ICL Research Profile

BONSAI

Overview

The goal of the Benchtuning OpeN Software Autotuning Infrastructure (BONSAI) project is to develop a software infrastructure for using scalable, parallel hybrid systems to carry out large, concurrent autotuning sweeps in order to dramatically accelerate the optimization process of computational kernels for GPU accelerators and many-core coprocessors.

Recent developments include a distributed benchmarking engine—already released and currently under active development—capable of scaling to tens of thousands of nodes. This enables benchmarking millions of kernel configurations for different problem sizes and many input datasets while collecting hundreds of performance metrics—including time, energy consumption, cache misses, and memory bandwidth.

Find out more at https://bitbucket.org/icl/bonsai/

Project Handout