The 58th annual edition of the TOP500 saw little change in the Top10. The Microsoft Azure system called Voyager-EUS2 was the only machine to shake up the top spots, claiming No. 10. Based on an AMD EPYC processor with 48 cores and 2.45GHz working together with an NVIDIA A100 GPU and 80 GB of memory, Voyager-EUS2 also utilizes a Mellanox HDR Infiniband for data transfer.

While there were no other changes to the positions of the systems in the Top10, Perlmutter at NERSC improved its performance to 70.9 Pflop/s. Housed at the Lawrence Berkeley National Laboratory, Perlmutter's increased performance couldn't move it from its previously held No. 5 spot.

KIOXIA America, Inc. today announced the production availability of its EM6 Series Enterprise NVMe-oF solid state drives (SSDs) for Ethernet Bunch of Flash (EBOF) systems. Using the Marvell 88SN2400 NVMe-oF SSD converter controller that converts an NVMe SSD into a dual-ported 25Gb NVMe-oF SSD, KIOXIA EM6 Series drives expose the entire SSD bandwidth to the network.

Due to their ability to scale performance of NVMe SSDs, native NVMe-oF architectures are well-suited for applications such as artificial intelligence (AI)/machine learning (ML), high performance computing (HPC) and storage expansion. In the case of HPC, leveraging the Lustre file system, which is used to provide high bandwidth and parallel access to compute clusters, is beneficial to NVMe-oF based storage, such as EBOF systems with EM6 SSDs, that enable high availability (HA) configurations. An HPC HA configuration example consists of multiple and redundant network connections between a compute host and an EBOF with 88SN2400-connected NVMe SSDs, to deliver scalable throughput based on the number of SSDs.

NVIDIA today announced NVIDIA Quantum-2, the next generation of its InfiniBand networking platform, which offers the extreme performance, broad accessibility and strong security needed by cloud computing providers and supercomputing centers.

The most advanced end-to-end networking platform ever built, NVIDIA Quantum-2 is a 400 Gbps InfiniBand networking platform that consists of the NVIDIA Quantum-2 switch, the ConnectX-7 network adapter, the BlueField-3 data processing unit (DPU) and all the software that supports the new architecture.

NVIDIA announced the first four Crypto Mining Processor (CMP) cards earlier this year with performance ranging from 26 MH/s to 86 MH/s. These cards were all based on existing Turing/Ampere silicon and featured board partner-designed cooling systems. NVIDIA appears to have introduced a new flagship model with the passively-cooled 170HX that is based on the NVIDIA A100 accelerator which features a GA100 GPU.

This new model is the first mining card to be designed by NVIDIA and features 4480 CUDA cores paired with 8 GB of HBM2E memory which are both considerably less than what is found in other GA100 based products. NVIDIA has also purposively limited the PCIe interface to Gen 1 x4 to ensure the card cannot be used for tasks outside of cryptocurrency mining. The 170HX has a TDP of 250 W and runs at a base clock of 1140 MHz with a locked-down BIOS that does not allow memory overclocking resulting in a hash rate of 164 MH/s when using the Etash algorithm.

Intel in its 2021 Architecture Day presentation put out fine technical details of its Xe HPC Ponte Vecchio accelerator, including some [very] preliminary performance claims for its current A0-silicon-based prototype. The prototype operates at 1.37 GHz, but achieves out at least 45 TFLOPs of FP32 throughput. We calculated the clock speed based on simple math. Intel obtained the 45 TFLOPs number on a machine running a single Ponte Vecchio OAM (single MCM with two stacks), and a Xeon "Sapphire Rapids" CPU. 45 TFLOPs sees the processor already beat the advertised 19.5 TFLOPs of the NVIDIA "Ampere" A100 Tensor Core 40 GB processor. AMD isn't faring any better, with its production Instinct MI100 processor only offering 23.1 TFLOPs FP32.

