Multi-LLM / prism-researchLinks
Research prototype of PRISM — a cost-efficient multi-LLM serving system with flexible time- and space-based GPU sharing.
☆20Updated 2 weeks ago
Alternatives and similar repositories for prism-research
Users that are interested in prism-research are comparing it to the libraries listed below
Sorting:
- Stateful LLM Serving☆81Updated 5 months ago
- [NeurIPS 2024] Efficient LLM Scheduling by Learning to Rank☆58Updated 10 months ago
- [OSDI'24] Serving LLM-based Applications Efficiently with Semantic Variable☆181Updated 11 months ago
- ☆131Updated last month
- High performance Transformer implementation in C++.☆129Updated 7 months ago
- ☆69Updated last year
- kvcached: Elastic KV cache for dynamic GPU sharing and efficient multi-LLM inference.☆66Updated this week
- A framework for generating realistic LLM serving workloads☆54Updated 2 months ago
- ☆78Updated 4 months ago
- ☆116Updated 9 months ago
- Artifact for "Marconi: Prefix Caching for the Era of Hybrid LLMs" [MLSys '25 Outstanding Paper Award, Honorable Mention]☆21Updated 5 months ago
- SpotServe: Serving Generative Large Language Models on Preemptible Instances☆128Updated last year
- ☆288Updated last week
- ☆123Updated 10 months ago
- ☆75Updated 3 years ago
- Compare different hardware platforms via the Roofline Model for LLM inference tasks.☆112Updated last year
- Scalable long-context LLM decoding that leverages sparsity—by treating the KV cache as a vector storage system.☆77Updated 2 weeks ago
- A tiny yet powerful LLM inference system tailored for researching purpose. vLLM-equivalent performance with only 2k lines of code (2% of …☆249Updated 2 months ago
- A record of reading list on some MLsys popular topic☆15Updated 5 months ago
- A prefill & decode disaggregated LLM serving framework with shared GPU memory and fine-grained compute isolation.☆105Updated 3 months ago
- FlexFlow Serve: Low-Latency, High-Performance LLM Serving☆56Updated this week
- gLLM: Global Balanced Pipeline Parallelism System for Distributed LLM Serving with Token Throttling☆39Updated last week
- Allow torch tensor memory to be released and resumed later☆118Updated this week
- PyTorch library for cost-effective, fast and easy serving of MoE models.☆228Updated last month
- ☆16Updated 2 months ago
- A lightweight design for computation-communication overlap.☆160Updated last week
- ☆42Updated last year
- Since the emergence of chatGPT in 2022, the acceleration of Large Language Model has become increasingly important. Here is a list of pap…☆266Updated 5 months ago
- NEO is a LLM inference engine built to save the GPU memory crisis by CPU offloading☆56Updated 2 months ago
- Systems for GenAI☆144Updated 4 months ago