feifeibear / LLMRooflineLinks
Compare different hardware platforms via the Roofline Model for LLM inference tasks.
☆100Updated last year
Alternatives and similar repositories for LLMRoofline
Users that are interested in LLMRoofline are comparing it to the libraries listed below
Sorting:
- ☆85Updated 2 months ago
- ☆95Updated 8 months ago
- A prefill & decode disaggregated LLM serving framework with shared GPU memory and fine-grained compute isolation.☆80Updated 2 weeks ago
- ☆76Updated last month
- A lightweight design for computation-communication overlap.☆132Updated 3 weeks ago
- ☆137Updated 2 months ago
- [USENIX ATC '24] Accelerating the Training of Large Language Models using Efficient Activation Rematerialization and Optimal Hybrid Paral…☆54Updated 10 months ago
- ☆127Updated 5 months ago
- ☆138Updated last year
- ☆59Updated last month
- A simple calculation for LLM MFU.☆38Updated 2 months ago
- Fast and memory-efficient exact attention☆72Updated 3 weeks ago
- DeepSeek-V3/R1 inference performance simulator☆134Updated 2 months ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆79Updated 3 weeks ago
- PyTorch distributed training acceleration framework☆49Updated 3 months ago
- QQQ is an innovative and hardware-optimized W4A8 quantization solution for LLMs.☆124Updated last month
- [DAC'25] Official implement of "HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference"☆50Updated 2 weeks ago
- High performance Transformer implementation in C++.☆124Updated 4 months ago
- ☆71Updated 2 weeks ago
- Sequence-level 1F1B schedule for LLMs.☆24Updated this week
- Stateful LLM Serving☆70Updated 2 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆37Updated 3 months ago
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆36Updated 2 months ago
- ☆148Updated 4 months ago
- An easy-to-use package for implementing SmoothQuant for LLMs☆99Updated last month
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆92Updated this week
- Standalone Flash Attention v2 kernel without libtorch dependency☆109Updated 8 months ago
- A GPU-optimized system for efficient long-context LLMs decoding with low-bit KV cache.☆36Updated last month
- Summary of system papers/frameworks/codes/tools on training or serving large model☆57Updated last year
- ☆99Updated 6 months ago