NVIDIA (NASDAQ: NVDA) today reported record revenue for the second quarter ended August 1, 2021, of $6.51 billion, up 68 percent from a year earlier and up 15 percent from the previous quarter, with record revenue from the company's Gaming, Data Center and Professional Visualization platforms. GAAP earnings per diluted share for the quarter were $0.94, up 276 percent from a year ago and up 24 percent from the previous quarter. Non-GAAP earnings per diluted share were $1.04, up 89 percent from a year ago and up 14 percent from the previous quarter.

"NVIDIA's pioneering work in accelerated computing continues to advance graphics, scientific computing and AI," said Jensen Huang, founder and CEO of NVIDIA. "Enabled by the NVIDIA platform, developers are creating the most impactful technologies of our time - from natural language understanding and recommender systems, to autonomous vehicles and logistic centers, to digital biology and climate science, to metaverse worlds that obey the laws of physics.

Hopper is an upcoming compute architecture from NVIDIA which will be the first from the company to feature a Multi-Chip-Module (MCM) design similar to Intel's Xe-HPC and AMD's upcoming CDNA2. The Hopper architecture has been teased for over 2 years but it would appear that it is nearing completion with a recent leak suggesting the product will tape out soon. This compute GPU will likely be manufactured on TSMC's 5 nm node and could feature two dies each with 288 Streaming Microprocessors which could theoretically provide a three-fold performance improvement over the Ampere-based NVIDIA A100. The first product to feature the GPU is expected to be the NVIDIA H100 data center accelerator which will serve as a successor to the A100 and could potentially launch in mid-2022.

During this year's ISC 2021 event, as a part of the company's exhibition portfolio, NVIDIA has decided to launch an updated version of the A100 accelerator. A couple of months ago, in November, NVIDIA launched an 80 GB HBM2E version of the A100 accelerator, on the SXM2 proprietary form-factor. Today, we are getting the same upgraded GPU in the more standard dual-slot PCIe type of card. Featuring a GA100 GPU built on TSMC's 7 nm process, this SKU has 6192 CUDA cores present. To pair with the beefy amount of computing, the GPU needs appropriate memory. This time, there is as much as 80 GB of HBM2E memory. The memory achieves a bandwidth of 2039 GB/s, with memory dies running at an effective speed of 3186 Gbps. An important note is that the TDP of the GPU has been lowered to 250 Watts, compared to the 400 Watt SXM2 solution.

To pair with the new upgrade, NVIDIA made another announcement today and that is an enterprise version of Microsoft's DirectStorage, called NVIDIA GPUDirect Storage. It represents a way of allowing applications to access the massive memory pool built on the GPU, with 80 GB of super-fast HBM2E memory.

Amulet Hotkey, a leader in design, manufacturing and system integration for mission-critical remote workstation and high-GPU-density solutions, is pleased to confirm it is now supplying HUT 8 Mining Corp. with its CoreServer CX4140 technology that hosts either four NVIDIA CMPs or four NVIDIA A100 GPUs in a 1U rack form factor server, representing unprecedented compute density for High Performance Computing (HPC) or the mining of blockchain networks.

Amulet Hotkey is trusted by partners and customers to deliver solutions that shatter the CMP or GPU rack density offered by other manufacturers. By using its knowledge and experience in managing heat dissipation, Amulet Hotkey has once again demonstrated how strategic partnerships deliver unprecedented benefits.

"A key element for HUT 8 Mining is to work with technology partners who can bring flexible enterprise-grade equipment and who are able to meet the timelines and needs of an industrial-size crypto currency miner," said Jason Zaluski, Head of Technology, HUT 8 Mining. "Amulet Hotkey's ability to engineer four NVIDIA CMPs into a 1U rack server is testament to the way their technology is able to combine performance and efficiency, both aspects are critical to our mining at scale"

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.

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.

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.

When NVIDIA introduced its Ampere family of graphics cards, the GPU lineup's first product was the A100 GPU. While not being a GPU used for gaming, the model is designed with compute-heavy workloads in mind. Even NVIDIA says that "NVIDIA A100 Tensor Core GPU delivers unprecedented acceleration at every scale to power the world's highest-performing elastic data centers for AI, data analytics, and HPC." However, it seems like the GA100 SKU, the base of the A100 GPU, could be used for another task that requires heavy computation and could benefit very much from the sheer core count that the biggest Ampere SKU offers.

