Artificial Intelligence Latest news Pro Server & Services

Intel Innovation 2023: Intel focuses on AI including AI PC, new Xeon processors and the OpenVINO Toolkit

Today’s news deals with the first day of Intel’s Innovation 2023 and uses Intel’s official press material for this (due to a lack of personal presence). Of course, the following content is written from Intel’s point of view and should be classified accordingly. Nevertheless, it is of course interesting to read how and where Intel classifies itself, what its plans look like and what efforts it is making in the field of AI in order not to miss the boat.

In this context, Intel also confirms that the so-called “five-node-in-four-years-process” technology plan (how should we actually translate that in a manner appropriate to the species?) remains on track and presents the world’s first multi-chiplet package with Universal Chiplet Interconnect Express (UCIe). The company also revealed new details on next-generation Intel® Xeon® processors, reporting significant advances in energy efficiency and performance. In addition, they are introducing an e-core processor with 288 cores. The 5th generation of Intel® Xeon® processors will be launched on December 14.

The so-called AI PC will be introduced with the launch of the Intel® Core™ Ultra processors on December 14. Featuring Intel’s first integrated Neural Processing Unit, Core Ultra is said to enable energy-efficient AI acceleration and local inference on the PC. A large AI supercomputer will be built on Intel Xeon processors and Intel® Gaudi®2 AI hardware accelerators, with Stability AI as the primary customer. The general availability of Intel® Developer Cloud for building and testing high-performance applications such as AI was also announced, including details that it is already being used by customers. New and upcoming Intel software solutions, including the 2023.1 release of the Intel® Distribution of OpenVINO™ Toolkit, will help developers unlock new AI capabilities.

OpenVino Fact Sheet_Innovation23_FINAL APPROVED FOR USE

 

But let’s get to the parts from Intel’s press release after this abbreviated overview and summary. Namely, at its third annual Intel Innovation event, Intel unveiled a set of technologies designed to make AI ubiquitous and accessible across all workloads, from client and edge to network and cloud.

“AI represents a generational shift, ushering in a new era of global expansion where computing is even more fundamental to a better future for all,” said Intel CEO Pat Gelsinger. “For developers, this opens up tremendous social and business opportunities to push the boundaries of what’s possible, create solutions to the world’s biggest challenges and improve the lives of every person on the planet.”

In a keynote presentation opening the event, Gelsinger showed how Intel is bringing AI capabilities on top of its hardware products and making them accessible through open, multi-architecture software solutions. He also highlighted how AI is helping to drive what he called the “Siliconomy,” a “growing economy enabled by the magic of silicon and software.” Today, silicon feeds a $574 billion industry, which in turn fuels a global tech economy worth nearly $8 trillion.

 

New advances in chips, packaging and multi-chiplet solutions

The work begins with innovations in silicon. Intel’s five-node-in-four-years process development program (what a term!) is progressing well, Gelsinger said, with Intel 7 already in high-volume production, Intel 4 ready for manufacturing, and Intel 3 on track for the end of this year. Gelsinger also showed an Intel 20A wafer with the first test chips for Intel’s Arrow Lake processor, destined for the client computing market in 2024. Intel 20A will be the first process node to integrate PowerVia, Intel’s backside power supply technology, and the new gate-all-around transistor design called RibbonFET. Intel 18A, which also uses PowerVia and RibbonFET, remains on track to be ready for production in the second half of 2024.

Another way Intel plans to push Moore’s Law is with new materials and new packaging technologies, such as glass substrates – a groundbreaking announcement from Intel this week. When introduced later this decade, glass substrates will enable continuous scaling of transistors on a package to meet the needs of data-intensive, high-performance workloads like AI and keep Moore’s Law alive well beyond 2030.

Intel also showed a test chip package with Universal Chiplet Interconnect Express (UCIe). The next wave of Moore’s Law will come with multi-chiplet packages, Gelsinger said, and will come faster if open standards can reduce the friction of integrating IP. Founded last year, the UCIe standard will allow chiplets from different vendors to work together, enabling new designs for extending different AI workloads. The open specification is supported by more than 120 companies. The test chip combines an Intel UCIe IP chiplet fabricated on Intel 3 and a Synopsys UCIe IP chiplet fabricated on the TSMC N3E process node. The chiplets are connected using advanced packaging technology called Embedded Multi-Die Interconnect Bridge (EMIB). The demonstration underscores the commitment of TSMC, Synopsys and Intel Foundry Services to support an open, standards-based chiplet ecosystem with UCIe.

Boosting performance and extending AI for all(s)

Gelsinger highlighted the range of AI technology available to developers today on Intel platforms – and how that range will expand dramatically in the coming year to make AI available everywhere. Gelsinger pointed to the planned launch of the 5th generation Intel® Xeon® processors, which will hit the market on December 14. These processors will offer significant advances in energy efficiency and performance, including an e-core processor with an impressive 288 cores.

Another highlight was the unveiling of the Intel® Core™ Ultra processor, which will also be launched on December 14. This processor will include Intel’s first integrated Neural Processing Unit and is expected to enable energy-efficient AI acceleration and local inference on the PC. Construction has also been announced to begin on a large AI supercomputer. This supercomputer will be based on Intel Xeon processors and Intel® Gaudi®2 AI hardware accelerators, with Stability AI as the primary customer.

To support the development of high-performance applications, including AI, Intel has announced the general availability of the Intel® Developer Cloud. The cloud is already being used by customers to build and test applications. In addition, Intel showcased new and upcoming software solutions, including the 2023.1 release of the Intel® Distribution of OpenVINO™ Toolkit, which will help developers unlock new AI capabilities.

