The SYCL Compiler and Runtimes 2019-09 release allow OpenCL offloading to accelerators (GPU/FPGA). Some OpenCLCL/SYCL FPGA extensions are now supported along with support for dumping the SYCL task graph to JSON. Lots of other improvements and fixes are included on the GitHub release page.
LLVM Clang 9.0 has been released and is now available for download. This is the first release to contain experimental support of C++ for OpenCL language mode in Clang. More details can be found in the Clang documentation. This new support will be discussed at the LLVM Developers meeting (October 2019) at the From C++ for OpenCL to C++ for accelerator devices talk by Khronos Member Anastasia Stulova.
Longtime Nouveau developer Karol Herbst has been leading the work on the Nouveau NIR/SPIR-V changes around OpenCL support since joining Red Hat almost two years ago. Arriving in Mesa recently is the SPIR-V support for Nouveau’s NVC0 Gallium3D driver.
In C++, especially in modern C++, function pointers are a legacy feature from the C language but they still exist in some code bases. SYCL does not provide support for function pointers as this is a limitation posed by the design of OpenCL v1.2 which is the basis of the current SYCL v1.2.1 definition. The good news is that we can use modern C++ to implement a solution that can be used with SYCL. Learn how to do this with examples from Codeplay.
Andrew Richards from Codeplay will be presenting “Using Industry-Standard Techniques to Accelerate AI Software” at this years Linley Fall Processor Conference in Santa Clara. You can learn more about this presentation and download a free white paper by Linley Gwennap, Principal Analyst at the Linley Group.
Heterogeneous-Compute Interface for Portability (HIP) is a runtime API and a conversion tool to help make CUDA programs more portable. It was originally contributed by AMD to the open source community with the intention to ease the effort of making CUDA applications also work on AMD’s ROCm platform.
While AMD and NVIDIA share the vast majority of the discrete GPU market, it is useful to make this “CUDA portability enhancement route” available to an even wider set of platforms. Since the Khronos OpenCL standard remains the most widely adopted cross-platform heterogeneous programming API/middleware, it is interesting to study whether HIP could be ported on top of it, expanding its scope potentially to all OpenCL supported devices. We in Customized Parallel Computing group, Tampere University, Finland, are happy to announce that to have worked on such a tool, known as HIPCL, for some time and it’s now published and available in Github.
The first release of HIPCL is a proof-of-concept, but is already useful for end-users. It can run most of the CUDA examples in the HIP repository and the list of supported CUDA applications will grow steadily as we add new features.
Andrew Richards, CEO and Co-founder of Codeplay Software, presents the “Can We Have Both Safety and Performance in AI for Autonomous Vehicles?” tutorial at the May 2019 Embedded Vision Summit. Andrews presentation includes discussion on how SYCL, OpenCL and Vulkan can play into Safety and Performance in vehicles.
The OpenCL working group has posted new OpenCL Reference pages. These OpenCL reference pages are now generated from open source AsciiDoctor – please let us know if you have fixes or suggestions on GitHub.
All of the presentations and videos from the Khronos OpenVX workshop at the 2019 Embedded Vision Summit are now online. If you were unable to attend this workshop, you may now watch the seven sessions online and follow along with the slide presentations:
- Introduction and OpenCL Overview & Update – Neil Trevett, NVIDIA: slides, video
- OpenCL & SYCL – Andrew Richards, Codeplay: slides, video
- Intel Open Source SYCL Compiler Project – Konstantin S. Bobrovsky, Intel: slides, video
- OpenVX Presentations – Frank Brill, Cadence / Niclas Danielsson & Mikael Pendse, Axis : here & here, video
- Inference with OpenVX – Mike Schmit, AMD: slides, video
- NNEF Presentation – Gergely Debreczeni, AImotive: slides, video
- OpenVX Hands-On - Part 1 – Rajy Rawther & Kiriti Nagesh Gowda, AMD: slides, video
Intel has released their 184.108.40.20661 graphics drivers. OpenGL has been bumped to 4.6, which adds an extension to use Vulkan’s SPIR-V shaders in OpenGL. Vulkan has been given four new extensions: VK_EXT_depth_clip_enable, VK_EXT_host_query_reset, VK_EXT_scalar_block_layout, and VK_KHR_shader_float16_int8. The OpenCL kernel compile times have been reduced.
Imagination Technologies announced that it has joined SiFive’s DesignShare ecosystem, giving system designers easy access to its industry-leading PowerVR GPU and neural network accelerator (NNA) IP cores. The PowerVR GPU will be the first fully featured GPU supporting the Vulkan applications programming interface (API) available via the DesignShare ecosystem. Mohit Gupta, vice president, SoC IP solutions, strategy and business development, SiFive; said: “We’re delighted to welcome Imagination into the DesignShare movement. Imagination is renowned for its industry-leading GPU IP and award-winning NNA IP. PowerVR’s support of APIs such as Vulkan, OpenCL, Android NNAPI, and SYCL will deliver new opportunities for customer growth for the ecosystem. Together, we hope to inspire customers to challenge the norm and develop groundbreaking custom silicon. We’re excited to see the creativity that our collaboration inspires.”
The recently formed Khronos OpenCL Tooling Subgroup has been focused on developing and enhancing open source tools and components, targeted at embedded systems and heterogeneous computation applications; the new tools and resources are available to the entire OpenCL ecosystem.
Michael Wong, VP of Research and Development, Charles Macfarlane, VP of Marketing and Rod Burns, Developer Relations Manager are attending this year’s AutoSens Conference in Detroit from 14th to 16th May, 2019. Michael Wong will be presenting “Ensuring safe AI in a car” highlighting the practical engineering challenges of turning deep learning, classical machine vision and sensor fusion algorithms from research prototypes into real-world automotive-grade systems. Codeplay will be exhibiting the benefits of open standards OpenCL and SYCL on Renesas R-Car V3H platform for rapid development and deployment of ADAS and AV functions.
LLVM 2019 wrapped up a couple of weeks ago, and some of the presentations are now rolling out. For a list of Khronos related session, visit our event page. For a complete list of presentations, please visit the LLVM website.
The European Space Agency have a GitLab project that ray traces an emulated black hole event horizon using Python and OpenCL (aptly named “PyHole”); they have included a WebGL visualization tool for the result.