Intel Updates Technology Roadmap with Data Center Processors and Game Streaming Service

[AleksandarK]

At Intel's 2022 Investor Meeting, Chief Executive Officer Pat Gelsinger and Intel's business leaders outlined key elements of the company's strategy and path for long-term growth. Intel's long-term plans will capitalize on transformative growth during an era of unprecedented demand for semiconductors. Among the presentations, Intel announced product roadmaps across its major business units and key execution milestones, including: Accelerated Computing Systems and Graphics, Intel Foundry Services, Software and Advanced Technology, Network and Edge, Technology Development, More: For more from Intel's Investor Meeting 2022, including the presentations and news, please visit the Intel Newsroom and Intel.com's Investor Meeting site.



ACCELERATED COMPUTING SYSTEMS AND GRAPHICS
The Accelerated Computing Systems and Graphics Group (AXG) is on track to ship products across its three segments and deliver more than $1 billion in revenue in 2022. As a growth engine for Intel, AXG's three segments together will approach $10 billion of revenue for Intel by 2026.

Visual Compute Roadmap and Strategy

• Intel Arc Graphics Timing and Roadmap Update - AXG expects to ship more than 4 million discrete GPUs in 2022. OEMs are introducing notebooks with Intel Arc graphics, code-named Alchemist, for sale in the first quarter of 2022. Intel will ship add-in cards for desktops in the second quarter and workstations by the third quarter. Architecture work has begun on Celestial, a product that will address the ultra-enthusiast segment.
• Project Endgame - Project Endgame will enable users to access Intel Arc GPUs through a service for an always-accessible, low-latency computing experience. Project Endgame will be available later this year.
• Super Compute Roadmap and Strategy - More than 85 percent of the world's supercomputers run on Intel Xeon processors. Building on this foundation, AXG is extending to higher compute and memory bandwidth and will deliver a leadership CPU and GPU roadmap to power high performance computing (HPC) and AI workloads. To date, Intel expects more than 35 HPC-AI design wins from top OEMs and CSPs. Additionally, AXG has set a course that paves the way to zetta-scale by 2027.
• Sapphire Rapids with High-Bandwidth Memory (HBM) - HBM integrated into the package with Sapphire Rapids offers applications up to 4x more memory bandwidth, providing a generational improvement of 2.8x over Intel 3rd Gen Xeon processors. On the same computational fluid dynamics application, Sapphire Rapids with HBM outperforms competing solutions by up to 2.8x.
• Ponte Vecchio - AXG is on track to deliver Ponte Vecchio GPUs for the Aurora supercomputer program later this year. Ponte Vecchio achieved leadership performance results with up to 2.6x more performance compared with the leading market solution on a complex financial services workload.
• Arctic Sound-M - Arctic Sound-M brings the industry's first hardware-based AV1 encoder into a GPU to provide 30% bandwidth improvement and includes the industry's only open-sourced media solution. The media and analytics supercomputer enables leadership transcode quality, streaming density and cloud gaming. Arctic-Sound M is sampling to customers and will ship by mid-2022.
• Falcon Shores - Falcon Shores is a new architecture that will bring x86 and Xe GPU together into a single socket. This architecture is targeted for 2024 and projected to deliver benefits of more than 5x performance-per-watt, 5x compute density, 5x memory capacity and bandwidth improvements.1
• Custom Compute Group - AXG's Custom Compute Group will build tailored products for emerging workloads such as blockchain, supercomputing at the edge, premium infotainment for cars, immersive displays and more.

