Super Micro Computer, Inc., a global leader in enterprise computing, storage, networking, and green computing technology, is announcing future Total IT Solutions for availability with Android Cloud Gaming and Media Processing & Delivery. These new solutions will incorporate the Intel Data Center GPU, codenamed Arctic Sound-M, and will be supported on several Supermicro servers. Supermicro solutions that will contain the Intel Data Center GPUs codenamed Arctic Sound-M, include the 4U 10x GPU server for transcoding and media delivery, the Supermicro BigTwin system with up to eight Intel Data Center GPUs, codenamed Arctic Sound-M in 2U for media processing applications, the Supermicro CloudDC server for edge AI inferencing, and the Supermicro 2U 2-Node server with three Intel Data Center GPUs, codenamed Arctic Sound-M per node, optimized for cloud gaming. Additional systems will be made available later this year.

"Supermicro will extend our media processing solutions by incorporating the Intel Data Center GPU," said Charles Liang, President, and CEO, Supermicro. "The new solutions will increase video stream rates and enable lower latency Android cloud gaming. As a result, Android cloud gaming performance and interactivity will increase dramatically with the Supermicro BigTwin systems, while media delivery and transcoding will show dramatic improvements with the new Intel Data Center GPUs. The solutions will expand our market-leading accelerated computing offerings, including everything from Media Processing & Delivery to Collaboration, and HPC."

GIGABYTE TECHNOLOGY Co. Ltd, the global leading brand of PC, launches GIGABYTE Gaming G5/G7 gaming laptops equipped with 10nm Intel 12th Gen Processor today. A laptop to meet the wide range of needs in multitasking, gaming, and entertainment, with 12th Gen Intel Core i5-12500H laptop CPU, which is comprised of 12-core, 16 threads, and a maximum clock rate of 4.5GHz, to meet the needs for telecommuting and online classes, the purchase of high-performance laptops has been made easier with the adoption of Core i5-12500H, the Core i5 processor is powerful enough to effortlessly handle users' routines. Equipped with the graphics cards of NVIDIA GeForce RTX 30 Series, also introduces MUX switch graphics card switching technology, discrete GPU can be directly output to the display with just one click, which can easily improve the game performance and increase the frame rate in fierce game battles. For offering authentic gaming specifications and flexible expandability of hardware. The series can satisfy the user's needs for playing multiple roles in life.

It's not only Intel that has been showing off new graphics cards recently, as Chinese company Glenfly has revealed more details about its Arise-GT10C0 graphics card. To be clear from the start, this is not a graphics card for gamers, but rather for the PRC government and its computers, as the nation is trying to become self-sufficient when it comes to computer hardware for its government agencies and other government backed organisations. The 28 nm GPU has a clock speed of a whopping 500 MHz and delivers 1.5 TFLOPs of FP32 performance, which places it firmly in yesteryear's performance category. Glenfly claims support for up to 4K resolution, althought this is most likely only for desktop use.

The GPU is paired with 2 or 4 GB of DDR4 memory with a clock speed of 1200 MHz, using either a 64 or a 128 bit memory interface. The actual cards have a PCIe 3.0 x8 interface and have support for unspecified HDMI and DP interfaces, as well as D-Sub VGA ports. Driver support includes DirectX 11, OpenGL 4.5 and OpenCL 1.2. The GPU is also said to have hardware offload support for HEVC and H.264 hardware encoding, as well as decoding for both formats, plus most other common video formats, although, oddly enough, support for AVS, which is China's homebrewed video codec, is missing. OS support includes various Chinese flavours of Linux, Ubuntu and Windows according to Glenfly and outside of the x86 processor world, MIPS and arm based processors are said be supported.

Graphics card prices continue to fall in China with NVIDIA & AMD holding excessive stock of RTX 3000/RX 6000 cards without enough consumers interested in buying them. The companies have resisted pressure to lower the official MSRP of these cards with most retailers now offering discounts of between 5% and 30%. The largest drops are with flagship cards such as the NVIDIA RTX 3090 Ti which is now available for 9499 RMB (1415 USD) which is 38% below MSRP while the AMD Radeon RX 6900 XT is 4999 RMB (744 USD) a 37.5% reduction. The prices for all graphics cards are now below MSRP with drops amounting to 20% on average for NVIDIA and 19% for AMD.

