#xeon6

[icnweb]

인텔-AWS, '제온 6' 프로세서로 클라우드 혁신 가속화

8세대 EC2 인스턴스 출시… 내장형 AI 가속과 압도적인 메모리 성능고성능 클라우드 컴퓨팅 시대 주도 인텔 제온 6 프로세서 (image. Intel) [아이씨엔 오승모 기자] 클라우드 컴퓨팅 시장의 두 거인, 인텔과 아마존웹서비스(AWS)가 다시 한번 손을 잡고 기술 혁신의 새로운 장을 열었다. 양사는 최근 P-코어를 탑재한 최신 ‘인텔® 제온® 6 프로세서’에 기반한 8세대 아마존 EC2(Amazon Elastic Compute Cloud) R8i 및 R8i-플렉스 인스턴스를 공식 출시한다고 발표했다. 이는 수년간 이어진 양사의 긴밀한 협력의 결과물로, AI와 실시간 데이터 처리 등 고성능을 요구하는 클라우드 환경에 최적화된 솔루션을 제공하는 것을 목표로 한다. 최근 기업 환경은 AI 모델 학습…

View On WordPress

[govindhtech]

Intel Xeon 6 E-core Processors for Gamers and Creators

Intel Xeon 6 E-core

Intel revealed state-of-the-art technologies and architectures today at Computex that have the potential to significantly accelerate the AI ecosystem from the data centre, cloud, and network to the edge and PC. With increased processing power, cutting-edge power efficiency, and an affordable total cost of ownership (TCO), clients may now take use of the full potential of AI systems.

AI Data Centres Benefit from Intel Xeon 6 Processors

Companies are under increasing pressure to update their outdated data centre systems in order to maximise physical floor and rack space, save costs, meet sustainability targets, and develop new digital capabilities throughout the organisation as digital transformations pick up speed.

With both Intel Xeon 6 E-core (Efficient-core) and P-core (Performance-core) SKUs to address the wide range of use cases and workloads, from AI and other high-performance compute needs to scalable cloud-native applications, the Xeon 6 platform and processor family as a whole were designed with these issues in mind. Built on a shared software stack and an open ecosystem of hardware and software manufacturers, E-cores and P-cores share a compatible architecture.

The Intel Xeon 6 E-core, code-named Sierra Forest, is the first of the Xeon 6 CPUs to be released and will be available starting today. The Xeon 6 P-cores, also known as Granite Rapids, should be released the following quarter.

The Intel Xeon 6 E-core processor offers good performance per watt and a high core density, allowing for efficient computing at much lower energy costs. For the most demanding high-density, scale-out workloads, such as cloud-native apps and content delivery networks, network microservices, and consumer digital services, the enhanced performance with higher power efficiency is ideal.

Furthermore, when compared to 2nd Gen Intel Xeon processors on media transcoding tasks, Intel Xeon 6 E-core enormous density advantages allow for rack-level consolidation of 3-to-1, giving clients a rack-level performance gain of up to 4.2x and performance per watt gain of up to 2.6×1. Xeon 6 processors free up computational capability and infrastructure for creative new AI applications by consuming less power and rack space.

Intel Gaudi AI Accelerators Improve GenAI Performance at Lower Cost

These days, it’s getting cheaper and quicker to use generative AI. The industry standard for infrastructure, x86 runs at scale in almost all data centre environments and provides the basis for integrating AI capabilities while guaranteeing affordable interoperability and the enormous advantages of an open community of developers and users.

When used in conjunction with Intel Gaudi AI accelerators, which are specifically intended for AI applications, Intel Xeon processors make the best CPU head node for AI workloads. When combined, these two provide a potent solution that blends in well with the current infrastructure.

For training and inference of large language models (LLM), the Gaudi architecture is the only MLPerf-bench marked substitute for Nvidia H100 that offers customers the desired GenAI performance at a price-performance advantage that offers choice, quick deployment, and a lower total cost of operation.

