ParCIS / Ok-TopkLinks
Ok-Topk is a scheme for distributed training with sparse gradients. Ok-Topk integrates a novel sparse allreduce algorithm (less than 6k communication volume which is asymptotically optimal) with the decentralized parallel Stochastic Gradient Descent (SGD) optimizer, and its convergence is proved theoretically and empirically.
☆27Updated 3 years ago
Alternatives and similar repositories for Ok-Topk
Users that are interested in Ok-Topk are comparing it to the libraries listed below
Sorting:
- Hi-Speed DNN Training with Espresso: Unleashing the Full Potential of Gradient Compression with Near-Optimal Usage Strategies (EuroSys '2…☆15Updated 2 years ago
- ☆14Updated 4 years ago
- A Cluster-Wide Model Manager to Accelerate DNN Training via Automated Training Warmup☆35Updated 3 years ago
- ☆25Updated 2 years ago
- Artifacts for our ASPLOS'23 paper ElasticFlow☆55Updated last year
- SHADE: Enable Fundamental Cacheability for Distributed Deep Learning Training☆35Updated 2 years ago
- Artifacts for our SIGCOMM'22 paper Muri☆43Updated 2 years ago
- Artifact for OSDI'23: MGG: Accelerating Graph Neural Networks with Fine-grained intra-kernel Communication-Computation Pipelining on Mult…☆41Updated last year
- ☆21Updated 3 years ago
- Artifact for PPoPP22 QGTC: Accelerating Quantized GNN via GPU Tensor Core.☆30Updated 3 years ago
- ☆52Updated 3 years ago
- ☆10Updated 2 years ago
- ☆31Updated last year
- ☆38Updated 6 months ago
- ☆17Updated last year
- Model-less Inference Serving☆93Updated 2 years ago
- Artifact for "Shockwave: Fair and Efficient Cluster Scheduling for Dynamic Adaptation in Machine Learning" [NSDI '23]☆46Updated 3 years ago
- DISB is a new DNN inference serving benchmark with diverse workloads and models, as well as real-world traces.☆58Updated last year
- A GPU-accelerated DNN inference serving system that supports instant kernel preemption and biased concurrent execution in GPU scheduling.☆44Updated 3 years ago
- Proteus: A High-Throughput Inference-Serving System with Accuracy Scaling☆12Updated last year
- Open-source implementation for "Helix: Serving Large Language Models over Heterogeneous GPUs and Network via Max-Flow"☆76Updated 3 months ago
- ☆56Updated 4 years ago
- ☆25Updated 3 years ago
- AlpaServe: Statistical Multiplexing with Model Parallelism for Deep Learning Serving (OSDI 23)☆92Updated 2 years ago
- Compiler for Dynamic Neural Networks☆45Updated 2 years ago
- ☆17Updated 3 years ago
- Tacker: Tensor-CUDA Core Kernel Fusion for Improving the GPU Utilization while Ensuring QoS☆33Updated 11 months ago
- Dual-way gradient sparsification approach for async DNN training, based on PyTorch.☆11Updated 3 years ago
- LLM serving cluster simulator☆132Updated last year
- Source code for the paper: "A Latency-Predictable Multi-Dimensional Optimization Framework forDNN-driven Autonomous Systems"☆22Updated 5 years ago