When designing integrated circuits, engineers aim to produce an efficient design that is easier to manufacture. If they manage to keep the circuit size down, the economics of manufacturing that circuit is also going down. NVIDIA has posted on its technical blog a technique where the company uses an artificial intelligence model called PrefixRL. Using deep reinforcement learning, NVIDIA uses the PrefixRL model to outperform traditional EDA (Electronics Design Automation) tools from major vendors such as Cadence, Synopsys, or Siemens/Mentor. EDA vendors usually implement their in-house AI solution to silicon placement and routing (PnR); however, NVIDIA's PrefixRL solution seems to be doing wonders in the company's workflow.

Creating a deep reinforcement learning model that aims to keep the latency the same as the EDA PnR attempt while achieving a smaller die area is the goal of PrefixRL. According to the technical blog, the latest Hopper H100 GPU architecture uses 13,000 instances of arithmetic circuits that the PrefixRL AI model designed. NVIDIA produced a model that outputs a 25% smaller circuit than comparable EDA output. This is all while achieving similar or better latency. Below, you can compare a 64-bit adder design made by PrefixRL and the same design made by an industry-leading EDA tool.

NVIDIA's next generation of graphics cards, codenamed RTX 40 series, Ada Lovelace, is expected to arrive sometime in October. However, the latest information from the YouTube channel "Moore's Law Is Dead" suggests that NVIDIA could postpone the arrival of the new GPU generation to December. Why, you might be wondering? The report claims that the current GPU market is flooded with used GeForce RTX 30 series GPUs. Thus, NVIDIA could postpone the availability of the latest GPUs to keep the demand high and ensure that the market is searching for additional graphics cards.

Retailers are experiencing smaller demand as the used GPU market is full of devices used for cryptocurrency mining, and the recent crypto crash has helped the situation. What we could see is NVIDIA announcing Ada Lovelace GPUs in October, with availability arriving later in December. Of course, these are just the current industry rumors, and we are yet to see how the market and NVIDIA will respond.

According to TrendForce investigations, foundries have seen a wave of order cancellations with the first of these revisions originating from large-size Driver IC and TDDI, which rely on mainstream 0.1X μm and 55 nm processes, respectively. Although products such as MCU and PMIC were previously in short supply, foundries' capacity utilization rate remained roughly at full capacity through their adjustment of product mix. However, a recent wave cancellations have emerged for PMIC, CIS, and certain MCU and SoC orders. Although still dominated by consumer applications, foundries are beginning to feel the strain of the copious order cancellations from customers and capacity utilization rate has officially declined.

Looking at trends in 2H22, TrendForce indicates, in addition to no relief from the sustained downgrade of driver IC demand, inventory adjustment has begun for smartphones, PCs, and TV-related peripheral components such as SoCs, CIS, and PMICs, and companies are beginning to curtail their wafer input plans with foundries. This phenomenon of order cancellations is occurring simultaneously in 8-inch and 12-inch fabs at nodes including 0.1X μm, 90/55 nm, and 40/28 nm. Not even the advanced 7/6 nm processes are immune.

Intel is preparing to launch its next-generation desktop platform codenamed Rocket Lake-S. According to the presentation held by Intel today in Shenzen, China, we have official information regarding some of the platform features that Raptor Lake is bringing. Starting with memory support, Raptor Lake is still carrying the transitional DDR4 and DDR5 support, as the full swing towards DDR5 is still in progress. Unlike the previous generation Alder Lake, which brought DDR5-4800 support, Raptor Lake's integrated memory controller can drive DDR5 modules with a 5600 MT/s configuration. As DDR4 support remains, it is limited to 3200 MT/s speed.

Interesting information from the leaked slide points out that support for CPU-attached NVMe storage remains PCIe Gen4. While AMD will provide an AM5 socket with CPU-attached NMVe storage on PCIe Gen5 protocol, Intel is taking a step back and holding on to Gen4. The CPU is outputting 16 PCIe Gen5 lanes on its own. Motherboard vendors for the upcoming 700-series boards for Raptor Lake can still provide a PCIe Gen5 NVMe slot; however, it will have to subtract eight Gen5 lanes from the PCI Express Graphics (PEG) slot and route them to NVMe storage. As our testing shows, this will affect GPU's performance by a few percent. AMD's upcoming AM5 platform has no such issues, as the CPU provides both the PEG and CPU-attached NVMe storage with sufficient PCIe Gen5 bandwidth.