INTEL FOUNDRY SERVICES
The automotive industry is undergoing a profound transformation, as vehicles become electric, safer, smarter and more connected. These trends are driving considerable growth, with automotive silicon revenues expected to nearly double to $115 billion by 2030. The fragmented supply chains and legacy process technologies of today will not be able to support the increasing demand and the transition to more compute-intensive applications. In response, Intel Foundry Services (IFS) is forming a dedicated automotive group to deliver a complete solution to automakers with three clear priorities:

• Open Central Compute Architecture - IFS will develop a high-performance open auto compute platform that enables automotive OEMs to build next-generation experiences and solutions. This open compute architecture will leverage chiplet-based building blocks, along with Intel's advanced packaging technologies, providing significant flexibility to build solutions optimized for technology nodes, algorithms, software and applications that address the compute needs of next-generation vehicles.
• Automotive-Grade Foundry Platform - Intel will enable manufacturing technologies that meet the stringent quality requirements of automotive applications and customers. IFS is targeting both leading-edge nodes and technologies optimized for microcontrollers and unique automotive needs, in combination with advanced packaging, to help customers design multiple types of automotive semiconductors. A partnership with Mobileye, a leader in advanced driver-assistance systems (ADAS) solutions with deep experience in automotive-grade products, is enabling IFS to deliver its advanced technology nodes for the automotive segment.
• Enabling Transition to Advanced Technologies - IFS will offer design services and Intel IP to automakers, enabling them to take advantage of Intel's expertise from silicon to system design. The IFS Accelerator automotive program, announced last year, is designed to help automotive chipmakers transition to advanced process and packaging technologies and innovate with Intel's custom and industry-standard IP portfolio.

SOFTWARE AND ADVANCED TECHNOLOGY
Software is a key component of Intel's competitive advantage that adds value across the stack for Intel's client, edge, cloud and data center businesses. Intel's approach to foster an open ecosystem ensures trust, choice and interoperability for our industry and acts as a catalyst for technological adoption and innovation. Intel's investments in software also present opportunities for growth that are both disruptive and transformative.

• Cross-Platform, Open Development - Intel oneAPI toolkits deliver a cross-platform, open programming model that unleashes developers to solve unique challenges with optimized performance.
• Solving Challenges with AI - The convergence of security and AI demonstrates the incredible promise of open and collaborative frameworks that help protect data while gathering insights. Intel Core processors and Intel vPRO systems using Intel Threat Detection Technology detect malware behavior below the operating system and feed those insights to endpoint detection and response solutions. For confidential computing in the cloud, 3rd Gen Intel Xeon processors with Intel Software Guard Extensions protect data and AI models so data can be aggregated and greater insights can be gathered to solve challenging problems, such as identifying brain tumors.

NETWORK AND EDGE
Networking and edge computing is a quickly growing industry. To accelerate growth in this market and drive the shift toward software-defined and fully programmable infrastructure, Intel created the Network and Edge Group (NEX) in 2021. Intel expects networking and edge revenue to grow at a faster rate than overall TAM over the next decade and be a significant contributor to the company's overall growth. To take advantage of this opportunity, NEX is producing programmable hardware and open software from the cloud through the internet and 5G networks to the intelligent edge.

• Intelligent Fabric - Intel Intelligent Fabric is programmable platform that allows customers to advance business opportunities by programming network behavior from end to end, through the infrastructure within the data center. It puts control in the hands of customers, giving them the means to program the network. This lets customers constantly evolve, improve and differentiate their infrastructure for themselves and creates a future where a new class of computing device, the Infrastructure Processing Unit (IPU), can be integrated into data centers, accelerating cloud infrastructure and maximizing performance.
• Mobile Network Transformation - Intel has led the transformation of telecommunications networks for more than a decade, propelling the world's networks away from legacy, fixed-function hardware and enabling them to be defined by open and interoperable software. Intel's ambition is to provide its customers with the industry's absolute best and broadest programmable platforms to advance business opportunities, putting control in the hands of the developers to support the buildout of 5G and beyond.
• Accelerating the Intelligent Edge - Intel offers a diverse portfolio of hardware and software offerings and a vast ecosystem of partners to help customers deliver intelligent edge platforms. Enabling new use cases and workloads across a range of vertical industry markets, NEX is positioned to meet the growing need for compute and analytics at the intelligent edge. AI - specifically inferencing at the edge - provides actionable intelligence where and when the data happens. As such, it is becoming the most prolific use case at the edge, transforming and automating factories, smart cities, hospitals and more.

