ai-dynamo / nixlLinks
NVIDIA Inference Xfer Library (NIXL)
☆770Updated this week
Alternatives and similar repositories for nixl
Users that are interested in nixl are comparing it to the libraries listed below
Sorting:
- Efficient and easy multi-instance LLM serving☆517Updated 3 months ago
- Perplexity GPU Kernels☆539Updated last month
- KV cache store for distributed LLM inference☆376Updated last month
- Virtualized Elastic KV Cache for Dynamic GPU Sharing and Beyond☆713Updated 3 weeks ago
- MSCCL++: A GPU-driven communication stack for scalable AI applications☆444Updated this week
- Disaggregated serving system for Large Language Models (LLMs).☆754Updated 8 months ago
- A low-latency & high-throughput serving engine for LLMs☆456Updated 2 months ago
- OME is a Kubernetes operator for enterprise-grade management and serving of Large Language Models (LLMs)☆341Updated this week
- Dynamic Memory Management for Serving LLMs without PagedAttention☆448Updated 6 months ago
- NVIDIA Resiliency Extension is a python package for framework developers and users to implement fault-tolerant features. It improves the …☆239Updated last week
- Materials for learning SGLang☆693Updated last week
- Byted PyTorch Distributed for Hyperscale Training of LLMs and RLs☆903Updated 3 weeks ago
- GLake: optimizing GPU memory management and IO transmission.☆491Updated 8 months ago
- A tool for bandwidth measurements on NVIDIA GPUs.☆583Updated 8 months ago
- Offline optimization of your disaggregated Dynamo graph☆128Updated this week
- torchcomms: a modern PyTorch communications API☆309Updated this week
- CUDA checkpoint and restore utility☆397Updated 3 months ago
- NVIDIA NCCL Tests for Distributed Training☆129Updated this week
- Distributed Compiler based on Triton for Parallel Systems☆1,280Updated this week
- ☆329Updated last month
- Genai-bench is a powerful benchmark tool designed for comprehensive token-level performance evaluation of large language model (LLM) serv…☆249Updated last week
- NCCL Fast Socket is a transport layer plugin to improve NCCL collective communication performance on Google Cloud.☆122Updated 2 years ago
- Zero Bubble Pipeline Parallelism☆442Updated 7 months ago
- A throughput-oriented high-performance serving framework for LLMs☆924Updated last month
- Microsoft Collective Communication Library☆377Updated 2 years ago
- A fast communication-overlapping library for tensor/expert parallelism on GPUs.☆1,208Updated 3 months ago
- NVIDIA Data Center GPU Manager (DCGM) is a project for gathering telemetry and measuring the health of NVIDIA GPUs☆631Updated 2 weeks ago
- NCCL Profiling Kit☆149Updated last year
- RDMA and SHARP plugins for nccl library☆217Updated last month
- Train speculative decoding models effortlessly and port them smoothly to SGLang serving.☆557Updated this week