ParCoreLab

Snoopie

Multi-GPU communication profiler and visualizer

ComScribe

ComScribe is a tool to identify communication among all GPU-GPU and CPU-GPU pairs in a single-node multi-GPU system.

CPU-Free-model

Source code for the CPU-Free model - a fully autonomous execution model for multi-GPU applications that completely excludes the involvement of the CPU beyond the initial kernel launch.

ReuseTracker

A fast and accurate reuse distance analyzer for multi-threaded applications. It leverages existing hardware features in commodity CPUs.

mixed-and-multi-spmv

Mixed and Multi-Precision SpMV for GPUs with Row-wise Precision Selection.

SpTRSV_Framework

The SpTRSV prediction framework is an automated prediction framework for the fastest sparse triangular solve (SpTRSV) algorithm for a given input sparse matrix on a CPU-GPU platform.

multi-GPU-comm-bench

Split_SpTRSV

The split execution framework can automatically determine the suitability of an SpTRSV for split-execution, find the appropriate split point, and execute SpTRSV in a split fashion using two SpTRSV algorithms while automatically managing any required inter-platform communication. The model is implemented as a C++/CUDA library supporting multiple CPU-GPU algorithms.

BeyondMoore

BeyondMoore has an ambitious goal to develop a software framework that performs static and dynamic optimizations, issues accelerator-initiated data transfers, and reasons about parallel execution strategies that exploit both processor and memory heterogeneity.

ParCoreTools

PES-artifact

aCG

GPU-accelerated linear solvers based on the conjugate gradient (CG) method, supporting NVIDIA and AMD GPUs with GPU-aware MPI, NCCL, RCCL or NVSHMEM

gpu-fusion

GPU fusion code and algorithm

pardnn

.github

Homepage README.

accuracy-verification-microbenchmarks

The microbenchmarks that are used to verify the accuracy of ComDetective.

cha-aware-result-parser

CPU-Free-Model-Compiler

DaCe - Data Centric Parallel Programming

hpctoolkit-externals

HPCToolkit performance tools: essential third party libraries for hpctoolkit

parcorelab.github.io

pes-benchs

AMD_IBS_Toolkit

AMD Research Instruction Based Sampling Toolkit

barnes

gpucommanalyzer

hpctoolkit

HPCToolkit performance tools: measurement and analysis components

snoopie-ucx-tracking-ucx

Modified ucx library to track communications

snoopie-visualiser

splash2

Splash 2 Benchmarks

Uniconn

Uniconn is a unified, portable high-level C++ communication library that supports both point-to-point and collective operations across GPU clusters. Uniconn enables seamless switching between backends and APIs (host or device) with minimal or no changes to application code.