According to a known leaker @kopite7kimi (Twitter), NVIDIA could repurpose the GA100 GPU SKU and launch it as a part of the CMP HX crypto mining series of graphics cards. As a reminder, the CMP series is specifically designed for the sole purpose of mining cryptocurrency, and CMP products have no video outputs. According to Kopite, the repurposed GPU SKU could be a "mining monster", which is not too hard to believe given the huge core count the SKU has and the fact that it was made for heavy computation workloads. While we do not the exact specifications of the rumored CMP HX SKU, you can check out the A100 GPU specifications here.

NVIDIA today announced NVIDIA AI Enterprise, a comprehensive software suite of enterprise-grade AI tools and frameworks optimized, certified and supported by NVIDIA, exclusively with VMware vSphere 7 Update 2, separately announced today.

Through a first-of-its-kind industry collaboration to develop an AI-Ready Enterprise platform, NVIDIA teamed with VMware to virtualize AI workloads on VMware vSphere with NVIDIA AI Enterprise. The offering gives enterprises the software required to develop a broad range of AI solutions, such as advanced diagnostics in healthcare, smart factories for manufacturing, and fraud detection in financial services.

GIGABYTE Technology, (TWSE: 2376), an industry leader in high-performance servers and workstations, today announced the G262-ZR0 for HPC, AI, and data analytics. Designed to support the highest-level of performance in GPU computing, the G262-ZR0 incorporates fast PCIe 4.0 throughput in addition to NVIDIA HGX technologies and NVIDIA NVLink to provide industry leading bandwidth performance.

GPU supply has been one of the most interesting things this year. With a huge demand for the new GPU generations like NVIDIA's Ampere and AMD's RDNA 2 "Big Navi" graphics cards, everyone is trying to grab a card for themselves. Besides the huge demand, there is also a big problem. The supply of these GPUs is just too low to satisfy the demand, driving up the prices, and increasing the scarcity of them. Companies like NVIDIA have their priorities set: all of the major production will go for the data center expansion and data center customers. However, even that plan is proving not to be good enough.

The scarcity of GPUs has now hit data centers, with NVIDIA unable to satisfy the demand for its A100 GPUs designed for high-performance computing. "It is going to take several months to catch up some of the demand," said Ian Buck, vice president of Accelerated Computing Business Unit at Nvidia. That is an indicator of just how huge the demand for these accelerators is. With the recent company announcement of A100 GPU with 80 GB memory, partners expect to have the first cards in their systems in the first half of 2021. That means that this situation and inadequate supply will hopefully resolve sometime around that timeframe.

The 56th edition of the TOP500 saw the Japanese Fugaku supercomputer solidify its number one status in a list that reflects a flattening performance growth curve. Although two new systems managed to make it into the top 10, the full list recorded the smallest number of new entries since the project began in 1993.

The entry level to the list moved up to 1.32 petaflops on the High Performance Linpack (HPL) benchmark, a small increase from 1.23 petaflops recorded in the June 2020 rankings. In a similar vein, the aggregate performance of all 500 systems grew from 2.22 exaflops in June to just 2.43 exaflops on the latest list. Likewise, average concurrency per system barely increased at all, growing from 145,363 cores six months ago to 145,465 cores in the current list.

NVIDIA today unveiled the NVIDIA A100 80 GB GPU—the latest innovation powering the NVIDIA HGX AI supercomputing platform—with twice the memory of its predecessor, providing researchers and engineers unprecedented speed and performance to unlock the next wave of AI and scientific breakthroughs. The new A100 with HBM2E technology doubles the A100 40 GB GPU's high-bandwidth memory to 80 GB and delivers over 2 terabytes per second of memory bandwidth. This allows data to be fed quickly to A100, the world's fastest data center GPU, enabling researchers to accelerate their applications even faster and take on even larger models and datasets.

