high performance computing on graphics processing units: hgpu.org

Holger R. Roth, Daniel J. Beutel, Yan Cheng, Javier Fernandez Marques, Heng Pan, Chester Chen, Zhihong Zhang, Yuhong Wen, Sean Yang, Isaac (Te-Chung)Yang, Yuan-Ting Hsieh, Ziyue Xu, Daguang Xu, Nicholas D. Lane, Andrew Feng

View

Download (PDF)

Source codes

Tags: AI, Algorithms, Computer science, nVidia, Package, Security

July 7, 2024 by hgpu

Chat AI: A Seamless Slurm-Native Solution for HPC-Based Services

Ali Doosthosseini, Jonathan Decker, Hendrik Nolte, Julian M. Kunkel

View

Download (PDF)

Source codes

Tags: AI, Cloud, Computer science, HPC, LLM, nVidia, nVidia H100, Package, PC cluster

July 7, 2024 by hgpu

Towards Unified Analysis of GPU Consistency

Haining Tong, Natalia Gavrilenko, Hernán Ponce de León, Keijo Heljanko

View

Download (PDF)

Source codes

Tags: Computer science, Memory, nVidia, OpenCL, Package, PTX, Vulkan

July 7, 2024 by hgpu

Automatic Code Rewriting for Performance Portability

Alister Johnson

View

Download (PDF)

Tags: Compilers, Computer science, HIP, OpenCL, OpenMP, performance portability, SYCL, Thesis

July 7, 2024 by hgpu

Composing Distributed Computations Through Task and Kernel Fusion

Rohan Yadav, Shiv Sundram, Wonchan Lee, Michael Garland, Michael Bauer, Alex Aiken, Fredrik Kjolstad

View

Download (PDF)

Tags: Computer science, CUDA, Distributed computing, MPI, nVidia, nVidia DGX-A100, Python

June 30, 2024 by hgpu

Over-synchronization in GPU Programs

profile_util: simple C++ library to report memory, timing, thread affinity

Profile Util library: A quick and easy way to get MPI, OpenMP and GPU runtime information

MaC: Mamba CPU performance predictor

MambaCPU: Enhanced Correlation Mining with State Space Models for CPU Performance Prediction

Mixed-precision finite element kernels

Mixed-precision finite element kernels and assembly: Rounding error analysis and hardware acceleration

SHMT: The Simultaneous and Heterogenous Multithreading project

General-Purpose Computing on Tensor Processors

Superpipeline: A Universal Approach for Reducing GPU Memory Usage in Large Models

EnergyUCB-Bandit

Online Energy Optimization in GPUs: A Multi-Armed Bandit Approach

See all packages

* * *

high performance computing on graphics processing units: hgpu.org

Applications

Bringing Auto-tuning to HIP: Analysis of Tuning Impact and Difficulty on AMD and Nvidia GPUs

Efficiently Training 7B LLM with 1 Million Sequence Length on 8 GPUs

Optimization of Large-Scale Sparse Matrix-Vector Multiplication on Multi-GPU Systems

Harnessing Integrated CPU-GPU System Memory for HPC: a first look into Grace Hopper

Automating Heterogeneous Parallelism in Numerical Differential Equations

The Impact of Modern Consumer GPUs on Commonly Used Secure Password Standards

Automated C/C++ Program Repair for High-Level Synthesis via Large Language Models

Supercharging Federated Learning with Flower and NVIDIA FLARE

Towards Unified Analysis of GPU Consistency

Automatic Code Rewriting for Performance Portability

Composing Distributed Computations Through Task and Kernel Fusion

Recent source codes

Kokkidio: Kokkos+Eigen - write fast, readable multi-platform code

LLaMA-Inference-Bench

MATAR

Over-synchronization in GPU Programs

profile_util: simple C++ library to report memory, timing, thread affinity

MaC: Mamba CPU performance predictor

Mixed-precision finite element kernels

SHMT: The Simultaneous and Heterogenous Multithreading project

Superpipeline: A Universal Approach for Reducing GPU Memory Usage in Large Models

EnergyUCB-Bandit

Most viewed papers (last 30 days)