Toward the end of 2024, Intel will update its client processor product stack with the introduction of the new "Arrow Lake" microarchitecture targeting both the desktop and mobile segments. On the desktop side of things, this will herald the new Socket LGA1851 with more SoC connectivity being shifted to the processor; and on the mobile side of things, there will be a much-needed increase in CPU core counts form the current 6P+8E+2LP. This low maximum core-count for "Meteor Lake" is the reason why Intel couldn't debut it on the desktop platform, and couldn't use it to power enthusiast HX-segment mobile processors, either—it had to tap into "Raptor Lake Refresh," and use the older 14th Gen Core nomenclature one last time.

All hopes are now pinned on "Arrow Lake," which could make up Intel's second Core Ultra mobile lineup; its first desktop Core Ultra, and possibly push "Meteor Lake" to the non-Ultra tier. "Arrow Lake" carries forward the Xe-LPG graphics architecture for the iGPU that Intel debuted with "Meteor Lake," but there's a key difference between the desktop- and mobile "Arrow Lake" chips concerning this iGPU, and it has not just to do with the Xe core counts. It turns out, that while the desktop "Arrow Lake-S" processor comes with an iGPU based on the Xe-LPG graphics architecture; the mobile "Arrow Lake" chips spanning the U-, P-, and H-segments will use a newer version of this architecture, called the Xe-LPG+.

Intel in an interview with Hardwareluxx shed more light on its second generation Xe graphics architecture, codenamed "Battlemage." There will be two key variants of "Battlemage,"—Xe2-LPG and Xe2-HPG. The Xe2-LPG (low-power graphics) architecture is a slimmed-down derivative of "Battlemage" that's optimized for low-power. It is meant for iGPUs (integrated graphics), particularly upcoming "disaggregated" Intel Core processors in which the iGPU exists on Graphics Tiles (chiplets). The iGPU powering the upcoming Core "Meteor Lake" processor is rumored to meet the full DirectX 12 Ultimate feature-set (something Xe-LP doesn't), and so it's likely that Xe2-LPG is getting its first outing with that processor. The Xe2-HPG (high performance graphics) architecture is designed squarely for discrete GPUs—either desktop graphics cards, or mobile discrete GPUs hardwired into laptops.

In the interview, Intel talked about how its first-generation Xe graphics IP had at least four separate product verticals based on the scalability of the product, and the specific application (Xe-LP for iGPUs and tiny dGPUs, Xe-HPG for client- and pro-vis discrete GPUs, Xe-HPC for scalar compute processors, and Xe-HP for data-center graphics). The company eventually axed Xe-HP as it felt the Xe-HPG and Xe-HPC architectures adequately addressed this segment. With AXG (accelerated compute group) being split up between the CCG (client computing group) and DCG (data-center group); Xe2-LPG and Xe2-HPG will be developed primarily under CCG, with a client and pro-visualization focus; while Xe-HPC will be developed as a scalar-compute architecture by DCG, which effectively leaves the Intel Arc Graphics team with just two verticals—to deliver a feature-rich iGPU for its next-generation Core processors, and a performance discrete GPU lineup so it can eat away market-share from NVIDIA and AMD—hopefully with better time-to-market.

In advance of Supercomputing '22 in Dallas, Intel Corporation has introduced the Intel Max Series product family with two leading-edge products for high performance computing (HPC) and artificial intelligence (AI): Intel Xeon CPU Max Series (code-named Sapphire Rapids HBM) and Intel Data Center GPU Max Series (code-named Ponte Vecchio). The new products will power the upcoming Aurora supercomputer at Argonne National Laboratory, with updates on its deployment shared today.

The Xeon Max CPU is the first and only x86-based processor with high bandwidth memory, accelerating many HPC workloads without the need for code changes. The Max Series GPU is Intel's highest density processor, packing over 100 billion transistors into a 47-tile package with up to 128 gigabytes (GB) of high bandwidth memory. The oneAPI open software ecosystem provides a single programming environment for both new processors. Intel's 2023 oneAPI and AI tools will deliver capabilities to enable the Intel Max Series products' advanced features.

The U.S. Government has imposed restrictions on the export of AI compute GPUs to China and Russia without Government-authorization in the form of a waiver or a license. This impacts sales of products such as the NVIDIA A100, H100; AMD Instinct MI100, MI200; and the upcoming Intel "Ponte Vecchio," among others. The restrictions came to light when NVIDIA on Wednesday disclosed that it has received a Government notification about licensing requirements for export of its AI compute GPUs to Russia and China.