"Achieving state-of-the-art results in HPC and AI research requires building the biggest models, but these demand more memory capacity and bandwidth than ever before," said Bryan Catanzaro, vice president of applied deep learning research at NVIDIA. "The A100 80 GB GPU provides double the memory of its predecessor, which was introduced just six months ago, and breaks the 2 TB per second barrier, enabling researchers to tackle the world's most important scientific and big data challenges."

AMD is eyeing a mid-November debut for its CDNA compute architecture with the Instinct MI100 compute accelerator card. CDNA is a fork of RDNA for headless GPU compute accelerators with large SIMD resources. An Aroged report pins the launch of the MI100 at November 16, 2020, according to leaked AMD documents it dug up. The Instinct MI100 will eye a slice of the same machine intelligence pie NVIDIA is seeking to dominate with its A100 Tensor Core compute accelerator.

It appears like the first MI100 cards will be built in the add-in-board form-factor with PCI-Express 4.0 x16 interfaces, although older reports do predict AMD creating a socketed variant of its Infinity Fabric interconnect for machines with larger numbers of these compute processors. In the leaked document, AMD claims that the Instinct MI100 is the "world's highest double-precision accelerator for machine learning, HPC, cloud compute, and rendering systems." This is an especially big claim given that the A100 Tensor Core features FP64 CUDA cores based on the "Ampere" architecture. Then again, given that AMD claims that the RDNA2 graphics architecture is clawing back at NVIDIA with performance at the high-end, the competitiveness of the Instinct MI100 against the A100 Tensor Core cannot be discounted.

TechPowerUp today released the latest version of TechPowerUp GPU-Z, the popular graphics sub-system information and diagnostic utility. Version 2.35.0 adds support for new GPUs, and fixes a number of bugs. To begin with, GPU-Z adds support for AMD Radeon RX 6000 series GPUs based on the "Navi 21" silicon. Support is also added for Intel DG1 GPU. BIOS extraction and upload for NVIDIA's RTX 30-series "Ampere" GPUs has finally been introduced. Memory size reporting on the RTX 3090 has been fixed. The latest Windows 10 Insider Build (20231.1000) made some changes to DirectML, which caused GPU-Z to report it as unavailable, this has been fixed.

TechPowerUp GPU-Z 2.35.0 also makes various improvements to fake GPU detection for cards based on NVIDIA GT216 and GT218 ASICs. Hardware detection for AMD Radeon Pro 5600M based on "Navi 12" has been fixed. Among the other GPUs for which support was added with this release are NVIDIA A100 Tensor Core PCIe, Intel UHD Gen9.5 graphics on the i5-10200H, and Radeon HD 8210E and Barco MXRT-6700. Grab GPU-Z from the link below.

DOWNLOAD: TechPowerUp GPU-Z 2.35.0The change-log follows.

A report straight from DigiTimes claims that NVIDIA is looking to upgrade their Ampere consumer GPUs from Samsung's 8 nm to TSMC's 7 nm. According to the source, the volume of this transition should be "very large", but most likely wouldn't reflect the entirety of Ampere's consumer-facing product stack. The report claims that TSMC has become more "friendly" to NVIDIA. This could be because TSMC now has available manufacturing capacity in 7 nm due to some of its clients moving to the company's 5 nm node, or simply because TSMC hadn't believed NVIDIA to consider Samsung as a viable foundry alternative - which it now does - and has thus lowered pricing.

There are various reasons being leveraged at this, none with substantial grounds other than "reported from industry sources". NVIDIA looking for better yields is one of the appointed reasons, as is its history as a TSMC customer. NVIDIA shouldn't have too high a cost porting its manufacturing to TSMC in terms of design changes to the silicon level so as to cater to different characteristics of TSMC's 7 nm, because the company's GA100 GPU (Ampere for the non-consumer market) is already manufactured at TSMC. The next part of this post is mere (relatively informed) speculation, so take that with a saltier disposition than what came before.