AMD's RDNA3 graphics IP is just around the corner, and we are hearing more information about the upcoming architecture. Historically, as GPUs advance, it is not unusual for companies to add dedicated hardware blocks to accelerate a specific task. Today, AMD engineers have updated the backend of the LLVM compiler to include a new instruction called Wave Matrix Multiply-Accumulate (WMMA). This instruction will be present on GFX11, which is the RDNA3 GPU architecture. With WMMA, AMD will offer support for processing 16x16x16 size tensors in FP16 and BF16 precision formats. With these instructions, AMD is adding new arrangements to support the processing of matrix multiply-accumulate operations. This is closely mimicking the work NVIDIA is doing with Tensor Cores.

AMD ROCm 5.2 API update lists the use case for this type of instruction, which you can see below:

rocWMMA provides a C++ API to facilitate breaking down matrix multiply accumulate problems into fragments and using them in block-wise operations that are distributed in parallel across GPU wavefronts. The API is a header library of GPU device code, meaning matrix core acceleration may be compiled directly into your kernel device code. This can benefit from compiler optimization in the generation of kernel assembly and does not incur additional overhead costs of linking to external runtime libraries or having to launch separate kernels.

rocWMMA is released as a header library and includes test and sample projects to validate and illustrate example usages of the C++ API. GEMM matrix multiplication is used as primary validation given the heavy precedent for the library. However, the usage portfolio is growing significantly and demonstrates different ways rocWMMA may be consumed.

This time last year, I wrote about how digital experiences had never been more important, from personal to business devices - they helped us stay connected and entertained at a time when we needed it most. Compute continues to define our experiences in the modern world, and now these experiences are becoming even more visual.

Smartphones are at the center of our connected lives. From gaming to productivity, through video calling, social media or virtual environments, it is the device that provides us the connection to everyone and everything, in real time. For developers, making these immersive real-time 3D experiences even more compelling and engaging requires more performance. Arm sets the standard for performance and efficient compute, and our latest suite of compute solutions for consumer devices will continue to raise the threshold of what's possible in the mobile market, shaping the visual experiences of tomorrow.

AMD is preparing to update its mobile sector with the latest IP in the form of Zen4 CPU cores and RDNA3 graphics. According to Red Gaming Tech, we have specifications of upcoming processor families. First, we have AMD Dragon Range mobile processors representing a downsized Raphael design for laptops. Carrying Zen4 CPU cores and RDNA2 integrated graphics, these processors are meant to power high-performance laptops with up to 16 cores and 32 threads. Being a direct competitor to Intel's Alder Lake-HX, these processors also carry an interesting naming convention. The available SKUs include AMD Ryzen 5 7600HX, Ryzen 7 7800HX, Ryzen 9 7900HX, and Ryzen 9 7980HX design with a massive 16-core configuration. These CPUs are envisioned to run along with more powerful dedicated graphics, with clock speeds of 4.8-5.0+ GHz.

Next, we have AMD Phoenix processors, which take Dragon Range's design to a higher level thanks to the newer graphics IP. Having Zen4 cores, Phoenix processors carry upgraded RDNA3 graphics chips to provide a performance level similar to NVIDIA's GeForce RTX 3060 Max-Q SKU, all in one package. These APUs will come in four initial configurations: Ryzen 5 7600HS, Ryzen 7 7800HS, Ryzen 9 7900HS, and Ryzen 9 7980HS. While maxing out at eight cores, these APUs will compensate with additional GPU compute units with a modular chiplet design. AMD Phoenix is set to become AMD's first chiplet design launching for the laptop market, and we can expect more details as we approach the launch date.

Intel's Arc Alchemist graphics cards launched in laptop/mobile space, and everyone is wondering just how well the first generation of discrete graphics performs in actual, GPU-accelerated workloads. Tellusim Technologies, a software company located in San Diego, has managed to get ahold of a laptop featuring an Intel Arc A370M mobile graphics card and benchmark it against other competing solutions. Instead of using Vulkan API, the team decided to use D3D12 API for tests, as the Vulkan usually produces lower results on the new 12th generation graphics. With the 30.0.101.1736 driver version, this GPU was mainly tested in the standard GPU working environment like triangles and batches. Meshlet size is set to 69/169, and the job is as big as 262K Meshlets. The total amount of geometry is 20 million vertices and 40 million triangles per frame.

