Intel Executing toward XPU Vision with oneAPI and Intel Server GPU

Intel Executing toward XPU Vision with oneAPI and Intel Server GPU

What’s New: Intel right now declared key milestones in its multiyear journey to provide a blend of architectures with a unified software experience. The firm declared the gold launch of Intel® oneAPI toolkits coming in December, and new capabilities in its software package stack as element of the Intel’s put together hardware and software package design and style approach. Intel also debuted its first discrete graphics processing unit (GPU) for the knowledge centre, Intel® Server GPU, based mostly on the Xe-LP microarchitecture and built especially for large-density, lower-latency Android cloud gaming and media streaming.

“Today is a vital instant in our bold oneAPI and XPU journey. With the gold release of our oneAPI toolkits, we have prolonged the developer encounter from familiar CPU programming libraries and instruments to include our vector-matrix-spatial architectures. We are also launching our to start with details centre GPU centered on Xe-LP microarchitecture targeted on the rapid-rising cloud gaming and media streaming segments.”
–Raja Koduri, Intel senior vice president, chief architect and general manager of Architecture, Graphics and Software program

Why It Matters: As the planet moves into an era of billions of intelligent gadgets and an exponential progress of knowledge, CPUs on your own have to have a change in concentration to a combine of architectures throughout CPUs, GPUs, FPGAs and other accelerators. Intel collectively describes this as “XPU” eyesight. The launch of the Intel Server GPU is a further move to extend Intel’s supplying in the XPU period.

This era of computing also calls for a in depth software package stack. Developers will be capable to accessibility a typical, open and expectations-based programming model across Intel® XPUs with Intel’s oneAPI toolkits. They supply on the general performance prospective of the fundamental components and decrease computer software development and routine maintenance prices even though cutting down dangers affiliated with deploying accelerated computing relative to proprietary, seller-certain remedies.

About the Gold Release of Intel oneAPI Toolkits: First introduced at SuperComputing 2019, the oneAPI market initiative is a daring eyesight for a unified and simplified cross-architecture programming model that delivers uncompromised effectiveness without having proprietary lock-in though enabling the integration of legacy code. With oneAPI, developers can choose the finest architecture for the certain issue they are attempting to clear up without the need of needing to rewrite program for the future architecture and system.

Intel oneAPI toolkits just take entire benefit of cutting-edge hardware abilities and instructions this sort of as Intel® AVX-512 and Intel® DL Raise on CPUs, together with features exceptional to XPUs. Designed on very long-standing and tested Intel developer resources, Intel oneAPI toolkits deliver familiar languages and criteria even though providing whole continuity with present code.

Currently Intel announced that the gold launch of Intel oneAPI toolkits will be transport in December, out there for absolutely free, regionally and in the Intel® DevCloud, alongside with industrial versions that incorporate worldwide help from Intel specialized consulting engineers. Intel will quickly transition Intel® Parallel Studio XE and Intel® Program Studio device suites to its oneAPI goods.

In addition, the Intel DevCloud, wherever builders can test code and workloads on a variety of Intel architectures, is expanding to incorporate new Intel® Iris® Xe graphics components. Intel Iris Xe MAX graphics is now available for general public accessibility and Intel Xe-HP is accessible to find developers.

oneAPI has received industry aid, which include modern endorsements from Microsoft Azure and TensorFlow top investigate companies, companies and universities also endorse oneAPI.

In addition, the College of Illinois Beckman Institute for Sophisticated Science and Know-how today announced the establishment of a new oneAPI center of excellence (CoE). It is working with the oneAPI programming model to lengthen the daily life sciences software nanoscale molecular dynamics (NAMD) to additional computing environments. NAMD, which simulate large biomolecular units, is assisting to tackle authentic-planet problems such as COVID-19. This CoE joins others at the University of Stockholm (SeRC) who are focusing on GROMACS, and Heidelberg University’s (URZ), the place they’re focusing on offering oneAPI help to GPUs from other sellers.

Intel toolkit oneAPI rendering

» Obtain all pictures (ZIP, 7 MB)

About the New Intel Server GPU: With its very first discrete GPU for the data heart, Intel is further expanding its loaded suite of platform innovations for improved cloud gaming and media encounters. The mix of Intel® Xeon® Scalable processors, open up-sourced and accredited software program substances and the new Intel Server GPU offers a large-density, reduced-latency resolution at reduce whole price tag of ownership (TCO) for Android cloud gaming and substantial-density media transcode and encode for genuine-time about-the-major movie streamingone.

The Intel Server GPU is dependent on Xe-LP microarchitecture, Intel’s most strength-efficient graphics architecture, offering a minimal-ability, discrete method-on-chip structure, with a 128-little bit pipeline and 8GB of committed onboard small-energy DDR4 memory.

By using Intel Server GPUs together with Intel Xeon Scalable processors, services vendors can scale graphics ability independently from the server depend, supporting more substantial quantities of streams and subscribers per process, when nonetheless obtaining minimal TCO. Packaging together 4 Intel Server GPUs in H3C’s 3-quarter-duration, full-top x16 PCIe Gen 3. add-in card can support a lot more than one hundred simultaneous Android cloud gaming buyers in a normal two-card technique and up to a hundred and sixty simultaneous consumers, depending on the particular activity title and server configuration.two Developers can leverage the common API in today’s Media SDK, which will be integrated into the oneAPI Movie Processing Library following yr. Intel is doing work with various computer software and services partners, which include Gamestream, Tencent Game titles and Ubitus, to convey the Intel Server GPU to industry.

“Intel is an crucial collaborator on our Android Cloud Gaming option. Intel Xeon Scalable processors and Intel Server GPUs provide a substantial-density, lower-latency, small-electricity, small-TCO resolution. We are capable to produce in excess of a hundred recreation circumstances for every two-card server for our most common games, King of Glory and Arena of Valor,” claimed Allen Fang, vice standard supervisor of Tencent XianYou Cloud Gaming System.

