high performance computing on graphics processing units: hgpu.org

hgpu.org » Computer science

KEET: Explaining Performance of GPU Kernels Using LLM Agents

Joshua H. Davis, Klaudiusz Rydzy, Srinivasan Ramesh, Aadit Nilay, Daniel Nichols, Swapna Raj, Nikhil Jain, Abhinav Bhatele

View

Download (PDF)

Tags: Computer science, CUDA, LLM, nVidia, nVidia H100, Performance

May 11, 2026 by hgpu

CuBridge: An LLM-Based Framework for Understanding and Reconstructing High-Performance Attention Kernels

Xing Ma, Yangjie Zhou, Wu Sun, Zihan Liu, Jingwen Leng, Yun Lin, Shixuan Sun, Minyi Guo, Jin Song Dong

View

Download (PDF)

Tags: Computer science, CUDA, Deep learning, LLM, nVidia, nVidia A100, nVidia H100

May 11, 2026 by hgpu

Microbenchmark-Driven Analytical Performance Modeling Across Modern GPU Architectures

Aaron Jarmusch, Sunita Chandrasekaran

View

Download (PDF)

Tags: AMD, AMD Radeon Instinct MI250X, AMD Radeon Instinct MI300A, Benchmarking, Computer science, CUDA, nVidia, nVidia B200

May 11, 2026 by hgpu

DITRON: Distributed Multi-level Tiling Compiler for Parallel Tensor Programs

Size Zheng, Xuegui Zheng, Hanshi Sun, Qi Hou, Wenlei Bao, Shiyu Li, Haojie Duanmu, Jin Fang, Chenli Xue, Chenhui Huang, Yuanqiang Liu, Renze Chen, Ningxin Zheng, Dongyang Wang, Li-Wen Chang, Liqiang Lu, Yun Liang, Jidong Zhai, Xin Liu

View

Download (PDF)

Source codes

Tags: Computer science, LLM, nVidia, nVidia H800, Package, Triton

May 11, 2026 by hgpu

Kerncap: Automated Kernel Extraction and Isolation for AMD GPUs

Cole Ramos, Keith Lowery

View

Download (PDF)

Source codes

Tags: AMD, AMD Radeon Instinct MI210, AMD Radeon Instinct MI300X, AMD Radeon Pro W7900, Computer science, LLM, Package, Python, ROCm, Triton

May 11, 2026 by hgpu

ARGUS: Agentic GPU Optimization Guided by Data-Flow Invariants

Haohui Mai, Xiaoyan Guo, Xiangyun Ding, Daifeng Li, Qiuchu Yu, Chenzhun Guo, Cong Wang, Jiacheng Zhao, Christos Kozyrakis, Binhang Yuan

View

Download (PDF)

Tags: AMD, AMD Radeon Instinct MI300X, Computer science, DSL, LLM, Triton

May 3, 2026 by hgpu

FACT: Compositional Kernel Synthesis with a Three-Stage Agentic Workflow

Sina Heidari, Dimitrios S. Nikolopoulos

View

Download (PDF)

Tags: Computer science, CUBLAS, CUDA, Deep learning, nVidia, nVidia A100

May 3, 2026 by hgpu

Revealing NVIDIA Closed-Source Driver Command Streams for CPU-GPU Runtime Behavior Insight

Yuang Yan, Ian Karlin, Ryan Grant

View

Download (PDF)

Tags: Computer science, CUDA, nVidia, nVidia A40, Performance

May 3, 2026 by hgpu

A Human–Machine Collaborative Tuning Framework for Triton Kernel Optimization on SIMD Platforms

Xulin Zhou, Hongbin Zhang, Mingjie Xing

View

Download (PDF)

Tags: Auto-Tuning, Computer science, Evolutionary Computations, Triton

May 3, 2026 by hgpu

Evaluating CUDA Tile for AI Workloads on Hopper and Blackwell GPUs

Divakar Kumar Yadav, Tian Zhao, Deepak Kumar

View

Download (PDF)

Source codes

Tags: Computer science, CUBLAS, CUDA, LLM, nVidia, nVidia B200, nVidia H100, nVidia RTX PRO 6000, Package, Performance, Triton

May 3, 2026 by hgpu

DVM: Real-Time Kernel Generation for Dynamic AI Models

Jingzhi Fang, Xiong Gao, Renwei Zhang, Zichun Ye, Lei Chen, Jie Zhao, Chengnuo Huang, Hui Xu, Xuefeng Jin

View

Download (PDF)

Source codes

Tags: Code generation, Computer science, LLM, Package

April 13, 2026 by hgpu

Agentic Code Optimization via Compiler-LLM Cooperation

Benjamin Mikek, Danylo Vashchilenko, Bryan Lu, Panpan Xu

View

Download (PDF)

Source codes

Tags: Code generation, Computer science, LLM, Package

April 13, 2026 by hgpu

high performance computing on graphics processing units: hgpu.org

KEET: Explaining Performance of GPU Kernels Using LLM Agents

CuBridge: An LLM-Based Framework for Understanding and Reconstructing High-Performance Attention Kernels

Microbenchmark-Driven Analytical Performance Modeling Across Modern GPU Architectures

DITRON: Distributed Multi-level Tiling Compiler for Parallel Tensor Programs

Kerncap: Automated Kernel Extraction and Isolation for AMD GPUs

ARGUS: Agentic GPU Optimization Guided by Data-Flow Invariants

FACT: Compositional Kernel Synthesis with a Three-Stage Agentic Workflow

Revealing NVIDIA Closed-Source Driver Command Streams for CPU-GPU Runtime Behavior Insight

A Human–Machine Collaborative Tuning Framework for Triton Kernel Optimization on SIMD Platforms

Evaluating CUDA Tile for AI Workloads on Hopper and Blackwell GPUs

DVM: Real-Time Kernel Generation for Dynamic AI Models

Agentic Code Optimization via Compiler-LLM Cooperation

Recent source codes

DITRON: Distributed Compiler based on Triton for Parallel Systems

IntelliKit: Agent-first tooling for AMD hardware

CuTile Benchmark Suite: Performance and Productivity Tradeoffs for GPU Kernel Programming on Blackwell Architecture

MegaTrain: Full Precision Training of 100B+ Parameter Large Language Models on a Single GPU

Device Virtual Machine (DVM)

Agentic Code Optimization via Compiler-LLM Cooperation

AutoKernel: Autoresearch for GPU kernels. Give it any PyTorch model, go to sleep, wake up to optimized Triton kernels

Triton-Sanitizer: A Fast and Device-Agnostic Memory Sanitizer for Triton with Rich Diagnostic Context

LLM.Q: Quantized LLM training in pure CUDA/C++

SOL-ExecBench: Speed-of-Light Benchmarking for Real-World GPU Kernels Against Hardware Limits

Most viewed papers (last 30 days)