Using the tests such as Single DIP (drawing 81 instances with u32 indices without going to Meshlet level), Mesh Indexing (Mesh Shader emulation), MDI/ICB (Multi-Draw Indirect or Indirect Command Buffer), Mesh Shader (Mesh Shaders rendering mode) and Compute Shader (Compute Shader rasterization), the Arc GPU produced some exciting numbers, measured in millions or billions of triangles. Below, you can see the results of these tests.

Graphics card prices have been on a steady decline in the past few months, following their peak in May of last year when we saw double and triple pricing compared to the baseline MSRP value. According to the 3DCenter.org report, which tracks graphics card prices in Germany and Austria, we have information that AMD GPU prices have dipped below MSRP, while NVIDIA GPUs are very close to baseline listed prices. The report tracks Ethereum mining profitability and displays it in the yellow line. As the line is declining, so are the GPU prices. For AMD, the prices are now 8% below the 100% of MSRP. At 92%, consumers can find AMD GPUs at a slight discount. While AMD cards are slightly cheaper, NVIDIA GPUs are now at 102% of the MSRP, the lowest price point since the launch.

Rumors of NVIDIA's upcoming Ada Lovelace graphics cards keep appearing. With every new update, it seems like the total power consumption is getting bigger, and today we are getting information about different SKUs, including mobile and desktop variants. According to a well-known leaker, kopite7kimi, we have information about the power limits of the upcoming GPUs. The new RTX 40 series GPUs will feature a few initial SKUs: AD102, AD103, AD104, and AD106. Every SKU, except the top AD102, will be available as well. The first in line, AD102, is the most power-hungry SKU with a maximum power limit rating of 800 Watts. This will require multiple power connectors and a very beefy cooling solution to keep it running.

Going down the stack, we have an AD103 SKU limited to 450 Watts on desktop and 175 Watts on mobile. The AD104 chip is limited to 400 Watts on desktop, while the mobile version is still 175 Watts. Additionally, the AD106 SKU is limited to 260 Watts on desktop and 140 Watts on mobile.

The recently announced Apple M2 processor which is set to feature in the new MacBook Air and 13-inch MacBook Pro models has been benchmarked. The processor appeared in numerous Geekbench 5 CPU & GPU tests where the chip scored a maximum single-core result of 1919 points and 8928 points in multi-core representing an 11% and 18% CPU performance improvement respectively from the M1. The chip brings significant GPU performance increases achieving a Geekbench Metal score of 30627 points which is a ~42% increase from the M1 partially due to a larger 10-core GPU compared to the 8-core GPU on the M1. These initial numbers largely align with claims from Apple of an 18% CPU and 35% GPU improvement over the original M1.

AMD is planning to debut its next-generation RDNA3 graphics architecture with the Radeon RX 7000 series desktop graphics cards, some time in late-October or early-November, 2022. This, according to Greymon55, a reliable source with AMD and NVIDIA leaks. We had known about a late-2022 debut for AMD's next-gen graphics, but now we have a finer timeline.

AMD claims that RDNA3 will repeat the feat of over 50 percent generational performance/Watt gains that RDNA2 had over RDNA. The next-generation GPUs will be built on the TSMC N5 (5 nm EUV) silicon fabrication process, and debut a multi-chip module design similar to AMD's processors. The logic dies with the GPU's SIMD components will be built on the most advanced node, while the I/O and display/media accelerators will be located in separate dies that can make do on a slightly older node.

A part of the process of building a graphics card is designing compatibility to execute the latest graphics APIs like DirectX, OpenGL, and Vulkan. Today, we have confirmation that Intel's Arc Alchemist discrete graphics cards will be compatible with Vulkan's latest iteration - version 1.3. In January, Khronos, the team behind Vulkan API, released their regular two-year update to the standard. Graphics card vendors like NVIDIA and AMD announced support immediately with their drivers. Today, the Khronos website officially lists Intel Arc Alchemist mobile graphics cards as compatible with Vulkan 1.3 with Intel Arc A770M, A730M, A550M, A370M, and A350M GPUs.

