A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on Hopper, Ada and Blackwell GPUs, to provide better performance with lower memory utilization in both training and inference.

cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.

Microsoft Automatic Mixed Precision Library

An innovative library for efficient LLM inference via low-bit quantization

Flux diffusion model implementation using quantized fp8 matmul & remaining layers use faster half precision accumulate, which is ~2x faster on consumer devices.

vLLM runtime patch-overlay for Qwen3.6 + Gemma4 on consumer NVIDIA (Ampere sm_86, 2x A5000/3090) — Qwen3.6-35B-A3B FP8 ~244 tok/s, 27B-int4 hybrid GDN+Mamba, Gemma4 26B/31B AWQ, 256K ctx. 321 patches: TurboQuant k8v4 KV, MTP/DFlash spec-decode, FULL cudagraph, hybrid GDN. vLLM pin dev424 + SNDR Control Center GUI.

FP8 Metal compute kernels for Apple Silicon MPS — fixing what PyTorch doesn't support yet. FLUX/SD3.5/ComfyUI on Mac.

JAX Scalify: end-to-end scaled arithmetics

Reproducible recipe: serve abliterated Gemma-4-12B (gemma4_unified) at 50-118 tok/s on no-NVLink Blackwell (SM120) via vLLM nightly + ModelOpt FP8/NVFP4 + MTP spec-decode.

Optimized vLLM setup for Qwen3.6-27B-FP8 on dual RTX PRO 6000 Blackwell (192 GB GDDR7, no NVLink) ; config, benchmark sweep results, and custom chat template with thinking mode off by default.

LLM inference engine for AMD RDNA4 — Rust + Vulkan compute shaders, gguf & native FP8.

Automated Triton w8a8 block FP8 kernel tuning tool for vLLM. Auto-detects model architecture, supports Qwen3-Coder-30B-A3B-Instruct-FP8/DeepSeek-V3/custom models, multi-GPU parallel tuning, and generates optimized kernel configs for quantization.

Source-code level analysis of LLM RL training infra: async RL, weight sync, FP8, MoE routing | LLM RL 训练基础设施源码级分析

An stress and benchmark utility for NVIDIA GPUs. Measures performance across various precisions (FP64, FP32, TF32, FP16, INT8) and monitors real-time vitals like power, temperature, and clock speeds.

Narrow precision floating point types

Pure Go machine learning framework. Train, run, and serve ML models with go build. Zero CGo.

Spike, a RISC-V ISA Simulator with added 8-bit vector floating point support

Systematic 24-hour benchmark study of Qwen3.6-27B inference on dual NVIDIA RTX PRO 6000 Blackwell SM120 (TP=2). 8 experiments comparing repne/vllm fork vs upstream vLLM across FP8/BF16/NVFP4/Q8_0 quants and MTP/DFlash speculative decoding. Peak: 2,083 tok/s at c=32. Quality: KLD vs BF16 = 0.0018 (noise floor).

Fused Triton kernels for Transformer inference: RMSNorm+RoPE, Gated MLP, FP8 GEMM — CPU-testable references, autotuning, and benchmarking

Hybrid Mamba-2 + Transformer 2.94B LLM (Nemotron-H style) — Korean 3B model pretrained from scratch on 7× NVIDIA B200 GPUs with SFT + DPO alignment