The Khronos Group, an open consortium of leading hardware and software companies creating advanced acceleration standards, today announces a request for industry feedback to gauge interest in developing a new low-level, heterogeneous communications open standard that could transform the way applications are developed for High Performance Embedded Computing (HPEC) systems. If there is industry support, then Khronos would create a working group and invite any interested company to participate for a voice and a vote in creating and evolving a multi-vendor standard under Khronos’ proven multi-company governance process. Those interested in finding out more and providing feedback are invited to visit the Khronos Heterogeneous Communication feedback page.
The traditional programming models and open standards for communications between processing elements have been left behind somewhat. The interconnect topologies have become more complex and fragmented. No longer is there a clear, unified communication standard that is suitable for all the new interfaces introduced by these heterogeneous architectures. Abaco felt no existing open standard API provided a unified solution. Abaco presented their case the Khronos Group proposing a new standard API called Takyon. Khronos has listened and created the “Heterogeneous Communications” Exploratory Group with a goal to gather industry support for development of a new standard and solicit additional proposals. Take a look and contribute here. For a more in-depth description of the problem, be sure to read Abaco's blog "Behind the times".
The new MoltenVK now allows run-time configuration of MoltenVK via environment variables, support for GPU switching was added and is enabled by default, queue family specialization was added but currently disabled by default, synchronous queue submits was added and enabled by default, there is now support for four queue families, support for the VK_AMD_gpu_shader_half_float and VK_KHR_variable_pointers extensions, and updated against the latest SPIR-V Cross. There is also some fixes and other internal improvements. Checkout the Change Log on GitHub and read more about this update on Phoronix.
The Khronos Group is running a survey on Vulkan learning materials and general education. Feedback will be used to guide our future planning in this area as we look to improve our resources. Take this opportunity to share your thoughts with us, it will effect the future of the Vulkan education and the findings will be presented to the Vulkan Working Group.
Illya Rudkin, Principal Engineer Safety-Critical Software, Alastair Murray, Principal Software Engineer and Victor Lomuller, Senior Software Engineer Compilers and SPIR-V, will be attending this year's Khronos F2F event in San Diego from 21st to the 25th January, 2019. Illya, Alastair and Victor would be delighted to speak to anyone attending the event interested in connecting Artificial Intelligence to Silicon. Please track them down via LinkedIn or using their contact form on Codeplay’s website or follow them on Twitter @codeplaysoft.
TensorFlow announced that you will now be able to leverage mobile GPUs for select models with the release of developer preview of the GPU backend for TensorFlow Lite. This new backend leverages. OpenGL ES 3.1 Compute Shaders on Android devices.
The Khronos Safety Critical Advisory Forum (KSCAF) gathers safety critical experts from a wide range of disciplines, such as transportation and medical imaging, who have experience developing software and products to widely adopting standards. The goal of KSCAF is to develop guidelines and recommendations for engineers creating open standard APIs within Khronos, and elsewhere in the industry, so that those standards can help streamline the product safety certification process. The Forum’s chair looks back on a successful 2018, with plans to expand in the new year ahead.
Intel compiler engineer Alexey Bader sent out a public "request for comments" on the idea of adding SYCL to the LLVM Clang compiler stack. He wrote, "We (Intel) would like to request to add SYCL programming model support to LLVM/Clang project to facilitate collaboration on C++ single-source heterogeneous programming for accelerators like GPU, FPGA, DSP, etc. from different hardware and software vendors." Learn more about this initiative or read the original letter.
Fifth-year student Adam Kelly focused on finding the most efficient way possible to simulate quantum computing. Adam's project QCGPU is a high performance, hardware accelerated quantum computer simulator written with Python and OpenCL. News coverage and short interview with Adam are online, as well as the research paper. Congratulations Adam!
The Khronos Group would like to welcome its newest Associate Member DGG (Darmstadt Graphics Group GmbH). DGG develops RapidCompact, a software that reduces the size of 3D data sets in a fully-automatic fashion, optimizing 3D models for different use cases such as VR, AR and 3D on the Web. Being a spinoff of the Fraunhofer Society, DGG embraces innovation as well as standardization efforts, which has made them a strong supporter of Khronos’ glTF format.
The Khronos Group is accepting proposals for an OpenVX project. The project will deliver a fully conformant implementation of the OpenVX 1.2.1 standard that is optimized for the Raspberry Pi 3 Model B+ (or similar) platform. The project will demonstrate the performance advantage of using the OpenVX API by implementing several optimizations that are enabled by OpenVX. Deadline for submissions is January 15, 2019. Complete details here.
Interested in heterogeneous programming for CPUs, GPUs and FPGAs in #OpenCL? Then submit a paper, technical presentation, poster, workshop or tutorial at the annual International Workshop on OpenCL. Deadline Jan 27th
Diligent Engine is a modern cross-platform abstraction layer for Vulkan, OpenGL, OpenGL ES, Direct3D11 and Direct3D12. In the latest release, Diligent Engine added support for Vulkan on MacOS, enabled by MoltenVK.
Yesterday AMD developers did their first AMDVLK open-source push of 2019. That first update in nearly a month updated against the Vulkan 1.1.96 headers, added GPU memory references to software compositing images, clean-ups for the barrier handling, various PAL and LLPC fixes, and other changes. Based upon that source code state from yesterday, an Ubuntu Debian package is now available of just the Vulkan driver and validated for at least 16.04/18.04 installations but should end up working too for e.g. 18.10. Read the entire story on Phoronix.