google / nccl-fastsocketLinks
NCCL Fast Socket is a transport layer plugin to improve NCCL collective communication performance on Google Cloud.
☆121Updated last year
Alternatives and similar repositories for nccl-fastsocket
Users that are interested in nccl-fastsocket are comparing it to the libraries listed below
Sorting:
- RDMA and SHARP plugins for nccl library☆209Updated last month
- NCCL Profiling Kit☆145Updated last year
- Microsoft Collective Communication Library☆367Updated 2 years ago
- Fine-grained GPU sharing primitives☆144Updated 2 months ago
- PArametrized Recommendation and Ai Model benchmark is a repository for development of numerous uBenchmarks as well as end to end nets for…☆153Updated last week
- Synthesizer for optimal collective communication algorithms☆118Updated last year
- ☆376Updated last year
- NVIDIA Resiliency Extension is a python package for framework developers and users to implement fault-tolerant features. It improves the …☆226Updated this week
- GPU-scheduler-for-deep-learning☆210Updated 4 years ago
- Microsoft Collective Communication Library☆66Updated 10 months ago
- MSCCL++: A GPU-driven communication stack for scalable AI applications☆425Updated this week
- ☆46Updated 10 months ago
- This is a plugin which lets EC2 developers use libfabric as network provider while running NCCL applications.☆186Updated this week
- Paella: Low-latency Model Serving with Virtualized GPU Scheduling☆62Updated last year
- An Efficient Pipelined Data Parallel Approach for Training Large Model☆76Updated 4 years ago
- PipeSwitch: Fast Pipelined Context Switching for Deep Learning Applications☆126Updated 3 years ago
- HierarchicalKV is a part of NVIDIA Merlin and provides hierarchical key-value storage to meet RecSys requirements. The key capability of…☆173Updated 3 weeks ago
- ☆83Updated 4 months ago
- NVIDIA NCCL Tests for Distributed Training☆114Updated last week
- pytorch ucc plugin☆23Updated 4 years ago
- NVIDIA Inference Xfer Library (NIXL)☆673Updated this week
- Offline optimization of your disaggregated Dynamo graph☆79Updated this week
- Ultra and Unified CCL☆595Updated this week
- Tiresias is a GPU cluster manager for distributed deep learning training.☆162Updated 5 years ago
- ☆83Updated 2 years ago
- CloudAI Benchmark Framework☆71Updated last week
- Code for "Heterogenity-Aware Cluster Scheduling Policies for Deep Learning Workloads", which appeared at OSDI 2020☆130Updated last year
- A validation and profiling tool for AI infrastructure☆341Updated this week
- Thunder Research Group's Collective Communication Library☆42Updated 3 months ago
- Splits single Nvidia GPU into multiple partitions with complete compute and memory isolation (wrt to performace) between the partitions☆159Updated 6 years ago