concepts
Learn concepts in NCCL.
concepts
- node: 一个服务器
- CUDA device:一个GPU
- rank: GPU的编号
- communicator: 这是一个抽象的概念,用于管理一组GPU之间的通信。
- IB GDA
temp
full-blown
multi-GPU parallelization model:
- single-threaded control of all GPUs
- multi-threaded, for example, using one thread per GPU
- multi-process, for example, MPI
SHARP
- NVLink Sharp
- IB Sharp
Algorithm
- NVLS
- IB Sharp
Feedback
Was this page helpful?
Glad to hear it! Please tell us how we can improve.
Sorry to hear that. Please tell us how we can improve.