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.

Today, Lenovo in partnership with Intel has announced that Leibniz Supercomputing Centre (LRZ) is building a supercomputer powered by Intel's next-generation technologies. Specifically, the supercomputer will use Intel's Sapphire Rapids CPUs in combination with the highly-teased Ponte Vecchio GPUs to power the applications running at Leibniz Supercomputing Centre. Along with the various processors, the LRZ will also deploy Intel Optane persistent memory to process the huge amount of data the LRZ has and is producing. The integration of HPC and AI processing will be enabled by the expansion of LRZ's current supercomputer called SuperMUG-NG, which will receive an upgrade in 2022, which will feature both Sapphire Rapids and Ponte Vecchio.

Mr. Raja Koduri, Intel graphics guru, has on Twitter teased that this supercomputer installment will represent a combination of Sapphire Rapids, Ponte Vecchio, Optane, and One API all in one machine. The system will use over one petabyte of Distributed Asynchronous Object Storage (DAOS) based on the Optane technologies. Then, Mr. Koduri has teased some Ponte Vecchio eye candy, which is a GIF of tiles combining to form a GPU, which you can check out here. You can also see some pictures of Ponte Vecchio below.

The NVIDIA RTX 3080 Ti and RTX 3070 Ti are set to be announced on May 31st after various delays. The two cards will then launch sometime in June with the review embargoes set as June 2nd for the RTX 3070 Ti and June 9th for the RTX 3080 Ti. The RTX 3070 Ti features the GA104-400 GPU 6144 CUDA cores and 8 GB GDDR6X memory. The RTX 3080 Ti will use the GA102-225 and include 10240 CUDA cores paired with 12 GB GDDR6X memory, We have already seen some retailers list custom RTX 3080 Ti cards from Gigabyte and MSI with prices ranging from 1338 USD to 2257 USD however these are likely placeholder values. The cards should be officially available to purchase sometime after the review embargoes lift in June.

The current supply of graphics cards has been very tight all over the world. Starting with the launch of the latest Radeon RX 6000 series of GPUs based on RDNA 2 architecture, AMD has found itself in big trouble when it comes to supply of the silicon, compared to the demand that exists for these new GPUs. We have discussed that many times in the past and saw that it represents a problem spanning everyone involved in getting the silicon chips to the hands of consumers. On Tuesday, April 27th, AMD held its Q1 2021 earnings call and webcast, where the company executives talked about the company's future, underlying problems, and ways of addressing them.

Among many topics covered in the call, AMD's President and CEO, Dr. Lisa Su, has talked about the GPU supply. According to Dr. Su, the company is ramping the production of its Radeon graphics cards, adding that the mobile Radeon GPU lineup is lurking. Here is a full quote from the earnings call.

Dr. Lisa Su—AMD Q1 2021 Earning CallAnd we're well-positioned for further growth as we have tripled our commercial notebook design wins with the largest OEMs this year. In graphics, revenue increased by a strong double-digit percentage year over year and sequentially, led by channel sales growth as revenue from our high-end Radeon 6000 GPUs more than doubled from the prior quarter. We introduced our Radeon 6700 XT desktop GPU with leadership 1440p gaming performance in March and are on track for the first notebooks featuring our leading-edge mobile RDNA 2 architecture to launch later this quarter. We expect Radeon 6000 Series GPU sales to grow significantly over the coming quarters as we ramp production.

Alphacool today presents the Eiswolf 2 AIO solution (Reference design with backplate) for AMD Radeon RX 6800/6800XT/6900 and NVIDIA RTX 3080/3090 graphics cards. As with all of our AIO models, we only use copper radiators, however, for the first time, the Alphacool 360 mm NexXxoS ST30 all-copper radiator will be used in one of our GPU "all-in-one" solutions. For the fans, we now install the new Alphacool Aurora Rise 120 mm fan. The fan impresses with a max. statistical pressure of 3.17 mm/H2O and offers a max. air flow of 119.8m3/h, at a speed range of 0-2500rpm which can be controlled via PWM.