TECHNOLOGY DEVELOPMENT
Intel remains on track to reclaim transistor performance per watt leadership by 2025. Intel's advanced test and packaging technologies give it unmatched industry leadership that benefits its products and foundry customers and will play a critical part in the pursuit of Moore's Law. Continuous innovation is the cornerstone of Moore's Law, and innovation is very much alive and well at Intel.

• Process - Intel 7 is in production and shipping in volume with the launch of 12th Gen Intel Core processors and additional products coming in 2022. Intel 4, our implementation of extreme ultraviolet (EUV) lithography, will be manufacturing-ready in the second half of 2022. It delivers an approximate 20% increase in transistor performance per watt. Intel 3, with additional features, delivers a further 18% performance per watt and will be manufacturing-ready in the second half of 2023. Ushering in the Angstrom Era with RibbonFET and PowerVia, Intel 20A will deliver up to a 15% performance per watt improvement and will be manufacturing-ready in the first half of 2024. Intel 18A delivers an additional 10% improvement and will be manufacturing-ready in the second half of 2024.
• Packaging - Our advanced packaging leadership gives designers options across thermals, power, high-speed signaling and interconnect density, to maximize and co-optimize product performance. In 2022, Intel will ship leadership packaging technologies in Sapphire Rapids and Ponte Vecchio and start risk production on Meteor Lake. Foveros Omni and Foveros Direct, our advanced packaging technologies unveiled at Intel Accelerated in July 2021, will be manufacturing-ready in 2023.
• Innovation - As Intel looks forward to technologies such as High-NA EUV, RibbonFET, PowerVia, and Foveros Omni and Direct, its leaders see no end to innovation and therefore no end to Moore's Law. Intel remains undeterred in achieving its aspiration of delivering approximately 1 trillion transistors in a single device by the end of the decade.

1.Falcon Shores performance targets based on estimates relative to current platforms as of February 2022.

View at TechPowerUp Main Site

 

[TheLostSwede]

The TL;DR version.

 

[Steevo]

So we have review samples out yet? No?

Smoke and mirrors, rinse and repeat.

 

[Deleted member 24505]

5700xt anyone?

 

[gasolina]

this is too painful to see......another day another post another roadmap another slide and so on.......why intel so obsessed with stuff they can't really made...... just my thought final review will be clear ......but my guess they only at 3060 level at best since it's Raja..... dude give ppl hype train of vega, rx480 blah blah and the last thing is intel alchemist

 

[Chrispy_]

So realistically, Intel is using today's announcement to bury the fact that desktop GPUs have slipped from Q1 to Q2 in amongst a whole bunch of other not-news.
Great execution there, classic Intel PR bullshit.

Intel competing with AMD and Nvidia in the GPU space has been a long-running meme for too long now; When Intel were touting their GTX1060 competitor, I don't think I'm alone in thinking that meant "whilst the GTX 1060 is Nvidia's current product", not one that was retired in 2018.

 

[v12dock]

As someone who uses Xe on their laptop there drivers have made 0 improvements since the launch of Xe. I have no faith that they will meet the Q2 launch and if they do it will be an awful experience.

 

[Chrispy_]

As someone who uses Xe on their laptop there drivers have made 0 improvements since the launch of Xe. I have no faith that they will meet the Q2 launch and if they do it will be an awful experience.

I don't regularly use one but there are a handful of i5-1135G7 laptops in the fleet and the bugfixes are happening but at a glacial pace - too slow to be relevant to the rapid pace of game releases.

In terms of game-ready drivers, Intel are about a decade behind AMD and Nvidia, though I woudn't call the experience "awful". As a mediocre IGP it handles itself acceptably and bugs aren't that common - it's not like the good old days of Intel GMA chipset graphics where finding a game that even ran properly at all was a real challenge.

 

[catulitechup]

• Arctic Sound-M - Arctic Sound-M brings the industry's first hardware-based AV1 encoder into a GPU to provide 30% bandwidth improvement and includes the industry's only open-sourced media solution.

seriously thinking about send nvidia jensen to hell

 

[Chrispy_]

seriously thinking about send nvidia jensen to hell

don't get your hopes up. Intel Quicksync has always been quick, and almost uselessly-dirty with terrible compression efficiency and quite poor actual encoding quality. It's fast but in almost every scenario I ever looked at (admittedly I haven't looked at it for a few generations now) it's unusable either for quality reasons or filesize/bitrate reasons. I'd defer to CPU encoding every time, without fail.