The Xe-LP-dependent Intel Server GPU is now delivery. Together with the a short while ago launched Intel Iris® Xe MAX graphics, the GPU will make improvements to the visual computing knowledge for end users all around the planet as Intel ramps its Xe architecture solutions and software initiatives.

What’s New with Intel® Graphics Program: One of Intel’s core procedures for achievements in scaling GPUs from entry-stage graphics to higher effectiveness computing (HPC) is the implementation of a single code foundation. Toward this eyesight, Intel’s computer software stack now supports many generations of graphics, including recently launched 11th Gen Intel® Core™ processors with integrated Intel Iris Xe graphics and Intel Iris Xe MAX discrete graphics. Growing the code base to support facts centre goods in which Linux is more widespread is a essential up coming move in its scalable Xe architecture strategy. Intel optimized the Linux driver to target on code reuse amongst operating programs, it has amplified aim on Linux 3D functionality and it now provides three completely validated and built-in distribution-prepared stacks.

Revealed now, Intel developed Job Flipfast to improve the Linux gaming practical experience. The Flipfast stack makes it possible for conclusion users to run a graphical software in a virtual equipment (VM) though retaining indigenous GPU performance and entire host integration with zero-duplicate sharing involving the VM and host. The Flipfast stack drives boosts in gaming overall performance, and the engineering specifically applies to data heart video game streaming purposes.

Intel right now announced that the Intel® Implicit SPMD Plan Compiler (ISPC) will run on major of oneAPI Level Zero, which offers reduced-level, immediate-to-metallic interfaces tailored to gadgets in the oneAPI platform and is the over-all hardware abstraction layer. ISPC, powered by oneAPI, is a variant of the C programming language, which enables solitary program, various facts programming and is utilised to speed up Intel® OSPRay ray tracing motor on Intel CPUs. Intel is introducing Xe support to ISPC to seamlessly accelerate Intel oneAPI Rendering Toolkit elements this sort of as OSPRay.

What is Upcoming: At the virtual oneAPI Developer Summit on Nov. 12-thirteen, innovators, scientists and builders are showcasing forty collaborations and jobs using oneAPI. Subjects assortment from computational drug repurposing screening for COVID-19 to predicting crop yields. At SuperComputing 2020 setting up this week, Intel, alongside with market leaders and analysis companies, are highlighting applying oneAPI and oneAPI resources as a result of a keynote, tech sessions, fireside chats, demos and other pursuits. Complete information about Intel’s presence at SuperComputing 2020 is accessible on

Intel oneAPI and graphics software program stack updates and the debut of the Intel Server GPU mark monumental methods in Intel’s journey towards XPU architectures. These developments set the foundation to provide a lot more meaningful activities by way of Intel’s six pillars of know-how innovation and heterogenous architectures that will be enabled through an open regular-primarily based unified scalable application abstraction layer in oneAPI.

Additional Context: Tencent Groups with Intel on XianYou Cloud Gaming Platform | oneAPI Ecosystem Aid | Intel oneAPI Toolkits Actuality Sheet

The Smaller Print:

one TCO evaluation is based mostly on interior Intel investigate. Pricing as of 10/01/2020. Assessment assumes standard server pricing, GPU checklist pricing and software program pricing based on believed Nvidia software license charges of $one for each yr for five many years.

two Performance may possibly differ dependent on the certain sport title and server configuration. To reference the total listing of Intel Server GPU system measurements, please refer to this link on​

Functionality varies by use, configuration and other aspects. Master much more at

All products ideas and roadmaps are matter to change devoid of see.

Intel technologies might involve enabled components, program or service activation.

No item or ingredient can be definitely protected.

Your charges and success could differ.

Intel does not management or audit 3rd-get together facts.  You must check with other sources to evaluate accuracy.

All solution options and roadmaps are topic to modify without having discover.

Statements in this presentation that refer to business enterprise outlook, foreseeable future options and anticipations are forward-searching statements that involve a selection of dangers and uncertainties. Terms these as “anticipates,” “expects,” “intends,” “goals,” “plans,” “believes,” “seeks,” “estimates,” “continues,“ “may,” “will,” “would,” “should,” “could,” and versions of these types of text and comparable expressions are supposed to identify such ahead-searching statements. Statements that refer to or are centered on projections, uncertain occasions or assumptions, like statements relating to long term products and solutions and technology and the anticipated availability and positive aspects of this kind of items and technology, also establish ahead-seeking statements. This kind of statements are centered on management’s current anticipations, except an before date is indicated, and include several dangers and uncertainties that could result in true benefits to differ materially from individuals expressed or implied in these ahead-wanting statements. Vital aspects that could bring about actual outcomes to vary materially from the company’s expectations are set forth in Intel’s earnings launch dated October 22, 2020, which is bundled as an exhibit to Intel’s Kind 8-K furnished to the SEC on this kind of date, and Intel’s SEC filings, which includes the company’s most the latest reports on Forms ten-K and ten-Q. Copies of Intel’s Kind ten-K, 10-Q and 8-K stories could be obtained by browsing our Trader Relations website at or the SEC’s web site at

All facts introduced during this occasion displays management’s views as of the date of this presentation, except an before day is indicated. Intel does not undertake, and expressly disclaims any responsibility, to update any statement manufactured in this presentation, whether as a outcome of new details, new developments or in any other case, except to the extent that disclosure might be demanded by law.

The article Intel Executing toward XPU Vision with oneAPI and Intel Server GPU appeared 1st on Intel Newsroom.

Add Comment

Your email address will not be published. Required fields are marked *