NCCL Fast Socket is a transport layer plugin to improve NCCL collective communication performance on Google Cloud.
☆122Nov 15, 2023Updated 2 years ago
Alternatives and similar repositories for nccl-fastsocket
Users that are interested in nccl-fastsocket are comparing it to the libraries listed below
Sorting:
- NCCL Profiling Kit☆152Jul 1, 2024Updated last year
- RDMA and SHARP plugins for nccl library☆224Jan 12, 2026Updated last month
- High performance NCCL plugin for Bagua.☆15Sep 15, 2021Updated 4 years ago
- This is a plugin which lets EC2 developers use libfabric as network provider while running NCCL applications.☆205Updated this week
- Microsoft Collective Communication Library☆385Sep 20, 2023Updated 2 years ago
- pytorch ucc plugin☆23Jul 8, 2021Updated 4 years ago
- [DEPRECATED] Moved to ROCm/rocm-systems repo☆411Feb 23, 2026Updated last week
- ☆26May 19, 2021Updated 4 years ago
- ☆47Dec 13, 2024Updated last year
- ☆68Mar 14, 2023Updated 2 years ago
- An external memory allocator example for PyTorch.☆16Aug 10, 2025Updated 6 months ago
- GPUDirect Async support for IB Verbs☆135Nov 10, 2022Updated 3 years ago
- NCCL Tests☆1,446Feb 9, 2026Updated 3 weeks ago
- ☆387Apr 23, 2024Updated last year
- Optimized primitives for collective multi-GPU communication☆4,474Updated this week
- MSCCL++: A GPU-driven communication stack for scalable AI applications☆475Updated this week
- TACCL: Guiding Collective Algorithm Synthesis using Communication Sketches☆80Jul 25, 2023Updated 2 years ago
- PSTensor provides a way to hack the memory management of tensors in TensorFlow and PyTorch by defining your own C++ Tensor Class.☆10Feb 10, 2022Updated 4 years ago
- A fast GPU memory copy library based on NVIDIA GPUDirect RDMA technology☆1,347Dec 17, 2025Updated 2 months ago
- Herald: Accelerating Neural Recommendation Training with Embedding Scheduling (NSDI 2024)☆23May 9, 2024Updated last year
- ☆26Feb 17, 2025Updated last year
- Prototype of OpenSHMEM for NVIDIA GPUs, developed as part of DoE Design Forward☆25Apr 26, 2018Updated 7 years ago
- Synthesizer for optimal collective communication algorithms☆124Apr 8, 2024Updated last year
- Collective communications library with various primitives for multi-machine training.☆1,400Feb 12, 2026Updated 2 weeks ago
- Microsoft Collective Communication Library☆66Nov 23, 2024Updated last year
- Hi-Speed DNN Training with Espresso: Unleashing the Full Potential of Gradient Compression with Near-Optimal Usage Strategies (EuroSys '2…☆15Sep 21, 2023Updated 2 years ago
- A tool for bandwidth measurements on NVIDIA GPUs.☆631Apr 15, 2025Updated 10 months ago
- [DEPRECATED] Moved to ROCm/rocm-systems repo☆86Feb 11, 2026Updated 2 weeks ago
- gossip: Efficient Communication Primitives for Multi-GPU Systems☆62Jul 1, 2022Updated 3 years ago
- A Micro-benchmarking Tool for HPC Networks☆34Sep 2, 2025Updated 6 months ago
- An IR for efficiently simulating distributed ML computation.☆32Jan 13, 2024Updated 2 years ago
- Fairring (FAIR + Herring) is a plug-in for PyTorch that provides a process group for distributed training that outperforms NCCL at large …☆65Mar 21, 2022Updated 3 years ago
- Switch-based Training Acceleration for Machine Learning (SwitchML)☆16Apr 13, 2021Updated 4 years ago
- NVIDIA Data Center GPU Manager (DCGM) is a project for gathering telemetry and measuring the health of NVIDIA GPUs☆671Feb 17, 2026Updated last week
- Fine-grained GPU sharing primitives☆148Jul 28, 2025Updated 7 months ago
- Ok-Topk is a scheme for distributed training with sparse gradients. Ok-Topk integrates a novel sparse allreduce algorithm (less than 6k c…☆27Dec 10, 2022Updated 3 years ago
- TransferBench is a utility capable of benchmarking simultaneous copies between user-specified devices (CPUs/GPUs)☆57Updated this week
- ☆16Apr 22, 2025Updated 10 months ago
- ☆27Feb 9, 2026Updated 3 weeks ago