System providers can purchase a basic AI kit for $65,000, which includes eight Intel Gaudi 2 accelerators and a universal baseboard (UBB). This kit is anticipated to be one-third less expensive than equivalent competitor platforms. Eight Intel Gaudi 3 accelerators with a UBB will be included in a kit that will retail for $125,000; this is around two-thirds less than comparable competition platforms.

With the help of Intel Gaudi 3 accelerators, businesses will be able to extract more value from their unique data by achieving notable performance gains for training and inference workloads on top GenAI models. According to projections, Intel Gaudi 3 in a 8,192-accelerator cluster will provide up to 15% better training throughput for a 64-accelerator cluster compared to Nvidia H100 on the Llama2-70B model and up to 40% faster time-to-train compared to the equal size Nvidia H100 GPU cluster. Furthermore, it is anticipated that Intel Gaudi 3 will provide up to two times quicker inferencing on average when compared to Nvidia H100 while running widely used LLMs like Mistral-7B and Llama-70B.

Intel is working with at least ten of the leading international system providers, including six new companies that just stated they will be releasing Intel Gaudi 3, to make these AI systems widely accessible. Leading system providers Dell, HPE, Lenovo, and Supermicro now have more production options thanks to new partners Asus, Foxconn, Gigabyte, Inventec, Quanta, and Wistron.

Revolutionary laptop AI architecture triples compute and power efficiency

Intel is expanding its AI presence outside of the data centre, both in the PC and at the edge. Intel has been enabling enterprise choice for decades with more than 200 million CPUs deployed to the ecosystem and more than 90,000 edge deployments.

Intel is leading the charge in this category-creating moment as the AI PC category is revolutionising every facet of the computing experience today. The goal now is to create edge devices that learn and change in real time, anticipating user requirements and preferences and ushering in a completely new era of productivity, efficiency, and creativity. It is no longer just about faster processing speeds or sleeker designs.

By 2028, 80% of PC sales are expected to come from AI models, according to Boston Consulting Group. Intel reacted swiftly, enabling over 100 independent software vendors (ISVs), 300 features, and support for 500 AI models throughout its Core Ultra platform, to provide the best hardware and software platform for the AI PC.

Building swiftly on these unparalleled benefits, the company today unveiled the Lunar Lake architecture, which serves as the flagship processor for the upcoming AI PC generation. Lunar Lake is expected to provide up to 40% reduced SoC power and more than three times the AI compute, thanks to a significant leap in graphics and AI processing power and an emphasis on capability-efficient compute performance for the thin-and-light market. It is anticipated to ship in 2024’s third quarter, just in time for the Christmas shopping season.

The brand-new architecture of Lunar Lake will allow for:

The new Performance-cores (P-cores) and Efficient-cores improve performance and energy efficiency.

A fourth-generation Intel NPU with 48 tera-operations per second (TOPS) AI capabilities. Improves in generative AI are made possible by this potent NPU, which provides up to 4x AI compute compared to the previous iteration.

The brand-new X2 GPU cores for visuals and the X Matrix Extension (XMX) arrays for AI are combined in the Battlemage GPU design. The new XMX arrays provide a second AI accelerator with up to 67 TOPS of performance for exceptional throughput in AI content production, while the X2 GPU cores increase gaming and graphics performance by 1.5x over the previous version.

Amazing laptop battery life is made possible by an innovative compute cluster, an advanced low-power island, and Intel innovation that manages background and productivity tasks extremely well.

Intel is already shipping at scale, delivering more AI PC processors through the first quarter of 2024 than all competitors combined, while others get ready to join the AI PC market. More than 80 distinct AI PC designs from 20 original equipment manufacturers (OEMs) will be powered by Lunar Lake. This year, Intel anticipates putting more than 40 million Core Ultra processors on the market.

Read more on govindhtech.com

[govindhtech]

Intel Computex Unveils AI : Xeon 6, Gaudi, Lunar Lake CPUs

Intel Computex 2024

In order to make AI affordable and available to everyone, Intel CEO Pat Gelsinger announced at Intel Computex 2024 the newest technologies that combine state-of-the-art performance and power efficiency in data centres, acceleration, and the AI PC experience. Intel is dedicated to driving AI potential for partners and customers, while also empowering open ecosystems. Intel is still at the forefront of driving the industry towards a sustainable and scalable future with its increased processing power, cutting-edge power efficiency, and affordable total cost of ownership.