The decision to use the larger radiator in conjunction with our new fans provides a significant increase in cooling performance. The AIO is completed with the Eisblock Aurora water cooler, which actively cools all relevant components such as graphics memory and voltage converters in addition to the GPU chip. All components used (radiator, fan, cooler, DC-LT 2 pump) including connectors and hoses are part of the Alphacool custom cooling range and are therefore available separately. This means that we're not just offering a simple AIO solution here, but a real custom water cooling system (pre-assembled & pre-filled). Only the pump housing for the DC-LT 2 pump is a custom design and not available separately. Today we present only the Reference designs. In the near future, more solutions for graphics cards in custom design from various manufacturers will be available.More pictures follow.

SAPPHIRE Technology announces the newest addition to the RDNA 2 family with the SAPPHIRE NITRO+ AMD Radeon RX 6900 XT Special Edition. With increased performance and an innovative spin on the award winning Tri-X Cooling Technology, this model is a powerhouse product for the ultra-aficionado 4K gamer. Designed with the classic NITRO+ minimalist aesthetic, additional fresh customizable RGB on the peripherals, and striking ARGB fans, the NITRO+ AMD Radeon RX 6900 XT Special Edition is sure to be a photogenic part to complete the theme of any PC.

Overclocked straight out of the box to hit every frame, the NITRO+ AMD Radeon RX 6900 XT Special Edition leads with 5120 stream processors and boasts a maximized Boost Clock of up to 2365 MHz and a Game Clock of up to 2135 MHz. It comes decked with 80 ray accelerators and 80 compute units to facilitate raytracing. The latest 16 GB of GDDR6 high-speed memory is clocked at 16 Gbps Effective with 128 MB of Infinity Cache, which dramatically reduces latency and power consumption, delivering higher overall gaming performance than traditional architectural designs.

AMD (NASDAQ:AMD) today announced revenue for the first quarter of 2021 of $3.45 billion, operating income of $662 million, net income of $555 million and diluted earnings per share of $0.45. On a non-GAAP* basis, operating income was $762 million, net income was $642 million and diluted earnings per share was $0.52.

"Our business continued to accelerate in the first quarter driven by the best product portfolio in our history, strong execution and robust market demand," said Dr. Lisa Su, AMD president and CEO. "We had outstanding year-over-year revenue growth across all of our businesses and data center revenue more than doubled. Our increased full-year guidance highlights the strong growth we expect across our business based on increasing adoption of our high-performance computing products and expanding customer relationships."

Recently, Apple has updated its macOS operating system to support the next generation of GPUs for Mac devices, coming from AMD. The upcoming RDNA 2 GPU lineup in Macs will bring all the enhancements AMD made to the architecture, and pack it inside Apple's signature designs. Today, we have received information that Apple could refresh its Mac Pro lineup with AMD's Radeon Pro GPU based on RDNA 2 architecture. On Geekbench 5, there was a test run that was conducted on Apple Mac Pro "7.1" revision, that features not only Intel Cascade Lake-X processor but AMD's unreleased Radeon Pro W6900X graphics card designed for professional users.

While we don't know much about the exact specifications, we know that it features a Navi 21 GPU SKU. Judging by the naming scheme, the Radeon Pro W6900X is representing an Apple-exclusive GPU variant designed only for Mac Pro devices. A Chiphell leak has given us a sneak peek at the alleged card look, which you can see below. When it comes to performance, the Geekbench result measures Metal API performance and the Radeon Pro W6900X has managed to score 171448 points, which is even higher than the Radeon RX 6900 XT GPU, which tops out at 164294 points. Of course, this is representing a professional SKU, so there could be some tuning present as well.