The notification doesn't specify the A100 and H100 by name, but defines AI inference performance thresholds to meet the licensing requirements. The Government wouldn't single out NVIDIA, and so competing products such as the AMD MI200 and the upcoming Intel Xe-HP "Ponte Vecchio" would fall within these restrictions. For NVIDIA, this is impacts $400 million in TAM, unless the Government licenses specific Russian and Chinese customers to purchase these GPUs from NVIDIA. Such trade restrictions usually come with riders to prevent resale or transshipment by companies outside the restricted region (eg: a distributor in a third waived country importing these chips in bulk and reselling them to these countries).

With AMD and NVIDIA launching its next-generation HPC compute architectures, "Hopper" and CDNA2, it began seeming like Intel's ambitious "Ponte Vecchio" accelerator based on the Xe-HP architecture, has missed the time-to-market bus. Intel doesn't think so, and in its Hot Chips 34 presentation, disclosed some of the first detailed performance claims that—at least on paper—put the "Hopper" H100 accelerator's published compute performance numbers to shame. We already had some idea of how Ponte Vecchio would perform this spring, at Intel's ISC'22 presentation, but the company hadn't finalized the product's power and thermal characteristics, which are determined by its clock-speed and boosting behavior. Team blue claims to have gotten over the final development hurdles, and is ready with some big numbers.

Intel claims that in classic FP32 (single-precision) and FP64 (double-precision) floating-point tests, its silicon is highly competitive with the H100 "Hopper," with the company claiming 52 TFLOP/s FP32 for the "Ponte Vecchio," compared to 60 TFLOP/s for the H100; and a significantly higher 52 TFLOP/s FP64 for the "Ponte Vecchio," compared to 30 TFLOP/s for the H100. This has to do with the SIMD units of the Xe-HP architecture all being natively capable of double-precision floating-point operations; whereas NVIDIA's architecture typically relies on FP64-specialized streaming multiprocessors.

So you thought your Arc A380 graphics card, or the Gen12 Xe iGPU in your 12th Gen Core processors were good enough to munch through your older games from the 2000s and early 2010s? Not so fast. Intel Graphics states that the Xe-LP and Xe-HPG graphics architectures, which power the Gen12 Iris Xe iGPUs and the new Arc "Alchemist" graphics cards, lack native support for the DirectX 9 graphics API. The two rely on API translation such as Microsoft D3D9On12, which attempts to translate D3D9 API commands to D3D12, which the drivers can recognize.

Older graphics architectures such as the Gen11 powering "Ice Lake," and Gen9.5 found in all "Skylake" derivatives, feature native support for DirectX 9, however when paired with Arc "Alchemist" graphics cards, the drivers are designed to engage D3D9On12 to accommodate the discrete GPU, unless the dGPU is disabled. API translation can be unreliable and buggy, and Intel points you to Microsoft and the game developers for support, Intel Graphics won't be providing any.

At Intel's 2022 investor meeting, the company has presented a technology roadmap update to give its clients an insight into what is to come. Today, team blue announced one of the first discrete data-centric graphics cards in the lineup, codenamed Arctic Sound-M GPU. Based on the DG2 Xe-HPG variation of Intel Xe GPUs, Arctic Sound-M is the company's first design to enter the data center space. The DG2 GPU features 512 Execution Units (EUs), which get passive cooling from the single-slot design of Arctic Sound's heatsink, envisioned for data center enclosures with external airflow.

One of the most significant selling points that Intel advertises is support for hardware-based AV1 encoding standard. This feature allows the card to achieve a 30% greater bandwidth, and it is the main differentiator between consumer-oriented Arc Alchemist GPUs and itself. The card is powered by PCIe power and an 8-pin EPS power connector. Arctic Sound-M is already sampling to select customers and it will become available in the middle of 2022.Below is Intel's teaser video.

Intel's Arc Alchemist discrete lineup of graphics cards is scheduled for launch this quarter. We are getting some performance benchmarks of the DG2-512EU silicon, representing the top-end Xe-HPG configuration. Thanks to a discovery of a famous hardware leaker TUM_APISAK, we have a measurement performed in the SiSoftware database that shows Intel's Arc Alchemist GPU with 4096 cores and, according to the report from the benchmark, just 12.8 GB of GDDR6 VRAM. This is just an error on the report, as this GPU SKU should be coupled with 16 GB of GDDR6 VRAM. The card was reportedly running at 2.1 GHz frequency. However, we don't know if this represents base or boost speeds.

When it comes to actual performance, the DG2-512EU GPU managed to score 9017.52 Mpix/s, while something like NVIDIA GeForce RTX 3070 Ti managed to get 8369.51 Mpix/s in the same test group. Comparing these two cards in floating-point operations, Intel has an advantage in half-float, double-float, and quad-float tests, while NVIDIA manages to hold the single-float crown. This represents a 7% advantage for Intel's GPU, meaning that Arc Alchemist has the potential for standing up against NVIDIA's offerings.