 

[catulitechup]

don't get your hopes up. Intel Quicksync has always been quick, and almost uselessly-dirty with terrible compression efficiency and quite poor actual encoding quality. It's fast but in almost every scenario I ever looked at (admittedly I haven't looked at it for a few generations now) it's unusable either for quality reasons or filesize/bitrate reasons. I'd defer to CPU encoding every time, without fail.

actually seems ok for my needs and send jensen to hell is priceless

 

[trsttte]

They are setting themselves for failure. Gpu stocks and prices are slowly improving and their potential customer pool is reducing as days go by. There's also the upcoming RDNA3 and Lovelace and the constant rumours of driver teething pains on a card that was already set to be upper mid range only.

They will need to be pretty agressive with pricing for this to have any chances which, I mean, it's great but it hurts to see Intel shoot themselves in the foot lol

 

[Chrispy_]

actually seems ok for my needs and send to hell jensen is priceless

I take it back. Quicksync has come a long way since I last tried it (Haswell-E).

TBH, it became redundant for me because while Intel were feeding prosumers shitty 6C/12T parts made from repurposed 1P Xeon dies, AMD succeeded that with 16-24 core Threadrippers. Suddenly highest-quality x264 software encodes were completing much faster than intel Quicksync.

Quicksync seems to be better than AMD at x264 and comparable to AMD at x265. I like that AMD's driver includes a decent consumer encoder utility negating the need for OBS but OBS is waaay more flexible for anything other than basic recording of a screen and AMD have always been shit there.

 

[dozenfury]

Time to market has been Intel's Achille's Heel for quite a while. They slept on the consumer cpu side for so long that they let AMD rise from the ashes and take a significant portion of market share. On the gpu side development has been so slow that they'll probably ending up missing out on the entire gpu gold rush by the time they get there. And Project Endgame? C'mon...that almost writes the "Intel shutting down Project Endgame" headline itself before it happens.

 

[MikeMurphy]

I've always had a good experience with Intel GPUs, except for some wanting in terms of performance.

I have faith in Intel to deliver a competitive dedicated GPU product and experience.

 

[R0H1T]

And Project Endgame? C'mon...that almost writes the "Intel shutting down Project Endgame" headline itself before it happens.

Maybe they timed it with the last Avengers movie & then 10/7nm happened

 

[Zubasa]

I take it back. Quicksync has come a long way since I last tried it (Haswell-E).

Quicksync seems to be better than AMD at x264 and comparable to AMD at x265. I like that AMD's driver includes a decent consumer encoder utility negating the need for OBS but OBS is waaay more flexible for anything other than basic recording of a screen and AMD have always been shit there.

Yeah and funny enough the best CPU to pair with the 6500XT is an Intel CPU with an IGPU due to Quicksync and having pcie 4.0/5.0 unlike AMD's own APUs.

 

[Steevo]

I've always had a good experience with Intel GPUs, except for some wanting in terms of performance.

I have faith in Intel to deliver a competitive dedicated GPU product and experience.

Which GPUs specifically? The integrated shitty low power ones they have made that had to be supplemented with Nvidia dare they use AMD ones? Or the non-existing dedicated ones?

Please let me know what excel or PowerPoint worked so much better with Intel than anyone else? Which 1080 YouTube static chill mix looked better on Intel than a phone or entry level TV. I await your intriguing response with high hopes good sir or madam. I’m almost breathless with expectations on your profound revelation!

Yeah and funny enough the best CPU to pair with the 6500XT is an Intel CPU with an IGPU due to Quicksync and having pcie 4.0/5.0 unlike AMD's own APUs.

Cause PCIe 5 is so profound and liberating. Without it mere computing ceases to exist with the millions, nay, BILLIONS of PCIe 5 devices capable of such computing magnitude.

