xxyux / SpInfer
SpInfer: Leveraging Low-Level Sparsity for Efficient Large Language Model Inference on GPUs
☆42Updated last month
Alternatives and similar repositories for SpInfer:
Users that are interested in SpInfer are comparing it to the libraries listed below
- A GPU-optimized system for efficient long-context LLMs decoding with low-bit KV cache.☆33Updated last month
- ☆60Updated this week
- ☆19Updated 6 months ago
- Implement Flash Attention using Cute.☆76Updated 4 months ago
- ☆55Updated 2 weeks ago
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆63Updated 8 months ago
- ☆15Updated 3 weeks ago
- Tutorials of Extending and importing TVM with CMAKE Include dependency.☆13Updated 6 months ago
- Artifacts of EVT ASPLOS'24☆23Updated last year
- MAGIS: Memory Optimization via Coordinated Graph Transformation and Scheduling for DNN (ASPLOS'24)☆50Updated 10 months ago
- Quantized Attention on GPU☆45Updated 5 months ago
- ☆68Updated 3 months ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆73Updated 3 weeks ago
- ☆92Updated 7 months ago
- play gemm with tvm☆90Updated last year
- ThrillerFlow is a Dataflow Analysis and Codegen Framework written in Rust.☆14Updated 5 months ago
- DeeperGEMM: crazy optimized version☆67Updated 3 weeks ago
- A Vectorized N:M Format for Unleashing the Power of Sparse Tensor Cores☆51Updated last year
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆36Updated 3 weeks ago
- A practical way of learning Swizzle☆18Updated 2 months ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆91Updated 3 weeks ago
- ☆30Updated last year
- Triton to TVM transpiler.☆19Updated 6 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆82Updated this week
- ArkVale: Efficient Generative LLM Inference with Recallable Key-Value Eviction (NIPS'24)☆36Updated 4 months ago
- Magicube is a high-performance library for quantized sparse matrix operations (SpMM and SDDMM) of deep learning on Tensor Cores.☆86Updated 2 years ago
- study of Ampere' Sparse Matmul☆18Updated 4 years ago
- ☆140Updated 9 months ago
- ☆53Updated last year
- ☆29Updated 10 months ago