance
   >>> python -m intel_pytorch_extension.launch --ninstances xxx --ncore_per_instance xx python_script args

3. Single-Node multi-process distributed training
   >>> python  -m intel_pytorch_extension.launch --distributed  python_script args

4. Multi-Node multi-process distributed training: (e.g. two nodes)
   rank 0: *(IP: 192.168.10.10, and has a free port: 295000)*"
   >>> python -m intel_pytorch_extension.launch --distributed --nproc_per_node=2
       --nnodes=2 --hostfile hostfile python_script args

5. >>> python -m intel_pytorch_extension.launch  --help

* Add psutil to requirment.txt
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant