unist-ssl / JABAS
"JABAS: Joint Adaptive Batching and Automatic Scaling for DNN Training on Heterogeneous GPUs" (EuroSys '25)
☆12Updated last month
Related projects ⓘ
Alternatives and complementary repositories for JABAS
- ☆12Updated last month
- Open-source repository for the paper "DyTIS: A Dynamic Dataset Targeted Index Structure Simultaneously Efficient for Search, Insert, and …☆14Updated last year
- Fast and Efficient Model Serving Using Multi-GPUs with Direct-Host-Access (ACM EuroSys '23)☆54Updated 7 months ago
- ☆11Updated 10 months ago
- Exploring the Design Space of Page Management for Multi-Tiered Memory Systems (USENIX ATC '21)☆43Updated 2 years ago
- ☆21Updated last year
- ☆32Updated last year
- ☆41Updated last year
- ☆23Updated 2 years ago
- LLMServingSim: A HW/SW Co-Simulation Infrastructure for LLM Inference Serving at Scale☆55Updated 2 weeks ago
- Ginex: SSD-enabled Billion-scale Graph Neural Network Training on a Single Machine via Provably Optimal In-memory Caching☆35Updated 4 months ago
- On Stacking a Persistent Memory File System on Legacy File Systems [FAST '23]☆15Updated last year
- Sources for the Multi-Clock system as described in the paper: MULTI-CLOCK: Dynamic Tiering for Hybrid Memory Systems, HPCA 2022.☆19Updated 2 years ago
- ☆48Updated last year
- SHADE: Enable Fundamental Cacheability for Distributed Deep Learning Training☆29Updated last year
- ☆14Updated 5 months ago
- ☆14Updated 3 years ago
- ☆23Updated last year
- ☆26Updated 4 months ago
- Proteus: A High-Throughput Inference-Serving System with Accuracy Scaling☆8Updated 8 months ago
- ☆12Updated last year
- Cluster Far Mem, framework to execute single job and multi job experiments using fastswap☆21Updated 10 months ago
- ☆60Updated last year
- ☆24Updated last year
- Nap - NUMA-Aware Persistent Indexes☆41Updated 3 years ago
- Canvas: Isolated and Adaptive Swapping for Multi-Applications on Remote Memory☆36Updated last year
- Hi-Speed DNN Training with Espresso: Unleashing the Full Potential of Gradient Compression with Near-Optimal Usage Strategies (EuroSys '2…☆15Updated last year
- Bamboo is a system for running large pipeline-parallel DNNs affordably, reliably, and efficiently using spot instances.☆46Updated last year
- Adaptive Page Migration Policy with Huge Pages in Tiered Memory Systems☆13Updated 3 years ago
- This is the respository that holds the artifacts of MICRO'23 -- Demystifying CXL Memory with True CXL-Ready Systems and CXL Memory Device…☆35Updated 7 months ago