Kernel Tuner

Kernel Tuner greatly simplifies the development of highly-optimized and auto-tuned CUDA, OpenCL, and C code, supporting many advanced use-cases and optimization strategies that speed up the auto-tuning process.

30
mentions
13
contributors

Cite this software

DOI:

10.5281/zenodo.1220113

What Kernel Tuner can do for you

  • Allows developers to easily unit test and auto-tune GPU code
  • Generic auto-tuning of user-defined parameters for CUDA, OpenCL, and C kernels
  • Supports more than 20 different search optimization methods to speedup tuning
  • Successfully used in 10+ different eScience projects, across various disciplines

Kernel Tuner simplifies the development of efficient GPU programs, or kernels. It does so by making kernels written in C/C++, OpenCL, or CUDA accessible from Python, while taking care of the required synchronization between data kept in host memory and data kept in device memory.

This has a number of advantages. First, it simplifies auto-tuning of the kernel parameters. In fact, Kernel Tuner comes standard with a variety of strategies for efficiently searching the parameter space, leading to greatly improved performance of tuned kernels. Second, it allows for unit testing of GPU code from within Python.

Kernel Tuner does not add any additional dependencies to the kernel code, and does not require extensive code changes. Furthermore, it is noteworthy that kernels tuned by Kernel Tuner do not require any changes after tuning to make them production ready--tuned kernels can be used as-is from any host programming language.

Keywords
  • Big data
  • GPU
  • High performance computing
  • Multi-scale & multi model simulations
  • Optimized data handling
  • Real time data analysis
Programming language
  • Python 99%
  • Cuda 1%
License
  • Apache-2.0
</>Source code

Participating organisations

ASTRON
CWI
Netherlands eScience Center

Mentions

Kernel Tuner tutorial at Supercomputing 2021

Author(s): Ben van Werkhoven
Published in 2021

Writing Testable GPU Code

Author(s): Ben van Werkhoven
Published in 2018

Testimonials

With Kernel Tuner, we were able to accelerate our CUDA kernels by a factor of 10 in just a few weeks
Chiel van Heerwaarden, Wageningen University & Research

Contributors

Contact person

Ben van Werkhoven

Ben van Werkhoven

Netherlands eScience Center
Mail Ben
AS
Alessio Sclocco
Netherlands eScience Center
Ben van Werkhoven
Ben van Werkhoven
Felipe Zapata
Felipe Zapata
Netherlands eScience Center
Floris-Jan Willemsen
Floris-Jan Willemsen
Netherlands eScience Center
Inti Pelupessy
Inti Pelupessy
Netherlands eScience Center
Jisk Attema
Jisk Attema
Johannes Hidding
Johannes Hidding
Netherlands eScience Center
Nicolas Renaud
Nicolas Renaud
Netherlands eScience Center
Patrick Bos
Patrick Bos
Netherlands eScience Center
RS
Richard Schoonhoven
CWI
Stijn Heldens
Stijn Heldens
Netherlands eScience Center
WP
Willem Jan Palenstijn
CWI

Related projects

RECRUIT

Reducing Energy Consumption in Radio-astronomical and Ultrasound Imaging Tools

Updated 1 month ago
Running

CORTEX

Self-learning machines hunt for explosions in the universe and speed up innovations in industry and...

Updated 2 months ago
Running

CHEOPS

Verified construction of correct and optimised parallel software

Updated 2 months ago
Running

ESiWACE2

For future exascale climate and weather predictions

Updated 2 months ago
Running

Retina COVID19

Social distancing measures carry economic and social costs

Updated 2 months ago
Finished

A methodology and ecosystem for many-core programming

Boosting the performance of current and future programs

Updated 2 months ago
Finished

DIRAC

Distributed radio astronomical computing

Updated 2 months ago
Finished

Triple-A 2

Accelerating astronomical applications 2

Updated 1 month ago
Finished

Parallelisation of multi point-cloud registration

Studying subcellular structures and functions

Updated 2 months ago
Finished

3D Geospatial Data Exploration for Modern Risk Management Systems

The country below sea level

Updated 2 months ago
Finished

Real-time detection of neutrinos from the distant Universe

Observing processes that are inaccessible to optical telescopes

Updated 2 months ago
Finished

A Jungle Computing Approach to Large-Scale Online Forensic Analysis

Programming tools that simplify application development and deployment

Updated 2 months ago
Finished

Related tools

AMBER

AM

A real-time pipeline to search for Fast Radio Bursts and other transient radio sources.

Updated 2 months ago
12 mentions, 3 contributors