The launch of NVIDIA's upcoming GeForce RTX 3080 Ti graphics card is upon us. The number of rumors circulating the web is getting greater and we have just received die pictures of the GA102 silicon and the specification of the specific SKU. Sources over at VideoCardz have provided the website with the first die picture of GA102-225 silicon, which powers the NVIDIA GeForce RTX 3080 Ti graphics card. Pictured below, it doesn't appear much different compared to the GA102-300 SKU found inside the RTX 3090 card, with the only obvious differentiator being the SKU ID. However, the difference only appears under the hood, with the GA102-225 SKU having 10240 CUDA cores instead of 10752 CUDA cores found inside GA102-300 of RTX 3090.

Paired with 12 GB of GDDR6X memory on a 384-bit bus, the memory will have run around 19 Gbps speeds. That will result in a bandwidth of 912 GB/s. If you are wondering about the performance of the card, it should remain within a few percent of its bigger brother RTX 3090. We have the first leak showing Ethereum mining performance and the GA102-225 silicon achieved a mining hash rate of 118.9 Mh/s with some tuning. The memory was overclocked to 21.5 Gbps, while the GPU TDP was limited to 278 Watts. The leak shows that the card has managed to achieve a 1365 MHz base and 1665 MHz boost frequency. While we don't have the exact launch date, the supposed MSRP will be anywhere from $999 to $1099, assuming you can get it at all at any price.

Overclocker extraordinnaire Roman "der8auer" Hartung has achieved a new world record for GPU clockspeed with the help of exotic cooling and PowerColor's RX 6900 XT Liquid Devil Ultimate. The PowerColor Red Devil Ultimate features the latest and greatest bin of any RX 6900 XT chip, featuring AMD's Navi 21 XTXH silicon (device ID 0x73AF), with improved clockspeed potential and lower leakage than the usual XT cadre of chips. Using LN2 cooling which dropped the operating temperatures of the card down to a finger-killing -87 °C, the overclocker managed to reach a 3,225 MHz clock - as reported by our very own GPU-Z, which continues to be the diagnostics tool of choice for graphics cards. Look after the break for the full video leading up to this overclocking feat.

NVIDIA is preparing the launch of an entry-level graphics card based on the Ampere architecture, sometimes later in the year, with possible dates being this or the following quarter. The GeForce RTX 3050 and RTX 3050 Ti are supposed to be the slowest SKUs in the whole Ampere lineup, representing the least expensive choice from the entire generation. However, we are wondering what the cards will look like and what are some design choices NVIDIA will introduce to these SKUs. Lenovo, the maker of the popular various kinds of laptops, has accidentally listed these exact SKUs in the specification section of its Legion gaming laptop.

The GPUs in question are NVIDIA GeForce RTX 3050 and RTX 3050 Ti mobile edition SKUs, probably based on the GA107 GPU SKU. Both versions are equipped with 4 GB of GDDR6 memory running on a 128-bit bus. The regular GeForce RTX 3050 SKU is configurable from a TGP of just 60 to 80 Watts. NVIDIA has a Dynamic Boost technology that will pump that by additional 15 Watts and max it out at 95 Watt SKU, which this laptop uses. The maximum boost clock frequency of these cards is 1695 MHz and 1740 MHz, for the RTX 3050 Ti and RTX 3050 GPUs respectively.

Graphics cards have been developed over the years so that they feature multi-level cache hierarchies. These levels of cache have been engineered to fill in the gap between memory and compute, a growing problem that cripples the performance of GPUs in many applications. Different GPU vendors, like AMD and NVIDIA, have different sizes of register files, L1, and L2 caches, depending on the architecture. For example, the amount of L2 cache on NVIDIA's A100 GPU is 40 MB, which is seven times larger compared to the previous generation V100. That just shows how much new applications require bigger cache sizes, which is ever-increasing to satisfy the needs.

Today, we have an interesting report coming from Chips and Cheese. The website has decided to measure GPU memory latency of the latest generation of cards - AMD's RDNA 2 and NVIDIA's Ampere. By using simple pointer chasing tests in OpenCL, we get interesting results. RDNA 2 cache is fast and massive. Compared to Ampere, cache latency is much lower, while the VRAM latency is about the same. NVIDIA uses a two-level cache system consisting out of L1 and L2, which seems to be a rather slow solution. Data coming from Ampere's SM, which holds L1 cache, to the outside L2 is taking over 100 ns of latency.