Intel has dropped plans to build Xe-HP server GPUs commercially. This line of products would have powered cloud-based graphics rendering instances, for cloud-gaming or cloud-rendering applications. An announcement to this effect came from Raja Koduri, overseeing the development and monetization of Xe. Koduri stated that Xe-HP based instances were originally set up to power Intel's oneAPI devcloud as a software development vehicle for oneAPI and the upcoming Aurora supercomputer of the Argonne National Laboratory.

The company will now focus on Xe as a compute accelerator in the form of Xe-HPC "Ponte Vecchio," and discrete graphics in the client segment, leveraging the Xe-HPG graphics architecture. The smallest derivatives, the Xe-LP, powers integrated graphics solutions found in the company's Core processors (11th Gen and later). Back in the August 2021 Architecture Day presentation, Intel's technical brief for Xe HPC revealed that the silicon itself features certain on-die hardware relevant to graphics rendering (more here). This would have gone on to power the Xe-HP server GPU solutions.

Intel's Arc Alchemist lineup of graphics cards, based on Xe-HPG GPU configuration, is nearing the launch. With the current situation with AMD and NVIDIA GPUs outputting graphics card prices over the default MSRP, we wonder how Intel would place pricing of its upcoming GPUs and fit inside the market. And today, we got the first round of speculations based on Intel's Arc Alchemist GPU giveaway called Xe-HPG Scavenger Hunt. There are two principal bundles: one worth $900 that includes Intel Arc merchandise, Xbox Game Pass PC for six months, Intel Premium Arc Alchemist graphics card, and one worth $700 that consists of three months of Xbox Game Pass PC, Intel Arc merchandise, and Intel Performance Arc Alchemist graphics card.

According to some preliminary calculations from Tom's Hardware, we assume that with the $900 bundle containing one Premium Arc Alchemist GPU and other prizes, the card will cost as much as $825 when all things get removed. Going down the ladder, Intel has paired a bundle worth $700 with a Performance Arc GPU, which is roughly worth $650 on its own. It indicates that the two Intel Performance and Premium Arch Alchemist graphics cards are respectfully worth $650 and $825. What will the final pricing look like? We don't know. However, we assume that it could be very similar to this. For more information we have to wait for the official launch.

At the 2021 International Supercomputing Conference (ISC) Intel is showcasing how the company is extending its lead in high performance computing (HPC) with a range of technology disclosures, partnerships and customer adoptions. Intel processors are the most widely deployed compute architecture in the world's supercomputers, enabling global medical discoveries and scientific breakthroughs. Intel is announcing advances in its Xeon processor for HPC and AI as well as innovations in memory, software, exascale-class storage, and networking technologies for a range of HPC use cases.

"To maximize HPC performance we must leverage all the computer resources and technology advancements available to us," said Trish Damkroger, vice president and general manager of High Performance Computing at Intel. "Intel is the driving force behind the industry's move toward exascale computing, and the advancements we're delivering with our CPUs, XPUs, oneAPI Toolkits, exascale-class DAOS storage, and high-speed networking are pushing us closer toward that realization."

Yesterday, we got information that Intel's upcoming DG2 discrete graphics card is "right around the corner". That means that we are inching closer to the launch of Intel's discrete GPU offerings, and we are going to get another major player in the current GPU market duopoly. Today, however, we are in luck because Igor from Igor's LAB has managed to get ahold of the specifications of Intel's Xe-HPG DG2 graphics card. For starters, it is important to note that DG2 GPU will first come to laptops later this year. More precisely, laptops powered by Alder Lake-P processors will get paired with DG2 discrete GPU in the second half of 2021. The CPU and GPU will connect using the PCIe 4.0 x12 link as shown in the diagram below, where the GPU is paired with the Tiger Lake-H processor. The GPU has its subsystem that handles the IO as well.

We have recently received pictures of any early engineering sample of Intel's upcoming DG2 GPU from YouTuber Moore's Law is Dead. The card features 512 Execution Units and will be the flagship model for Intel's upcoming Xe-HPG lineup reportedly targeting performance between the RTX 3070 and RTX 3080. The final product is rumored to feature a base clock of 2.2 GHz along with 16 GB GDDR6 memory and a 256-bit bus. The sample has a TDP of 275 W with 8 + 6 pin power connectors up from original targets of 225 W - 250 W.

The report also notes that Intel is still deciding between three cooler designs with the finished card potentially featuring a white shroud. Intel also appears to be working on a NVIDIA DLSS/AMD FidelityFX Super Resolution competitor codenamed XeSS which confirms support for hardware-accelerated raytracing and resolution upscaling tech. The card is unlikely to launch until Q4 2021 with wider availability in 2022, lower end 128 EU, and 256 EU cards will follow shortly afterward. The full report can be viewed below.

