POCL 1.0 RC1 Adds Experimental CUDA Backend, Full OpenCL 1.2 Support

Written by Michael Larabel in Free Software on 6 December 2017 at 09:46 AM EST. 3 Comments
One of the most exciting open-source OpenCL projects we have been following in recent years is POCL as "Portable C" for having an LLVM-based portable OpenCL implementation to run on CPUs as well as GPUs now via AMD HSA back-end and a new experimental NVIDIA CUDA back-end. The POCL 1.0 release is finally near.

It's been some months since last having any news to report on with POCL, but it's exciting to see this morning their 1.0 release candidate was just issued. POCL 1.0 contains some significant changes and improvements.

First and foremost, POCL 1.0 is the first release to contain their new experimental NVIDIA CUDA back-end for running their OpenCL implementation off CUDA with NVIDIA graphics processors.

POCL 1.0 RC1 also features automatic local work-group sizing on kernel enqueue for taking into account SIMD width for vectorization and other factors. POCL 1.0 also has LLVM 5.0 back-end support, improved scalability in the CPU driver, full conformance for the OpenCL 1.2 standard by default, an updated format of POCL binaries, improved AVX-512 support, and much more.

POCL 1.0 RC1 can be downloaded via GitHub. Test results and build guide via their Wiki. Those unfamiliar with the Portable Computing Language can learn more at PortableCL.org.
Related News
About The Author
Author picture

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter, LinkedIn, or contacted via MichaelLarabel.com.

Popular News This Week