Intel shills are such a bunch of twats. Go home, your poor attempt is poor. Take the pittance you have been paid and leave.

 

[Zubasa]

Cause PCIe 5 is so profound and liberating. Without it mere computing ceases to exist with the millions, nay, BILLIONS of PCIe 5 devices capable of such computing magnitude.

Intel shills are such a bunch of twats. Go home, your poor attempt is poor. Take the pittance you have been paid and leave.

Calm down dude. It is not my decision that AMD stayed on PCIE 3.0 on their Zen3 APUs.
Then they released the 6500XT that actually loses significant performance dropping from PCIE 4.0 to 3.0 due to the limited vram and x4 link width.
On top of that the 6500XT lacks an hardware encoder, which is arguable more useful for a low end card, since these will most likely be paired with a low end CPU, and Quicksync solves that.

It is almost as if their GPU and CPU divisions were trying to sabotage each other in this case.
It is the combination of 4GB vram and x4 link that makes it problematic, if it had either more vram or x8 link width it wouldn't have been an embarrassing launch.

 

[Chrispy_]

Calm down dude. It is not my decision that AMD stayed on PCIE 3.0 on their Zen3 APUs.
Then they released the 6500XT that actually loses significant performance dropping from PCIE 4.0 to 3.0 due to the limited vram and x4 link width.
On top of that the 6500XT lacks an hardware encoder, which is arguable more useful for a low end card, since these will most likely be paired with a low end CPU, and Quicksync solves that.
View attachment 237109
It is almost as if their GPU and CPU divisions were trying to sabotage each other in this case.
It is the combination of 4GB vram and x4 link that makes it problematic, if it had either more vram or x8 link width it wouldn't have been an embarrassing launch.

The 6500XT was never supposed to be a desktop GPU, the lack of PCIe lanes, hardware encoder, and display outputs are specifically because it was supposed to be paired with mobile Ryzen APUs that already had direct CPU access over infinity fabric instead of PCIe, a full hardware encoder already (no need waste silicon area and power budget on duplicating that!), and two more outputs bringing the paired combination to a respectable four.

The messed up market condition and a need to get something at sub-$200 meant that this auxiliary laptop bolt-on got shoved onto a PCB and turned into a dGPU even though I suspect nobody at AMD was particulary proud of it.

The GPU and CPU divisions are clearly talking to each other and in this case the 6500XT was launched because AMD had no other silicon in production that could compete in the entry segment that Nvidia were entering with the 3050. Better to release a complete turd and expect poor sales than to completely hand the entire market segment to Nvidia on a golden platter - even poor sales are better than none.

Yeah and funny enough the best CPU to pair with the 6500XT is an Intel CPU with an IGPU due to Quicksync and having pcie 4.0/5.0 unlike AMD's own APUs.

The original plan for Navi24 was (and perhaps still is) to work in Crossfire. Presumably we'll see thin-and-light laptops next that have 680M IGP and a 6500XT paired up.

 

[MikeMurphy]

Which GPUs specifically? The integrated shitty low power ones they have made that had to be supplemented with Nvidia dare they use AMD ones? Or the non-existing dedicated ones?

Please let me know what excel or PowerPoint worked so much better with Intel than anyone else? Which 1080 YouTube static chill mix looked better on Intel than a phone or entry level TV. I await your intriguing response with high hopes good sir or madam. I’m almost breathless with expectations on your profound revelation!



Cause PCIe 5 is so profound and liberating. Without it mere computing ceases to exist with the millions, nay, BILLIONS of PCIe 5 devices capable of such computing magnitude.

Intel shills are such a bunch of twats. Go home, your poor attempt is poor. Take the pittance you have been paid and leave.

Most people relying on integrated don't need high performance. When you consider the performance/video decode and encode to power consumption and die size it's actually quite an engineering feat.

I've never once had driver issues with my Intel iGPUs.

Intel is certainly capable of scaling this up with a new GPU architecture.

 

[Zubasa]

Cause PCIe 5 is so profound and liberating. Without it mere computing ceases to exist with the millions, nay, BILLIONS of PCIe 5 devices capable of such computing magnitude.