high performance computing on graphics processing units: hgpu.org

hgpu.org » nVidia GeForce RTX 5090

AgentServe: Algorithm-System Co-Design for Efficient Agentic AI Serving on a Consumer-Grade GPU

Yuning Zhang, Yan Yan, Nan Yang, Dong Yuan

View

Download (PDF)

Tags: Computer science, CUDA, LLM, nVidia, nVidia GeForce RTX 5090, nVidia RTX 5000 Ada

March 15, 2026 by hgpu

Diagnosing FP4 inference: a layer-wise and block-wise sensitivity analysis of NVFP4 and MXFP4

Musa Cim, Burak Topcu, Mahmut Taylan Kandemir

View

Download (PDF)

Tags: AMD, Computer science, LLM, nVidia, nVidia GeForce RTX 5090, nVidia RTX PRO 6000, Precision

March 15, 2026 by hgpu

Private LLM Inference on Consumer Blackwell GPUs: A Practical Guide for Cost-Effective Local Deployment in SMEs

Jonathan Knoop, Hendrik Holtmann

View

Download (PDF)

Source codes

Tags: Cloud, Computer science, CUDA, LLM, nVidia, nVidia GeForce RTX 5060 Ti, nVidia GeForce RTX 5070 Ti, nVidia GeForce RTX 5090, Package

February 2, 2026 by hgpu

Home-made Diffusion Model from Scratch to Hatch

Shih-Ying Yeh

View

Download (PDF)

Source codes

Tags: Computer science, Computer vision, nVidia, nVidia GeForce RTX 5090, Package, Python, PyTorch

September 14, 2025 by hgpu

KernelSkill: A Multi-Agent Framework for GPU Kernel Optimization

MSKernelBench & CUDAMaster

Making LLMs Optimize Multi-Scenario CUDA Kernels Like Experts

EvoScientist: Harness Vibe Research with Self-evolving AI Scientists

EvoScientist: Towards Multi-Agent Evolving AI Scientists for End-to-End Scientific Discovery

RepoLaunch: Automating Build and Test Pipeline of Code Repositories on ANY Language and ANY Platform

RepoLaunch: Automating Build & Test Pipeline of Code Repositories on ANY Language and ANY Platform

RepoLaunch: Automating Build and Test Pipeline of Code Repositories on ANY Language and ANY Platform

RepoLaunch: Automating Build & Test Pipeline of Code Repositories on ANY Language and ANY Platform

CONCUR: a benchmark designed to evaluate multithreaded Java code generated by LLMs

CONCUR: Benchmarking LLMs for Concurrent Code Generation

HIPRT: Ray Tracing using HIP

Ray Tracing using HIP

MXFP4 Training Support Codebase

Practical FP4 Training for Large-Scale MoE Models on Hopper GPUs

CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation

CUDABench: Benchmarking LLMs for Text-to-CUDA Generation

See all packages

* * *

high performance computing on graphics processing units: hgpu.org

AgentServe: Algorithm-System Co-Design for Efficient Agentic AI Serving on a Consumer-Grade GPU

Diagnosing FP4 inference: a layer-wise and block-wise sensitivity analysis of NVFP4 and MXFP4

Home-made Diffusion Model from Scratch to Hatch

Recent source codes

KernelSkill: A Multi-Agent Framework for GPU Kernel Optimization

MSKernelBench & CUDAMaster

EvoScientist: Harness Vibe Research with Self-evolving AI Scientists

RepoLaunch: Automating Build and Test Pipeline of Code Repositories on ANY Language and ANY Platform

RepoLaunch: Automating Build and Test Pipeline of Code Repositories on ANY Language and ANY Platform

CONCUR: a benchmark designed to evaluate multithreaded Java code generated by LLMs

HIPRT: Ray Tracing using HIP

MXFP4 Training Support Codebase

CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation

CUDABench: Benchmarking LLMs for Text-to-CUDA Generation

Most viewed papers (last 30 days)