UofT-EcoSystem / hftaLinks
Boost hardware utilization for ML training workloads via Inter-model Horizontal Fusion
☆32Updated last year
Alternatives and similar repositories for hfta
Users that are interested in hfta are comparing it to the libraries listed below
Sorting:
- ☆92Updated 3 years ago
- PipeSwitch: Fast Pipelined Context Switching for Deep Learning Applications☆127Updated 3 years ago
- AlpaServe: Statistical Multiplexing with Model Parallelism for Deep Learning Serving (OSDI 23)☆92Updated 2 years ago
- ☆83Updated 3 years ago
- An Efficient Pipelined Data Parallel Approach for Training Large Model☆76Updated 5 years ago
- Model-less Inference Serving☆92Updated 2 years ago
- Compiler for Dynamic Neural Networks☆46Updated 2 years ago
- An experimental parallel training platform☆56Updated last year
- FTPipe and related pipeline model parallelism research.☆43Updated 2 years ago
- ☆42Updated 2 years ago
- ☆14Updated 3 years ago
- PET: Optimizing Tensor Programs with Partially Equivalent Transformations and Automated Corrections☆122Updated 3 years ago
- ☆38Updated 5 months ago
- ☆41Updated 5 years ago
- DISB is a new DNN inference serving benchmark with diverse workloads and models, as well as real-world traces.☆57Updated last year
- ☆77Updated 4 years ago
- DietCode Code Release☆64Updated 3 years ago
- ☆145Updated 10 months ago
- 🔮 Execution time predictions for deep neural network training iterations across different GPUs.☆63Updated 3 years ago
- (NeurIPS 2022) Automatically finding good model-parallel strategies, especially for complex models and clusters.☆43Updated 3 years ago
- Benchmark for matrix multiplications between dense and block sparse (BSR) matrix in TVM, blocksparse (Gray et al.) and cuSparse.☆23Updated 5 years ago
- Synthesizer for optimal collective communication algorithms☆121Updated last year
- Cavs: An Efficient Runtime System for Dynamic Neural Networks☆15Updated 5 years ago
- Artifact of OSDI '24 paper, ”Llumnix: Dynamic Scheduling for Large Language Model Serving“☆64Updated last year
- Bamboo is a system for running large pipeline-parallel DNNs affordably, reliably, and efficiently using spot instances.☆54Updated 3 years ago
- Artifacts for our ASPLOS'23 paper ElasticFlow☆55Updated last year
- ☆41Updated last year
- Supplemental materials for The ASPLOS 2025 / EuroSys 2025 Contest on Intra-Operator Parallelism for Distributed Deep Learning☆24Updated 7 months ago
- A GPU-accelerated DNN inference serving system that supports instant kernel preemption and biased concurrent execution in GPU scheduling.☆44Updated 3 years ago
- A Cluster-Wide Model Manager to Accelerate DNN Training via Automated Training Warmup☆35Updated 2 years ago