Gelsinger gave a keynote address at the June 4–7 event in Taipei, Taiwan, where he introduced the Intel Xeon 6 processors with Efficient-cores (E-cores), revealed the cost of the Intel Gaudi 2 and 3 AI accelerator kits, and unveiled the ground-breaking Lunar Lake client processor architecture a ground-breaking design that expands the AI PC category even further.

Lunar Lake

For its forthcoming Lunar Lake client computing processor, which was completely rebuilt to deliver industry-leading core and graphics performance with unparalleled AI, Intel unveiled the architecture specifics.

In comparison to the previous generation, the new Performance-cores (P-cores) and Efficient-cores (E-cores) give incredible performance at up to 40% reduced system-on-chip power. Compared to the previous generation, a new neural processing unit is up to 4 times quicker, allowing for equivalent advancements in generative AI. Moreover, performance in gaming and graphics is 1.5 times better with the new Xe2 graphics processing unit cores compared to the previous version.

Starting in the third quarter of 2024, Lunar Lake will power more than 80 new AI PC designs from more than 20 partners.

Lunar Lake Structure: A Motion Picture

An animated film demonstrates the various technological levels that comprise Lunar Lake. Intel unveiled Lunar Lake architecture at Intel Computex in Taipei on June 4, 2024. When compared to the previous generation, the thin-and-light PC processor offers notable gains in performance and efficiency in the neural processing unit, central processing unit, and graphics processing unit. Intel Corporation is credited.

Pictures of Lunar Lake

Image credit to Intel

Lunar Lake Provides the AI PC with a New Architecture

Intel unveiled the Lunar Lake architecture at Intel Computex 2024, revealing the flagship chip for the upcoming AI PC generation. Lunar Lake focuses on power-efficient compute performance for the thin-and-light segment while providing a significant boost in graphics and AI processing capabilities. Intel Corporation is credited.

Intel Core 6

The Intel Xeon 6 family of processors, which comes with E-core and P-core variants, was introduced by Intel to handle a wide range of workloads and use cases, including scalable cloud-native apps and high-performance computation requirements for artificial intelligence.

The first member of the series, code-named Sierra Forest, is the Intel Xeon 6 CPU with efficient cores, making its debut at Intel Computex 2024. When compared to 2nd Gen Intel Xeon processors on media transcoding workloads, it offers a performance per watt boost of up to 2.6x and a rack-level performance gain of up to 4.2x. Its high core density and remarkable performance also allow for rack-level consolidation of 3-to-1.

The third quarter of 2024 will see Intel Xeon 6 processors with P-cores increase AI, high-performance computing, image processing, and data analytics.

Xeon 6 Picture

Image credit to Intel

An Overview of Intel Xeon 6 Processors in 60 Seconds

Three key points concerning Intel Xeon 6 processors with Efficient-cores (E-cores) are outlined in this one-minute summary. Known by the code name Sierra Forest, they were introduced on June 4, 2024, at Intel Computex in Taipei, Taiwan. Designed with E-cores, they enhance performance and reduce power consumption for applications that require high densities and scaling. The Intel Xeon 6 processors with Performance-cores (P-cores), code-named Granite Rapids, will shortly trail this initial member of the family. Workloads requiring additional computation, such AI and computer vision, will be supported. Intel Corporation is credited.

Kulim, Malaysia; San Jose, Costa Rica: Intel Manufacturing (B-Roll)

The Intel Kulim Assembly Test (KuAT) plant in Kulim, Malaysia, and the Costa Rica Assembly Test (CRAT) site in San Jose, Costa Rica, are depicted in B-roll film from May 2024. It features footage of Intel Xeon 6 processors with efficient cores (code-named Sierra Forest) and Intel personnel inside the buildings. Intel Corporation is credited.

