Khronos Releases OpenCL 2.1 and SPIR-V 1.0 Specifications for Heterogeneous Parallel Programming
Supercomputing 2015– November 16th 2015 – Austin, TX –The Khronos Group, an open consortium of leading hardware and software companies, today announced the ratification and public release of the OpenCL™ 2.1 and SPIR-V™ 1.0 specifications for heterogeneous parallel computation. Consumption of the new SPIR-V cross-API intermediate language is guaranteed in the core OpenCL 2.1 specification. Khronos has released open source utilities and extensions to enable use of SPIR-V in OpenCL 1.2 and 2.0, as well as the upcoming Vulkan™ graphics API, ensuring widespread availability of its powerful runtime capabilities for developers of parallel computation languages and frameworks. The OpenCL C++ kernel language released in the OpenCL 2.1 provisional specification is being finalized and will be released imminently, also using SPIR-V for run-time execution. The OpenCL 2.1 specification is available for immediate download at www.khronos.org/opencl/ and SPIR-V 1.0 is available at http://www.khronos.org/spir/.
SPIR-V 1.0 is an intermediate language fully defined by Khronos with unique, native support for graphics shaders and computational kernels. By providing a well formed compiler target, SPIR-V enables splitting of the compiler chain in graphics and compute intensive environments, so that high-level language and framework front-ends can emit programs to be efficiently executed by Vulkan or OpenCL drivers. Eliminating the need for a built-in high-level language source compiler significantly reduces GPU driver complexity and encourages a diversity of language front-ends. Additionally, a standardized intermediate language provides a measure of kernel IP protection, accelerated kernel load times and enables developers to use a common language front-end, improving kernel reliability and portability across multiple runtime implementations.
“The use of SPIR-V by Vulkan and OpenCL will fundamentally reshape the graphics and compute ecosystem by enabling diverse language and middleware front-ends to leverage the hardware community’s investment in optimized back-end drivers,” said Neil Trevett, president of the Khronos Group and chair of the OpenCL working group and vice president at NVIDIA. “OpenCL 2.1 places the power of SPIR-V into the hands of developers as quickly as possible while we put the finishing touches to the OpenCL C++ kernel language, which we are working to finalize and release as early as mid-2016.”
In parallel with the availability of the OpenCL 2.1 and SPIR-V 1.0 specifications, Khronos has released a number of open source utilities on GitHub to catalyze the use of SPIR-V:
- A bi-directional translator between LLVM to SPIR-V to enable flexible use of both intermediate languages in tool chains;
- An OpenCL C to LLVM compiler that generates SPIR-V through the above translator;
- A SPIR-V assembler and disassembler.
In addition to these Khronos initiatives, there is already significant open source community activity using SPIR-V. Further details on these open source projects can be found here: https://www.khronos.org/spir/resources.
In addition to SPIR-V 1.0 support, OpenCL 2.1 brings enhancements to the OpenCL API, including:
- Subgroups, which enable finer grain control of hardware threading, are now in core, together with additional subgroup query operations for increased flexibility;
- clCloneKernel enables copying of kernel objects and state for safe implementation of copy constructors in wrapper classes;
- Low-latency device timer queries for alignment of profiling data between device and host code.
Industry Support for OpenCL 2.1 and SPIR-V 1.0
“AMD is thrilled to see OpenCL™ include a standard intermediate language as a core component. We believe this opens the door for further innovation in high-level programming languages which can fully leverage the compute power of OpenCL capable devices,” says Greg Stoner, senior director at AMD. “We applaud the efforts of both the compute and graphics communities for defining a standard that can support both - this represents real progress for the industry.”
“We are excited to be part of this move forward with SPIR-V and OpenCL 2.1,” said Andrew Richards, CEO of Codeplay. “This also benefits C++ application developers since the Khronos SYCL™ framework for single source C++ programming will therefore be available on all OpenCL 2.1 enabled platforms.”
