awslabs / Lancet-Accelerating-MoE-Training-via-Whole-Graph-Computation-Communication-OverlappingLinks
Official implementation for the paper Lancet: Accelerating Mixture-of-Experts Training via Whole Graph Computation-Communication Overlapping, published in MLSys'24.
☆12Updated 9 months ago
Alternatives and similar repositories for Lancet-Accelerating-MoE-Training-via-Whole-Graph-Computation-Communication-Overlapping
Users that are interested in Lancet-Accelerating-MoE-Training-via-Whole-Graph-Computation-Communication-Overlapping are comparing it to the libraries listed below
Sorting:
- ☆64Updated last year
- LLM serving cluster simulator☆107Updated last year
- AlpaServe: Statistical Multiplexing with Model Parallelism for Deep Learning Serving (OSDI 23)☆82Updated 2 years ago
- A lightweight design for computation-communication overlap.☆148Updated 3 weeks ago
- SpotServe: Serving Generative Large Language Models on Preemptible Instances☆123Updated last year
- nnScaler: Compiling DNN models for Parallel Training☆113Updated last week
- Compiler for Dynamic Neural Networks☆46Updated last year
- ☆106Updated 8 months ago
- ☆79Updated 2 years ago
- High performance Transformer implementation in C++.☆126Updated 6 months ago
- An interference-aware scheduler for fine-grained GPU sharing☆142Updated 5 months ago
- Artifact of OSDI '24 paper, ”Llumnix: Dynamic Scheduling for Large Language Model Serving“☆61Updated last year
- Chimera: bidirectional pipeline parallelism for efficiently training large-scale models.☆67Updated 3 months ago
- A resilient distributed training framework☆95Updated last year
- ☆80Updated 3 months ago
- DeepSeek-V3/R1 inference performance simulator☆155Updated 3 months ago
- A ChatGPT(GPT-3.5) & GPT-4 Workload Trace to Optimize LLM Serving Systems☆186Updated 9 months ago
- Since the emergence of chatGPT in 2022, the acceleration of Large Language Model has become increasingly important. Here is a list of pap…☆259Updated 4 months ago
- FlexFlow Serve: Low-Latency, High-Performance LLM Serving☆48Updated 2 months ago
- Thunder Research Group's Collective Communication Library☆38Updated last week
- ☆149Updated 11 months ago
- REEF is a GPU-accelerated DNN inference serving system that enables instant kernel preemption and biased concurrent execution in GPU sche…☆95Updated 2 years ago
- Stateful LLM Serving☆76Updated 4 months ago
- A baseline repository of Auto-Parallelism in Training Neural Networks☆144Updated 3 years ago
- ☆32Updated last year
- NEO is a LLM inference engine built to save the GPU memory crisis by CPU offloading☆44Updated last month
- LLM Inference analyzer for different hardware platforms☆80Updated last week
- ☆23Updated last year
- Artifacts for our ASPLOS'23 paper ElasticFlow☆52Updated last year
- Synthesizer for optimal collective communication algorithms☆110Updated last year