Early Access to the SYCL Open Standard for C++ Acceleration
A new generation of technology is coming for low power devices with embedded intelligence. At Codeplay, we believe that the full performance of today's highly parallel devices should not be vendor-locked to one platform but should be available to all programmers using open standards.
To help steer this evolution, Codeplay is giving developers free, early access to ComputeCpp™ with a pre-conformance beta implementation of the SYCL™ open standard, along with an open-source preview of the latest Parallel Technical Specification to be adopted into C++17. Other open-source projects being made available are VisionCpp, a machine vision library demonstrating C++ techniques for performance-portability, and an early version of the Eigen C++ library that uses SYCL for acceleration on OpenCL devices. Eigen is used in projects like TensorFlow.
The SYCL standard provides acceleration using OpenCL™ devices meaning you can accelerate your C++ software on a wide range of platforms into many market segments such as automotive ADAS, cloud compute, IoT vision and cellular. This leads to better performance portability at the higher layer for Machine Vision, Neural Networks, and Deep Learning.
The SYCL specification from the Khronos™ Group is for developers who want to take software written using C++ single-source programming models such as CUDA® or C++AMP, and port to a wide range of OpenCL™devices. The current early-access ComputeCpp Community Edition Beta release provides pre-conformance SYCL v1.2 support for AMD® and Intel® OpenCL GPUs and CPUs. Further operating system and device support is on its way. More information on the specification from Khronos is available here: https://www.khronos.org/sycl
“SYCL is designed to not lock in your software development to one platform” said Andrew Richards, CEO of Codeplay. “Our ComputeCpp Community Edition Beta will let developers work on acceleration of open-source C++ software such as TensorFlow, Eigen and the C++ 17 Parallel STL.”
"SYCL is one of the strongest basis from which the ISO C++ Standard will adapt future C++ support for heterogeneous computing, and it is one of the few candidates that is designed from the base up to work with C++ templates and abstractions for low power and low-latency massive parallel dispatch" Michael Wong, VP R&D at Codeplay, Chair of C++ Standard SG14 on Low Latency and SG5 on Transactional Memory, ISOCPP VP and Director
To further stimulate the SYCL community, Codeplay has also setup http://sycl.tech where anyone can post projects, videos, news or tutorials. This site is for everything related to SYCL whether it uses ComputeCpp or not.
Codeplay is presenting and available for discussions this week at AutoSens (http://auto-sens.com/) and CppCon (http://cppcon.org/) and will also be at the British Machine Vision Conference 2016, in York, UK. To download ComputeCpp Community Edition Beta or contact Codeplay, visit https://codeplay.com/
Khronos, SPIR and SYCL are trademarks of the Khronos Group Inc. OpenCL is a trademark of Apple Inc. used under license by Khronos. CUDA is a trademark of NVIDIA Corporation. Intel is a trademark of Intel Corporation. AMD is a trademark of Advanced Micro Devices, Inc.