Gaudi Intel

With a price-performance advantage that offers customers options, quick deployment, and a lower total cost of operation, the Intel Gaudi architecture delivers the generative AI performance they desire.

Intel revealed at Intel Computex that a typical AI kit, which includes eight Intel Gaudi 2 accelerators and a universal baseboard (UBB), is expected to cost system providers about one-third as much as comparable competition systems. The kit is available to system providers for $65,000. Eight Intel Gaudi 3 accelerators with a UBB will be included in a kit that will retail for $125,000, which is roughly two-thirds less than comparable competitor platforms2.

Additionally, Intel disclosed that the launch of Intel Gaudi 3 systems is anticipated from six new system vendors. With plans to sell Intel Gaudi 3 systems, ASUS, Foxconn , the Gigabyte, Inventec, Quanta, and Wistron join Dell, HPE, Lenovo, and Supermicro.

Intel Tech Tour

Intel held its third-annual Technology Tour in Taiwan just a few days before Intel Computex, providing worldwide media and analysts with presentations on Xeon 6 and Gaudi accelerators as addition to an in-depth look at Lunar Lake architecture. Bringing next-generation technologies to life is the responsibility of business and technical leaders from throughout Intel, and these leaders delivered keynote addresses and technical deep dives over the two-day event.

Read more on Govindhtech.com

[teknolojihaber]

Intel, daha önce Sierra Forest ve Granite Rapids olarak adlandırılan Xeon 6 işlemcilerini duyurdu

Yüksek performanslı P-çekirdeklerini temel alan yeni Intel Sierra Forest işlemcileri ve  enerji tasarruflu E-çekirdeklerini temel alan Granite Rapids, aynı aile olan Xeon 6 içinde üretilecek. Intel bunu, gerçekleşen Vision 2024 etkinliğinin bir parçası olarak duyurdu. Phoenix, Arizona'da üretilecek, işlemciler adına Scalable markasını terk ederek  yalnızca seri numarasına sahip yeni çipleri piyasaya sürecek. Resmi olarak Sierra Forest ve Granite Rapids işlemcilerinin duyurusu geçen yıl gerçekleşti. Bu çiplerin toplu teslimatlarının bu yılın ikinci yarısında başlaması bekleniyor. Intel, Skylake mimarisi üzerine inşa edilen Platinum 8100 serisi çiplerin piyasaya sürülmesiyle 2017 yılında Xeon işlemcileri için Scalable markasını kullanmaya başladı. Intel, 2024 yılında veri merkezi, uç ve kişisel bilgi işlem segmentlerindeki mevcut müşteri taleplerini karşılamak için Intel Kurumsal Yapay Zeka ürünleri ve hizmetlerinden oluşan bir ekosistem oluşturmayı hedefliyor. Bu ekosistemin bir parçası olarak şirket, geniş bir çerçeve ve donanım yelpazesi kullanmayı planlıyor ve ayrıca bağımsız yazılım ve donanım satıcılarıyla da yakın işbirliği içinde çalışacak. Kurumsal pazar için tasarlanan yeni Xeon 6 işlemci ailesi bu girişimde önemli bir rol oynamaktadır. Granite Rapids ve Sierra Forest işlemcilerinin piyasaya sürülmesiyle Intel, her iki seriyi birleştiren yeni ve daha basit bir terminoloji kullanmaya karar verdi. Çip modelleri kurumsal segmente, iş istasyonlarına, sunuculara ve veri merkezlerine yönelik kategorilere ayrılmaya devam edecek. Ölçeklenebilirlik marka isminden kaybolacak, sadece nesil numarası kalacak. Sierra Forest Xeon 6 işlemciler piyasaya ilk olarak çıkacak ve bu yılın ikinci çeyreğinde piyasaya sürülecek. Bu yongalar 288'e kadar enerji tasarruflu çekirdek sunacak ve büyük performanslı P çekirdekleriyle donatılmayacak. Buna karşılık, P-çekirdeklerine dayalı Granite Rapids Xeon 6 işlemciler bu yılın ikinci yarısında piyasaya çıkacak. Read the full article