CPU-Z, the CPU monitoring tool used to gather information about the processor your system is running on, has been updated with version 1.96. This new update doesn't change the software much but rather brings support for new hardware. Starting from this revision, Intel's upcoming Alder Lake CPUs have received preliminary support in the tool. To go along with CPUs, the software is now also enabled to recognize Intel's Z6xx motherboards that pair with Alder Lake processors. Alongside that, the software now also brings support for next-generation DDR5 memory, which is supposed to feature speeds anywhere from 4800 to 8400 MT/s. When it comes to AMD, the tool received an update that enables it to read information about AMD's Ryzen 5700G, 5600G, and 5300G APUs, and Radeon RX 6900 XT, 6800 (& XT), 6700 XT GPUs.Download CPU-Z Version 1.96 Here.

Some time ago, NVIDIA introduced its GeForce RTX 3060 graphics card with GA106-300 Ampere GPU SKU. The GPU was the first to feature NVIDIA's latest additions like Resizable BAR and crypto mining algorithm limiter that limited the performance of the card while mining. However, despite NVIDIA's intention to keep the card out of the hands of crypto miners, there has been a lot of flaws in the plan. A lot of people discovered that the card still managed to turn in profits with the limiter enables. Later, NVIDIA accidentally released a driver that actually removes the limiter and enables the GPU to mine at full capacity, making the company's efforts useless.

Today we have new information that NVIDIA will launch an updated GeForce RTX 3060 GPU SKU that features a different ID, in the quest to limit card's mining performance. According to HKEPC, NVIDIA is producing updated GeForce RTX 3060 GKU SKUs with GA106-302 ID that should launch sometime in May, which are supposed to replace the GA106-300 SKUs now present. The software and the drivers will use the new ID to identify new SKUs and limit the performance of the card at mining tasks such as Ethereum mining. That way, it ensures that no driver version or bypass can trick the software to enable the card to use its full mining power and it shall render it unprofitable. Additionally, kopite7kimi, a known hardware leaker, claims that NVIDIA is also preparing updated GPU SKU IDs for GA104 and GA102 GPUs, with GA102-302/202 and GA104-302/202 variants supposedly coming.

NVIDIA today introduced a new class of NVIDIA-Certified Systems, bringing AI within reach for organizations that run their applications on industry-standard enterprise data center infrastructure. These include high-volume enterprise servers from top manufacturers, which were announced in January and are now certified to run the NVIDIA AI Enterprise software suite—which is exclusively certified for VMware vSphere 7, the world's most widely used compute virtualization platform.

Further expanding the NVIDIA-Certified servers ecosystem is a new wave of systems featuring the NVIDIA A30 GPU for mainstream AI and data analytics and the NVIDIA A10 GPU for AI-enabled graphics, virtual workstations and mixed compute and graphics workloads, also announced today.

NVIDIA today announced a range of eight new NVIDIA Ampere architecture GPUs for next-generation laptops, desktops and servers that make it possible for professionals to work from wherever they choose, without sacrificing quality or time. For desktops, the new NVIDIA RTX A5000 and NVIDIA RTX A4000 GPUs feature new RT Cores, Tensor Cores and CUDA cores to speed AI, graphics and real-time rendering up to 2x faster than previous generations. For professionals on the go needing thin and light devices, the new NVIDIA RTX A2000, NVIDIA RTX A3000, RTX A4000 and RTX A5000 laptop GPUs deliver accelerated performance without compromising mobility.

For the data center, there are the new NVIDIA A10 GPU and A16 GPU. The A10 provides up to 2.5x the virtual workstation performance of the previous generation for designers and engineers, while the A16 GPU provides up to 2x user density with lower total cost of ownership and an enhanced virtual desktop infrastructure experience over the previous generation.