Intel CEO Pat Gelsinger emphasized the central role of AI in today’s society, stating that it “represents a generational shift and ushers in a new era of global expansion.” He pointed out that AI is paving the way for the “Siliconomy,” an economy enabled by “the magic of silicon and software.” In summary, Intel unveiled a series of technological innovations aimed at making AI available everywhere and helping developers unlock new AI capabilities. With a clear focus on AI, Intel unveiled the next generation of processors and software solutions designed to drive the “Siliconomy” and change the world through the power of silicon and software.

Kommentar

Lade neue Kommentare

hansmuff

Mitglied

48 Kommentare 32 Likes

Interessant und ich wünsche viel Glück. Kaufe ich gerne, wenn's gut wird.

Aaaaaaber... dieses ganze "AI" Zeug geht mir echt gegen den Strich. Da ist überhaupt nichts "intelligent", in keiner Hinsicht. Es ist eher so 'ne Art Nachgeplapper, aber mit dem Versuch, Zusammenhang zwischen menschlich generierten Artifakten zu finden, kombinieren und ausspucken. Mit keiner Ahnung ob es denn jetzt stimmt.

Das mal voraus.

So, jetzt ist da so'n Modul auf meiner CPU. Was kann das denn? Was GENAU? Lokal eine "AI" erstellen? Nein, braucht ja ein riesiges Modell, hat keiner außer halt den "Großen." Also was dann jetzt? Es ist ja nicht intelligent. Soll das dann das Windows Start Menu unterstützen?
Zur Suche ist es total ungeeignet, da braucht man keine "AI", es ist eine brute force Sache die Durchsatz braucht, jede CPU kann den Rest.

Also was DANN? Irgendwie mit Bildmanipulation oder Streaming oder Videokonferenz was neues raushauen? Wird mit so 'ner kleinen Einheit nichts, die hat ja auch keine Quelldaten. Und GPUs können solche Sachen eh wesentlich schneller.

Ich persönlich finde ChatGPT supergeil. Echt Klasse um Programmieraufgaben etwas aufzuknacken, oder SQL oder RegEx zu erstellen, es gibt wirklich Probleme die das Ding super löst.

Aber dieses ganze Gemache um "AI" herum ist Quatsch und Augenblenderei. Das kleine Modul bringt nichts und der ganze KI Ramsch wird vom Marketing erzwungen, denn mittlerweile ist ja auch der Kackbeutel für die Hundescheiße "KI betrieben."

Antwort 1 Like

Klicke zum Ausklappem
e
eastcoast_pete

Urgestein

1,556 Kommentare 883 Likes

Jetzt gibt's in der Tat in letzter Zeit viel Hype um KI, allerdings kann ich mir durchaus ein paar Dinge vorstellen, die von einer NPU (oder wie immer die dann auch heißt) profitieren könnten; zB Beam Forming bei 5G Mobilfunk und auch neuen WiFi Standards.
Außerdem: selbst wenn sich Sachen wie Bild und Audio Optimierung in Videokonferenzen auch mit der CPU machen lassen, kann eine dedizierte NPU das oft sehr viel effizienter. Gerade bei Laptops durchaus eine Erwägung. Aber ja, jetzt sind sowohl Hardware wie auch Software Anbieter im Zugzwang, und sollen zeigen, was KI "vor Ort" denn bringt.

Antwort Gefällt mir

hansmuff

Mitglied

48 Kommentare 32 Likes

Wie kann denn die NPU das besser als eine kleine GPU oder die normale CPU? Ich benutze NV Broadcast und das funktioniert super, was kann denn diese NPU besser? Warum ist die effizienter, wer hat denn da benchmarks? Keiner. Welche API gibt's dafür und wer kann die benutzen?

Will Dich nicht dumm anmachen, aber Beweise für die Kraft dieser neuen Einheit muß ich erstmal finden. Und danach die praktische Anwendung, und ob der Befehlssatz universell wird.

Antwort Gefällt mir

G
Guest

BITTE KEINE KI IN HARDWARE!

Haben wir schon vergessen, was diese Northsections waren..etz?

Ich will die pure CPU und pure Rechenleistung. ( und Pures Os)
Wird nicht kommen.
Das Ende dieser Entwicklung sind digital voll betreute semibetrunkene Honks, die sogar einen Entscheidungs-ASI haben
ob ES jetzt gut wäre gaxxi zu gehen oder vorher noch.. was red ich..

Das wäre mir lieber. Eine Intel CPU die als pure 8 Kern einen Cb von 80.000 hat dabei 90 Watt braucht und es ist gut.

Aber das kapiert keiner. Statt der Zeitenwende, die 10 fache Leistung bedeute baut man den Nonsens ein.

nvidia amd arm intel...etz.. only your data mind set not power..

EIN BILD und ein TON ist vorher perfekt her zu richten. Net irgendwie und nachher optimieren. vergiss es.

TAU DAU VEGAN WOK ISTAN = T D V W I = so heißt unser Zeitalter! = in short lol²

Antwort Gefällt mir

Danke für die Spende



Du fandest, der Beitrag war interessant und möchtest uns unterstützen? Klasse!

Hier erfährst Du, wie: Hier spenden.

Hier kannst Du per PayPal spenden.

About the author

Igor Wallossek

Editor-in-chief and name-giver of igor'sLAB as the content successor of Tom's Hardware Germany, whose license was returned in June 2019 in order to better meet the qualitative demands of web content and challenges of new media such as YouTube with its own channel.

Computer nerd since 1983, audio freak since 1979 and pretty much open to anything with a plug or battery for over 50 years.

Follow Igor:
YouTube Facebook Instagram Twitter

Werbung

Werbung