At the time of writing, there is no official announcement for the desktop cards yet. However, given that the mobile SKUs are supporting the latest standard, it is extremely likely that the desktop variants will also carry the same level of support.

Based on a report in the Taiwanese media, AMD is quickly becoming a key customer for TSMC and is expected to become its third largest customer in 2023. This is partially due to new orders that AMD has placed with TSMC for its 5 nm node. AMD is said to become TSMC's single largest customer for its 5 nm node in 2023, although it's not clear from the report how large of a share of the 5 nm node AMD will have.

The additional orders are said to be related to AMD's Zen 4 based processors, as well as its upcoming RDNA3 based GPUs. AMD is expected to be reaching a production volume of some 20,000 wafers in the fourth quarter of 2022, although there's no mention of what's expected in 2023. Considering most of AMD's products for the next year or two will all be based on TSMC's 5 nm node, this shouldn't come as a huge surprise though, as AMD has a wide range of new CPU and GPU products coming.

Jon Peddie Research reports that the global PC-based graphics processor units (GPU) market reached 96 million units in Q1'22 and PC GPUs shipments decreased 6.2% due to disturbances in China, Ukraine, and the pullback from the lockdown elsewhere. However, the fundamentals of the GPU and PC market are solid over the long term, JPR predicts GPUs will have a compound annual growth rate of 6.3% during 2022-2026 and reach an installed base of 3.3 million units at the end of the forecast period. Over the next five years, the penetration of discrete GPUs (dGPU) in the PC market will grow to reach a level of 46%.

AMD's overall market share percentage from last quarter increased 0.7%, Intel's market share decreased by -2.4%, and Nvidia's market share increased 1.69%, as indicated in the following chart.

Supercomputing game has been chasing various barriers over the years. This has included MegaFLOP, GigaFLOP, TeraFLOP, PetaFLOP, and now ExaFLOP computing. Today, we are witnessing for the first time an introduction of an Exascale-level machine contained at Oak Ridge National Laboratory. Called the Frontier, this system is not really new. We have known about its upcoming features for months now. What is new is the fact that it was completed and is successfully running at ORNL's facilities. Based on the HPE Cray EX235a architecture, the system uses 3rd Gen AMD EPYC 64-core processors with a 2 GHz frequency. In total, the system has 8,730,112 cores that work in conjunction with AMD Instinct MI250X GPUs.

As of today's TOP500 supercomputers list, the system is overtaking Fugaku's spot to become the fastest supercomputer on the planet. Delivering a sustained HPL (High-Performance Linpack) score of 1.102 Exaflop/s, it features a 52.23 GigaFLOPs/watt power efficiency rating. In the HPL-AI metric, dedicated to measuring the system's AI capabilities, the Frontier machine can output 6.86 exaFLOPs at reduced precisions. This alone is, of course, not a capable metric for Exascale machines as AI works with INT8/FP16/FP32 formats, while the official results are measured in FP64 double-precision form. Fugaku, the previous number one, scores about 2 ExaFLOPs in HPL-AI while delivering "only" 442 PetaFlop/s in HPL FP64 benchmarks.

AMD's upcoming Radeon RX 7000 series of graphics cards based on the RDNA 3 architecture are supposed to feature next-generation protocols all over the board. Today, according to a patch committed to the Linux kernel, we have information about display output choices AMD will present to consumers in the upcoming products. According to a Twitter user @Kepler_L2, who discovered this patch, we know that AMD will bundle DisplayPort 2.0 technology with UHBR 20 transmission mode. The UHBR 20 standard can provide a maximum of 80 Gbps bi-directional bandwidth, representing the highest bandwidth in a display output connector currently available. With this technology, a sample RDNA 3 GPU could display 16K resolution with Display Stream Compression, 10K without compression, or two 8K HDR screens running at 120 Hz refresh rate. All of this will be handled by Display Controller Next (DCN) engine for media.

The availability of DisplayPort 2.0 capable monitors is a story of its own. VESA noted that they should come at the end of 2021; however, they got delayed due to the lack of devices supporting this output. Having AMD's RDNA 3 cards as the newest product to support these monitors, we would likely see the market adapt to demand and few available products as the transition to the latest standard is in the process.

