wangsiping97 / FastGEMV
High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.
☆90Updated 4 months ago
Related projects ⓘ
Alternatives and complementary repositories for FastGEMV
- ☆169Updated 4 months ago
- Standalone Flash Attention v2 kernel without libtorch dependency☆98Updated 2 months ago
- llama INT4 cuda inference with AWQ☆48Updated 4 months ago
- Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.☆50Updated 2 months ago
- ☆48Updated this week
- An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).☆211Updated 3 weeks ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆85Updated 8 months ago
- ☆79Updated 2 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆53Updated 3 weeks ago
- ☆45Updated 2 weeks ago
- ☆156Updated last year
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆29Updated 2 months ago
- Applied AI experiments and examples for PyTorch☆168Updated 3 weeks ago
- Simple and fast low-bit matmul kernels in CUDA / Triton☆147Updated this week
- TiledCUDA is a highly efficient kernel template library designed to elevate CUDA C’s level of abstraction for processing tiles.☆157Updated this week
- ☆47Updated 2 months ago
- extensible collectives library in triton☆72Updated 2 months ago
- Official PyTorch implementation of FlatQuant: Flatness Matters for LLM Quantization☆66Updated last week
- QQQ is an innovative and hardware-optimized W4A8 quantization solution for LLMs.☆89Updated last month
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆52Updated 3 months ago
- ☆132Updated 4 months ago
- Benchmark code for the "Online normalizer calculation for softmax" paper☆59Updated 6 years ago
- ☆88Updated 2 months ago
- A fast communication-overlapping library for tensor parallelism on GPUs.☆226Updated 3 weeks ago
- Dynamic Memory Management for Serving LLMs without PagedAttention☆240Updated last week
- SparseTIR: Sparse Tensor Compiler for Deep Learning☆131Updated last year
- Boosting 4-bit inference kernels with 2:4 Sparsity☆51Updated 2 months ago
- ☆55Updated 6 months ago
- A collection of memory efficient attention operators implemented in the Triton language.☆219Updated 5 months ago
- Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity☆181Updated last year