NVIDIA today announced its first data center CPU, an Arm-based processor that will deliver 10x the performance of today's fastest servers on the most complex AI and high performance computing workloads.

The result of more than 10,000 engineering years of work, the NVIDIA Grace CPU is designed to address the computing requirements for the world's most advanced applications—including natural language processing, recommender systems and AI supercomputing—that analyze enormous datasets requiring both ultra-fast compute performance and massive memory. It combines energy-efficient Arm CPU cores with an innovative low-power memory subsystem to deliver high performance with great efficiency.

NVIDIA today unveiled the world's first cloud-native, multi-tenant AI supercomputer—the next-generation NVIDIA DGX SuperPOD featuring NVIDIA BlueField -2 DPUs. Fortifying the DGX SuperPOD with BlueField-2 DPUs—data processing units that offload, accelerate and isolate users' data—provides customers with secure connections to their AI infrastructure.

The company also announced NVIDIA Base Command, which enables multiple users and IT teams to securely access, share and operate their DGX SuperPOD infrastructure. Base Command coordinates AI training and operations on DGX SuperPOD infrastructure to enable the work of teams of data scientists and developers located around the globe.

The global supply of semiconductor processors has been at risk lately. Starting from GPUs to CPUs, the demand for both has been much greater than the available supply. Manufacturing companies, such as TSMC, have been expanding capacities, however, they have not yet been able to satisfy the demand. We have seen the results of that demand in a form of the scarcity of the latest generation of graphics cards, covering NVIDIA's GeForce RTX 3000 series Ampere, and AMD' Radeon RX 6000 series Big Navi graphics cards. Consumers have had a difficult time sourcing them and they have seen artificial price increase that is much higher than their original MSRP.

However, it doesn't seem like the situation will improve. According to the latest reporting from Bloomberg, the next victim of global chip shortage is... you guessed it, your home internet router. The cited sources have noted that the waiting list to get a batch of ordered routers has doubled the waiting time, from the regular 30 weeks to 60-week waiting time. This represents a waiting list that is more than a year long. With the global COVID-19 pandemic still going strong, there is an increased need for better home router equipment, and delays can only hurt broadband providers that supply routers. Taiwan-based router manufacturer Zyxel Communications, notes that the company has seen massive demand for their equipment. Such a massive demand could lead to insufficient supply, which could increase prices of routers well above their MSRP and bring scarcity of them as well.

We have received various leaks and benchmarks for AMD's upcoming Ryzen 5000G processors, these were all from engineering samples but we now have our first look at the retail 5700G. The AMD Ryzen 7 5700G features the model number 100-000000263 attributed to earlier rumors and has been tested in CPU-Z scoring 631 points in single-threaded performance along with 6782 points in multi-threaded, and in Cinebench R20 it scored 6040 points. The integrated Vega graphics lack any official drivers but GPU-Z reports a Vega 8 processor with 12 Streaming Multiprocessors and a base clock of 2 GHz. AMD is yet to officially announce any Ryzen 5000G processors so it is unclear how far away their launch is and whether or not they will be made available to the DIY market.

Alphacool introduces four new extremely large radiators. The first two radiators are for 180 or 200 mm fans and are a thickness of 86 mm. The 200 and 400 mm NexXxoS Monsta radiators offer a huge cooling surface. Both radiators have 7x standard G1/4" ports. 6 of the ports are for inlet/outlet and the 7th port is located at the end chamber and can only be used for filling, draining, or bleeding the radiator. The fin density has been reduced from 16 to 12 FPI to allow the use of lower static pressure fans and lower RPMs.

The new Nova 1080 UT60 radiator is a thicker version of the Nova 1080 XT45 radiator. The enormous cooling surface is comparable to three 360 mm UT60 radiators. This makes it easy to cool more than 1000 W of excess heat with extremely quiet to silent spinning fans. This is equivalent to a 64 core AMD Threadripper of the latest generation and at least two GeForce RTX 3090 graphics cards.