“Achieving high compute performance within a limited power budget is essential to the new wave of functions and apps in embedded and mobile systems. Heterogeneous compute is now firmly established as the way to achieve that, and Imagination is delighted that Khronos is directly addressing this objective with the OpenCL 2.1 and SPIR-V 1.0 standards,” said Peter McGuinness, director of multimedia technology marketing, Imagination Technologies. “GPU compute is the bedrock of massively parallel power-efficient acceleration, and these new standards will stimulate the rapid deployment of more functional, compelling apps across a wide range of platforms.”
“Mobica is excited by all of the recent developments from the Khronos Group in graphics and parallel computing. The bringing together of OpenCL 2.1 and Vulkan to use SPIR-V 1.0 as an intermediate language in core is excellent news for the graphics and parallel computing industry,” said Jim Carroll, CTO Mobica.
OpenCL at Supercomputing 2015
There are OpenCL-related presentations and activities at Supercomputing 2015 in Austin on November 15-20th:
OpenCL Booth #285
The OpenCL Booth is the go-to place at SC’15 to engage in technical conversations about OpenCL, SYCL, and SPIR. Visit the booth to receive free OpenCL, SPIR™, and SYCL stickers and OpenCL 2.1 and SYCL 1.2 reference guides.
Tutorial: Portable Programs for Heterogeneous Computing: A Hands-on Introduction
Monday, Nov 16 | 8:30am - 5:00pm | Room 17B
This tutorial will provide lectures and exercises where students can use their own laptops (Windows, Linux or OS/X) to log into a remote OpenCL server. More information and calendar links.
OpenCL BOF: Flocking Together: Experience the Diverse OpenCL Ecosystem
Wednesday, Nov 18 | 5:30pm – 7:00pm | Room 17AB
The strength of OpenCL is in how it was created and maintained by a consortium of like-minded organizations. This session will start with an overview of the newly released OpenCL 2.1 and SPIR-V 1.0 specifications and the SYCL 1.2 abstraction layers. Attendees will have the opportunity to experiment with implementations and tools from multiple vendors, including Altera, AMD, Codeplay Software, Intel, and Xilinx. We invite attendees to bring their code and their toughest questions and join the OpenCL many-core mashup. More information and calendar links.
About The Khronos Group
The Khronos Group is an industry consortium creating open standards to enable the authoring and acceleration of parallel computing, graphics, vision, sensor processing and dynamic media on a wide variety of platforms and devices. Khronos standards include Vulkan™, OpenGL®, OpenGL® ES, WebGL™, OpenCL™, SPIR™, SPIR-V™, SYCL™, WebCL™, OpenVX™, EGL™, COLLADA™, and glTF™. All Khronos members are enabled to contribute to the development of Khronos specifications, are empowered to vote at various stages before public deployment, and are able to accelerate the delivery of their cutting-edge media platforms and applications through early access to specification drafts and conformance tests. More information is available at www.khronos.org.
|
Related News
- Khronos Releases OpenCL 2.1 Provisional Specification for Public Review
- Khronos Releases Vulkan SC 1.0 Open Standard for Safety-Critical Accelerated Graphics and Compute
- Khronos Releases OpenXR 1.0 Specification Establishing a Foundation for the AR and VR Ecosystem
- Khronos Group Releases NNEF 1.0 Standard for Neural Network Exchange
- Khronos Releases Vulkan 1.0 Specification
Breaking News
- Arm loses out in Qualcomm court case, wants a re-trial
- Jury is out in the Arm vs Qualcomm trial
- Ceva Seeks To Exploit Synergies in Portfolio with Nano NPU
- Synopsys Responds to U.K. Competition and Markets Authority's Phase 1 Announcement Regarding Ansys Acquisition
- Alphawave Semi Scales UCIe™ to 64 Gbps Enabling >20 Tbps/mm Bandwidth Density for Die-to-Die Chiplet Connectivity
Most Popular
E-mail This Article | Printer-Friendly Page |