Hot Chips 34, the upcoming semiconductor conference from Sunday, August 21 to Tuesday, August 23, 2022, will feature many significant contributions from folks like Intel, AMD, Tesla, and NVIDIA. Today, thanks to Intel's registration at the event, we discovered that the company would present its work on Meteor Lake and Arrow Lake processors with the novel Foveros 3D packaging. The all-virtual presentation from Intel will include talks about Ponte Vecchio GPU and its architecture, system, and software; Meteorlake and Arrowlake 3D Client Architecture Platform with Foveros; and some Xeon D and FPGA presentations. You can see the official website here for a complete list of upcoming talks.

As a little reminder, Meteor Lake is supposed to arrive next year, replacing the upcoming Raptor Lake design, and it has already ahs been pictured, which you can see below. The presentation will be recorded and all content posted on Hot Chips's website for non-attendees to catch up on.

Today AMD announced the Kria KR260 Robotics Starter Kit, the latest addition to the Kria portfolio of adaptive system-on-modules (SOMs) and developer kits. A scalable and out-of-the-box development platform for robotics, the Kria KR260 offers a seamless path to production deployment with the existing Kria K26 adaptive SOMs. With native ROS 2 support, the standard framework for robotics application development, and pre-built interfaces for robotics and industrial solutions, the new SOM starter kit enables rapid development of hardware-accelerated applications for robotics, machine vision and industrial communication and control.

"The Kria KR260 Robotics Starter Kits builds on the success of our Kria SOMs and KV260 Vision AI Starter Kit for AI and embedded developers, providing roboticists with a complete, out-of-the-box solution for this rapidly growing application space," said Chetan Khona, senior director of Industrial, Vision, Healthcare and Sciences Markets at AMD. "Roboticists will now be able to work in their standard development environment on a platform that has all the interfaces and capabilities needed to be up and running in less than an hour. The KR260 Starter Kit is an ideal platform to accelerate robotics innovation and easily take ideas to production at scale."

Frank Azor, Chief Architect of Gaming Solutions & Marketing at AMD, has posted an interesting slide on Twitter, claiming that AMD Radeon products possess higher FPS/$ value than NVIDIA's graphics offerings. According to the slide, AMD Radeon graphics cards are the best solutions for gamers looking at performance per dollar ratings and performance per watt. This means that AMD claims that Radeon products are inherently higher-value products than NVIDIA's offerings while also more efficient. As the chart shows, which you can see below, some AMD Radeon cards are offering up to 89% better FPS/$ value with up to 123% better FPS/Watt metric. This highest rating is dedicated to Radeon RX 6400 GPU; however, there are all GPUs included in comparison with up to the latest Radeon RX 6950 XT SKU.

Compared to TechPowerUp's own testing of AMD's Radeon cards and multiple reviews calculating the performance per dollar metric, we could not see numbers as high as AMD's. This means that AMD's marketing department probably uses a different selection of games that may perform better on AMD Radeon cards than NVIDIA GeForce RTX. Of course, as with any company marketing material, you should take it with a grain of salt, so please check some of our reviews for a non-biased comparison.

Ten years ago, in 2012, NVIDIA introduced its Kepler series of graphics cards based on the TSMC 28 nm node. Architecture has been supported for quite a while now by NVIDIA's drivers, and the last series to carry support was the 470 driver class. Today, NVIDIA pushed a security update in the form of a 473.47 WHQL driver that brings fixes to various CVE vulnerabilities that can cause anything from issues that may lead to denial of service, information disclosure, or data tampering. This driver version has no fixed matters and doesn't bring any additional features except the fix for vulnerabilities. With CVEs rated from 4.1 to 8.5, NVIDIA has fixed major issues bugging Kepler GPU users. With a high risk for code execution, denial of service, escalation of privileges, information disclosure, and data tampering, the 473.47 WHQL driver is another step for supporting Kepler architecture until 2024, when NVIDIA plans to drop the support for this architecture. Supported cards are GT 600, GT 700, GTX 600, GTX 700, Titan, Titan Black, and Titan Z.

The updated drivers are available for installation on NVIDIA's website and for users of TechPowerUp's NVCleanstall software.