microsoft / tokenweaveLinks
Efficient Compute-Communication Overlap for Distributed LLM Inference
☆68Updated 2 months ago
Alternatives and similar repositories for tokenweave
Users that are interested in tokenweave are comparing it to the libraries listed below
Sorting:
- NVSHMEM‑Tutorial: Build a DeepEP‑like GPU Buffer☆153Updated 4 months ago
- ☆50Updated 8 months ago
- ☆82Updated 3 months ago
- DeepXTrace is a lightweight tool for precisely diagnosing slow ranks in DeepEP-based environments.☆86Updated last month
- ☆65Updated 8 months ago
- DeeperGEMM: crazy optimized version☆73Updated 8 months ago
- DLSlime: Flexible & Efficient Heterogeneous Transfer Toolkit☆91Updated last week
- A lightweight design for computation-communication overlap.☆213Updated 3 weeks ago
- Stateful LLM Serving☆94Updated 10 months ago
- Tile-based language built for AI computation across all scales☆116Updated last week
- Nex Venus Communication Library☆72Updated 2 months ago
- ☆47Updated last year
- ☆78Updated this week
- Microsoft Collective Communication Library☆66Updated last year
- ☆92Updated 9 months ago
- Multi-Level Triton Runner supporting Python, IR, PTX, and cubin.☆83Updated 2 weeks ago
- Thunder Research Group's Collective Communication Library☆47Updated 6 months ago
- Accelerating Large-Scale Reasoning Model Inference with Sparse Self-Speculative Decoding☆79Updated last month
- SpInfer: Leveraging Low-Level Sparsity for Efficient Large Language Model Inference on GPUs☆59Updated 9 months ago
- Aims to implement dual-port and multi-qp solutions in deepEP ibrc transport☆73Updated 8 months ago
- gLLM: Global Balanced Pipeline Parallelism System for Distributed LLM Serving with Token Throttling☆52Updated last week
- FractalTensor is a programming framework that introduces a novel approach to organizing data in deep neural networks (DNNs) as a list of …☆31Updated last year
- ☆85Updated 7 months ago
- DISB is a new DNN inference serving benchmark with diverse workloads and models, as well as real-world traces.☆58Updated last year
- High-performance distributed data shuffling (all-to-all) library for MoE training and inference☆107Updated 3 weeks ago
- ☆74Updated last year
- A prefill & decode disaggregated LLM serving framework with shared GPU memory and fine-grained compute isolation.☆123Updated 3 weeks ago
- [DAC'25] Official implement of "HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference"☆96Updated last month
- A NCCL extension library, designed to efficiently offload GPU memory allocated by the NCCL communication library.☆80Updated last month
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆104Updated 6 months ago