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.

When AMD introduced RDNA 2 architecture and higher-end Navi 21 GPU SKUs, it was only a matter of time before the company launches these GPUs inside professional-grade graphics cards. Today, thanks to the Chiphell forums, we have pictures and some specifications of AMD's upcoming Radeon Pro workstation graphics card. Pictured below is a new RDNA 2 based design that features AMD's Navi 21 GLXL GPU SKU. The new GLXL GPU SKU is supposed to be rather similar to the Navi 21 XL GPU found inside AMD's Radeon RX 6800 XT graphics card, judging by the number and arrangement of capacitors on the back of the card.

When it comes to memory, the upcoming Radeon Pro workstation card is featuring 16 GB of VRAM, likely a variant of GDDR6 found on gaming-oriented graphics cards from RDNA 2 generation. When it comes to cooler design, the Radeon Pro graphics card has a blower-type cooler helping tame the Navi 21 GLXL GPU. Given that blower-type coolers are suitable for situations with less airflow, the TDP of this card could be around or under 250 Watts. You can take a look at the card below, however, do note that it is an engineering sample and the final product can look a bit different.

AMD on April 1st published a new patent application that seems to show the way its chiplet GPU design is moving towards. Before you say it, it's a patent application; there's no possibility for an April Fool's joke on this sort of move. The new patent develops on AMD's previous one, which only featured a passive bridge connecting the different GPU chiplets and their processing resources. If you want to read a slightly deeper dive of sorts on what chiplets are and why they are important for the future of graphics (and computing in general), look to this article here on TPU.

The new design interprets the active bridge connecting the chiplets as a last-level cache - think of it as L3, a unifying highway of data that is readily exposed to all the chiplets (in this patent, a three-chiplet design). It's essentially AMD's RDNA 2 Infinity Cache, though it's not only used as a cache here (and for good effect, if the Infinity Cache design on RDNA 2 and its performance uplift is anything to go by); it also serves as an active interconnect between the GPU chiplets that allow for the exchange and synchronization of information, whenever and however required. This also allows for the registry and cache to be exposed as a unified block for developers, abstracting them from having to program towards a system with a tri-way cache design. There are also of course yield benefits to be taken here, as there are with AMD's Zen chiplet designs, and the ability to scale up performance without any monolithic designs that are heavy in power requirements. The integrated, active cache bridge would also certainly help in reducing latency and maintaining chiplet processing coherency.

Tianshu Zhixin, a Chinese startup company dedicated to designing advanced processors for accelerating various kinds of tasks, has officially entered the production of its latest GPGPU design. Called "Big Island" GPU, it is the company's entry into the GPU market, currently dominated by AMD, NVIDIA, and soon Intel. So what is so special about Tianshu Zhixin's Big Island GPU, making it so important? Firstly, it represents China's attempt of independence from the outside processor suppliers, ensuring maximum security at all times. Secondly, it is an interesting feat to enter a market that is controlled by big players and attempt to grab a piece of that cake. To be successful, the GPU needs to represent a great design.

And great it is, at least on paper. The specifications list that Big Island is currently being manufactured on TSMC's 7 nm node using CoWoS packaging technology, enabling the die to feature over 24 billion transistors. When it comes to performance, the company claims that the GPU is capable of crunching 37 TeraFLOPs of single-precision FP32 data. At FP16/BF16 half-precision, the chip is capable of outputting 147 TeraFLOPs. When it comes to integer performance, it can achieve 317, 147, and 295 TOPS in INT32, INT16, and IN8 respectively. There is no data on double-precision floating-point numbers, so the chip is optimized for single-precision workloads. There is also 32 GB of HBM2 memory present, and it has 1.2 TB of bandwidth. If we compare the chip to the competing offers like NVIDIA A100 or AMD MI100, the new Big Island GPU outperforms both at single-precision FP32 compute tasks, for which it is designed.Pictures of possible solutions follow.