NVIDIA today announced that it is building the United Kingdom's most powerful supercomputer, which it will make available to U.K. healthcare researchers using AI to solve pressing medical challenges, including those presented by COVID-19.

Expected to come online by year end, the "Cambridge-1" supercomputer will be an NVIDIA DGX SuperPOD system capable of delivering more than 400 petaflops of AI performance and 8 petaflops of Linpack performance, which would rank it No. 29 on the latest TOP500 list of the world's most powerful supercomputers. It will also rank among the world's top 3 most energy-efficient supercomputers on the current Green500 list.

AMD has been preparing to launch its MI100 accelerator and fight NVIDIA's A100 Ampere GPU in machine learning and AI horizon, and generally compute-intensive workloads. According to some news sources over at AdoredTV, the GPU alleged specifications were listed, along with some slides about the GPU which should be presented at the launch. So to start, this is what we have on the new Radeon MI100 "Arcturus" GPU based on CDNA architecture. The alleged specifications mention that the GPU will feature 120 Compute Units (CUs), meaning that if the GPU keeps the 64-core per CU configuration, we are looking at 7680 cores powered by CDNA architecture.

The leaked slide mentions that the GPU can put out as much as 42 TeraFLOPs of FP32, single-precision compute. This makes it more than twice as fast compared to NVIDIA's A100 GPU at FP32 workloads. To achieve that, the card would need to have all of its 7680 cores running at 2.75 GHz, which would be a bit high number. On the same slide, the GPU is claimed to have 9.5 TeraFLOPs of FP64 dual-precision performance, while the FP16 power is going to be around 150 TeraFLOPs. For comparison, the A100 GPU from NVIDIA features 9.7 TeraFLOPS of FP64, 19.5 TeraFLOPS of FP32, and 312 (or 634 with sparsity enabled) TeraFLOPs of FP16 compute. AMD GPU is allegedly only more powerful for FP32 workloads, where it outperforms the NVIDIA card by 2.4 times. And if that is really the case, AMD has found its niche in the HPC sector, and it plans to dominate there. According to AdoredTV sources, the GPU could be coming in December of this year.

When NVIDIA announced its Ampere lineup of the graphics cards, the A100 GPU was there to represent the higher performance of the lineup. The GPU is optimized for heavy computing workloads as well as machine learning and AI tasks. Today, NVIDIA has submitted the MLPerf results on the A100 GPU to the MLPerf database. What is MLPerf and why it matters you might think? Well, MLPerf is a system benchmark designed to test the capability of a system for machine learning tasks and enable comparability between systems. The A100 GPU got benchmarked in the latest 0.7 version of the benchmark.

The baseline for the results was the previous generation king, V100 Volta GPU. The new A100 GPU is average 1.5 to 2.5 times faster compared to V100. So far A100 GPU system beats all offers available. It is worth pointing out that not all competing systems have been submitted, however, so far the A100 GPU is the fastest.The performance results follow:

When NVIDIA introduced its Ampere A100 GPU, it was said to be the company's fastest creation yet. However, we didn't know how fast the GPU exactly is. With the whopping 6912 CUDA cores, the GPU can pack all that on a 7 nm die with 54 billion transistors. Paired with 40 GB of super-fast HBM2E memory with a bandwidth of 1555 GB/s, the GPU is set to be a good performer. And how fast it exactly is you might wonder? Well, thanks to the Jules Urbach, the CEO of OTOY, a software developer and maker of OctaneRender software, we have the first benchmark of the Ampere A100 GPU.

Scoring 446 points in OctaneBench, a benchmark for OctaneRender, the Ampere GPU takes the crown of the world's fastest GPU. The GeForce RTX 2080 Ti GPU scores 302 points, which makes the A100 GPU up to 47.7% faster than Turing. However, the fastest Turing card found in the benchmark database is the Quadro RTX 8000, which scored 328 points, showing that Turing is still holding well. The result of Ampere A100 was running with RTX turned off, which could yield additional performance if RTX was turned on and that part of the silicon started working.