high performance computing on graphics processing units: hgpu.org

Applications

hgpu.org » Applications

Stencil Computation at the Intersection of AI and HPC

Timothee Ewart, Mauricio Araya-Polo

View

Tags: AI, AMD, AMD Radeon Instinct MI325X, Computer science, HPC, Intel Arc B580, Intel B70, Intel Data Center GPU Max 1550, nVidia, nVidia A100, nVidia H100, nVidia RTX PRO 6000, performance portability, Stencil computation, SYCL, Triton

September 14, 2026 by hgpu

Hardware-Aware FP4 FlashAttention-4

Robert Hu

View

Download (PDF)

Source codes

Tags: Computer science, CUDA, nVidia, nVidia B300, nVidia GB200, Package, PTX

September 14, 2026 by hgpu

Every Kernel Is a Join: Automatic Multi-GPU Parallelism for AI Computations in Einsummable

Zhimin Ding, Chen-Kuan Liao, Chima Adiole, Brianna Barrow, Fangzhou Du, Yu Hsiao, Ge Huang, Yicheng Jin, Ismail Syed, Chris Jermaine

View

Download (PDF)

Source codes

Tags: AI, Computer science, CUDA, LLM, nVidia, nVidia DGX-A100, Package, Triton

September 14, 2026 by hgpu

MaxKernel: Agentic Kernel Generation for TPUs

Shangkun Wang, Nina Cai, Charles Hoong, Julian Walker, Gerson Kroiz, George Vanica, Deepak Patil, Andi Gavrilescu, Hassan Sipra, Sethu Sankaran

View

Download (PDF)

Source codes

Tags: Computer science, Package, TPU

September 14, 2026 by hgpu

Taming Bitwise Behavior in GPU Kernels with Tensor Core: Black-Box Reconstruction, Compiler Enforcement, and Static Verification

Ziteng Yang, Nicholas J. Riasanovsky, Warren Deng, Vivek Sarkar

View

Download (PDF)

Tags: Computer science, CUBLAS, CUDA, nVidia, nVidia GB300, nVidia H100, PTX, Triton

September 14, 2026 by hgpu

An HPC Approach to Accelerate Tensor Decompositions

Markus Hellgren, Erna Begovic Kovac, Hans O. Karlsson, Roman Iakymchuk

View

Download (PDF)

Tags: Computer science, CUDA, HPC, nVidia, nVidia H100

August 30, 2026 by hgpu

Concurrency Response of Plain Global Loads on the NVIDIA H100

Somashekar Manjunath, Rahul Ramachandra M

View

Download (PDF)

Tags: Computer science, CUDA, Hardware Architecture, nVidia, nVidia H100, Performance

August 30, 2026 by hgpu

Benchmarking Confidential Computing Performance on NVIDIA Blackwell GPUs

Daniyal Khan, Amean Asad, Ansgar Grunseid

View

Download (PDF)

Tags: Benchmarking, Computer science, CUDA, nVidia, nVidia B200, Security

August 30, 2026 by hgpu

PTXBench: Benchmark and Adapt LLMs for GPU Kernel Optimization with Architecture-specific PTX

Genghan Zhang, Yixin Dong, Chengze Fan, Zhichen Zeng, Yueming Yuan, Shaowei Zhu, Kunle Olukotun

View

Download (PDF)

Source codes

Tags: Benchmarking, Computer science, LLM, nVidia, nVidia B200, nVidia H100, Package, PTX

August 30, 2026 by hgpu

RealisticTritonBench: A Benchmark for Triton-Kernel Generation in Real-World AI Frameworks

Jinjun Huang, Zhongzhen Wen, Tongtong Xu, Meng Yan, Xin Xia, Zhongxin Liu

View

Download (PDF)

Source codes

Tags: Benchmarking, Computer science, CUDA, LLM, nVidia, nVidia GeForce RTX 3090, Package, Triton

August 30, 2026 by hgpu

A Contract-Grade Verifier for LLM-Generated GPU Kernels, and a Native Blackwell Backward for the Gated-Linear-Recurrence Family

Rishi Shah, Rishav Shrestha

View

Download (PDF)

Tags: Computer science, LLM, nVidia, nVidia B200, nVidia GB200, Triton

August 17, 2026 by hgpu

Hand-Written PTX Tensor-Core GEMM Kernels: A Multi-Precision Study on NVIDIA L4

Matt J. Borowski, Blazej Osinski

View

Download (PDF)

Source codes

Tags: Computer science, nVidia, nVidia L4, Package, PTX

August 17, 2026 by hgpu

PTXBench: an open-source environment for kernel agents with architecture-specific CUDA/PTX

PTXBench: Benchmark and Adapt LLMs for GPU Kernel Optimization with Architecture-specific PTX

RealisticTritonBench: A Benchmark for Triton-Kernel Generation in Real-World AI Frameworks

TensorCorePTX: PTX Tensor Core GEMM Exploration

Hand-Written PTX Tensor-Core GEMM Kernels: A Multi-Precision Study on NVIDIA L4

FlashPDE: Triton-fused PDE stencil kernels for physics-informed neural networks

FlashPDE: A Drop-In Fused Triton Operator Library for Neural PDE Solvers

PortLBM: Portable GPU-accelerated Lattice Boltzmann Method Simulations

PortLBM: A Portable Lattice Boltzmann Tool Leveraging SYCL on AMD, NVIDIA, and Intel GPUs

Harness Engineering for LLM-Driven GPU Kernel Generation

CANNBot Skills

Compiler-Grounded Hierarchical Diagnosis for LLM-Based Triton Kernel Optimization

See all packages

* * *

high performance computing on graphics processing units: hgpu.org

Applications

Stencil Computation at the Intersection of AI and HPC

Hardware-Aware FP4 FlashAttention-4

Every Kernel Is a Join: Automatic Multi-GPU Parallelism for AI Computations in Einsummable

MaxKernel: Agentic Kernel Generation for TPUs

Taming Bitwise Behavior in GPU Kernels with Tensor Core: Black-Box Reconstruction, Compiler Enforcement, and Static Verification

An HPC Approach to Accelerate Tensor Decompositions

Concurrency Response of Plain Global Loads on the NVIDIA H100

Benchmarking Confidential Computing Performance on NVIDIA Blackwell GPUs

PTXBench: Benchmark and Adapt LLMs for GPU Kernel Optimization with Architecture-specific PTX

RealisticTritonBench: A Benchmark for Triton-Kernel Generation in Real-World AI Frameworks

A Contract-Grade Verifier for LLM-Generated GPU Kernels, and a Native Blackwell Backward for the Gated-Linear-Recurrence Family

Hand-Written PTX Tensor-Core GEMM Kernels: A Multi-Precision Study on NVIDIA L4

Recent source codes

FP4 FlashAttention forward, backward, and TorchTitan reproduction code

EinsummableV3 - graph-based computation with tensor operations support

MaxKernel: Intelligent TPU Kernel Engineering Agents

PTXBench: an open-source environment for kernel agents with architecture-specific CUDA/PTX

RealisticTritonBench: A Benchmark for Triton-Kernel Generation in Real-World AI Frameworks

TensorCorePTX: PTX Tensor Core GEMM Exploration

FlashPDE: Triton-fused PDE stencil kernels for physics-informed neural networks

PortLBM: Portable GPU-accelerated Lattice Boltzmann Method Simulations

Harness Engineering for LLM-Driven GPU Kernel Generation

CANNBot Skills

Most viewed papers (last 30 days)