Intel is still in the work of testing and certification for their more complex graphics products based on their Xe microarchitectures, and that means that some tests are being done in well-known benchmarking platforms. Case in point, an Intel Xe GPU with a reported 128 EUs (1024 shading units), 3 GB of memory, and a clockspeed of 1.4 GHz ran through Geekbench, where it scored an... interesting 9,311 points in the OpenCL test.

This is more likely than not an engineering sample, considering that Intel's Iris Xe MAX wrapped in its Tiger Lake package can score up to 23,000 points. It is currently unclear if this particular Xe manifestation is running on Intel's Xe-LP or Xe-HPG architecture. This might be Intel's DG-2 product, which offers higher performance than their DG-1 discrete graphics card that is only available for system integrators.

Raja Koduri senior vice president, chief architect, and general manager of Architecture, Graphics, and Software at Intel Corporation has recently teased Intel's upcoming Xe-HP accelerator alongside its in production HC3 XG310 server card. The HC3 solution was Intel's first Xe-based product utilizing the Xe-LP architecture. The Intel Xe-LP products are Intel's lowest power efficiency optimized Xe processors while the Xe-HP products should offer improved performance and scaling. The upcoming Xe-HP accelerator appears to be a single-slot passively card with a single 8-pin power connector. Raja Koduri expects developers will begin receiving Xe-HP, Xe-HPG, and Xe-HPC products in 2021. He also declared that we are in the GPU golden age with new launches from NVIDIA, AMD, Intel, and Apple.

Intel and Argonne National Laboratory are collaborating on the co-design and validation of exascale-class applications using graphics processing units (GPUs) based on Intel Xe-HP microarchitecture and Intel oneAPI toolkits. Developers at Argonne are tapping into Intel's latest programming environments for heterogeneous computing to ensure scientific applications are ready for the scale and architecture of the Aurora supercomputer at deployment.

"Our close collaboration with Argonne is enabling us to make tremendous progress on Aurora, as we seek to bring exascale leadership to the United States. Providing developers early access to hardware and software environments will help us jumpstart the path toward exascale so that researchers can quickly start taking advantage of the system's massive computational resources." -Trish Damkroger, Intel vice president and general manager of High Performance Computing.

Intel today announced key milestones in its multiyear journey to deliver a mix of architectures with a unified software experience. The company announced the gold release of Intel oneAPI toolkits coming in December, and new capabilities in its software stack as part of the Intel's combined hardware and software design approach. Intel also debuted its first discrete graphics processing unit (GPU) for the data center, Intel Server GPU, based on the Xe-LP microarchitecture and designed specifically for high-density, low-latency Android cloud gaming and media streaming.

"Today is a key moment in our ambitious oneAPI and XPU journey. With the gold release of our oneAPI toolkits, we have extended the developer experience from familiar CPU programming libraries and tools to include our vector-matrix-spatial architectures. We are also launching our first data center GPU based on Xe-LP microarchitecture focused on the fast-growing cloud gaming and media streaming segments," said Raja Koduri, Intel senior vice president, chief architect and general manager of Architecture, Graphics and Software.

Intel is preparing to flood the market with its Xe GPU lineup, covering the entire vector from low-end to high-end consumer graphics cards. Just a few days ago, the company has announced its Iris Xe MAX GPU, the first discrete GPU from Intel, aimed at 1080p gamer and content creators. However, that seems to be only the beginning of Intel's GPU plan and just a small piece of the entire lineup. Next year, the company is expected to launch two GPU families - Xe-HP and Xe-HPG. With the former being a data-centric GPU codenamed Arctic Sound, and the latter being a gaming-oriented GPU called DG2. Today, thanks to the GeekBench listing, we have some information on the Xe-HP GPU.

Being listed with 512 EUs (Execution Units), translating into 4096 shading units, the GPU is reportedly a Xe-HP variant codenamed "NEO Graphics". This is not the first time that the NEO graphics has been mentioned. Intel has called a processor Neo graphics before, on its Architecture day when the company was demonstrating the FP32 performance. The new GeekBench leak shows the GPU running at 1.15 GHz clock speed, where at the Architecture day the same GPU ran at 1.3 GHz frequency, indicating that this is only an engineering sample. The GPU ran the GeekBench'es OpenCL test and scored very low 25,475 points. Compared to NVIDIA's GeForce RTX 3070 GPU that scored 140,484, the Intel GPU is at least four times slower. That is possibly due to the non-optimization of the benchmark, which could greatly improve in the future. In the first picture below, this Xe-HP GPU would represent the single-tile design.