ruipeterpan / marconiLinks
Artifact for "Marconi: Prefix Caching for the Era of Hybrid LLMs" [MLSys '25 Outstanding Paper Award, Honorable Mention]
☆37Updated 7 months ago
Alternatives and similar repositories for marconi
Users that are interested in marconi are comparing it to the libraries listed below
Sorting:
- kvcached: Elastic KV cache for dynamic GPU sharing and efficient multi-LLM inference.☆94Updated last week
- 16-fold memory access reduction with nearly no loss☆105Updated 6 months ago
- ☆72Updated last year
- ☆82Updated 8 months ago
- Efficient Compute-Communication Overlap for Distributed LLM Inference☆58Updated this week
- DLSlime: Flexible & Efficient Heterogeneous Transfer Toolkit☆66Updated last week
- DeeperGEMM: crazy optimized version☆71Updated 5 months ago
- A GPU-optimized system for efficient long-context LLMs decoding with low-bit KV cache.☆60Updated last month
- ☆64Updated 5 months ago
- ☆56Updated last year
- [NeurIPS 2024] Efficient LLM Scheduling by Learning to Rank☆59Updated 11 months ago
- SpInfer: Leveraging Low-Level Sparsity for Efficient Large Language Model Inference on GPUs☆59Updated 6 months ago
- NVSHMEM‑Tutorial: Build a DeepEP‑like GPU Buffer☆132Updated 2 weeks ago
- [DAC'25] Official implement of "HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference"☆72Updated 3 months ago
- Scalable long-context LLM decoding that leverages sparsity—by treating the KV cache as a vector storage system.☆83Updated 2 weeks ago
- FlexFlow Serve: Low-Latency, High-Performance LLM Serving☆62Updated 3 weeks ago
- Distributed MoE in a Single Kernel [NeurIPS '25]☆49Updated last week
- Debug print operator for cudagraph debugging☆13Updated last year
- A lightweight design for computation-communication overlap.☆177Updated 2 weeks ago
- A Suite for Parallel Inference of Diffusion Transformers (DiTs) on multi-GPU Clusters☆50Updated last year
- ☆39Updated 2 months ago
- ☆78Updated 5 months ago
- Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity☆221Updated 2 years ago
- NEO is a LLM inference engine built to save the GPU memory crisis by CPU offloading☆62Updated 3 months ago
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆143Updated 4 months ago
- ☆50Updated 4 months ago
- ☆98Updated 4 months ago
- [NeurIPS 2025] ClusterFusion: Expanding Operator Fusion Scope for LLM Inference via Cluster-Level Collective Primitive☆41Updated last week
- ☆151Updated last year
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆128Updated 10 months ago