NVIDIA today introduced the next generation of NVIDIA Mellanox 400G InfiniBand, giving AI developers and scientific researchers the fastest networking performance available to take on the world's most challenging problems.

As computing requirements continue to grow exponentially in areas such as drug discovery, climate research and genomics, NVIDIA Mellanox 400G InfiniBand is accelerating this work through a dramatic leap in performance offered on the world's only fully offloadable, in-network computing platform. The seventh generation of Mellanox InfiniBand provides ultra-low latency and doubles data throughput with NDR 400 Gb/s and adds new NVIDIA In-Network Computing engines to provide additional acceleration.

AMD has won a contract to empower the LUMI supercomputer, designed for the EuroHPC Joint Undertaking (EuroHPC JU) in conjunction with 10 European countries. The contract will see AMD provide both the CPU and GPU innards of the LUMI, set to be populated with next-generation AMD Epyc CPUs and AMD Instinct GPUs. The supercomputer, which is set to enter operation come next year, will deliver an estimated 552 petaflop/s - higher than the world's current fastest supercomputer, Fugaku in Japan, which reaches peak performance of 513 petaflop/s - and is an Arm-powered affair.

The contract for LUMI's construction has been won by Hewlett Packard Enterprise (HPE), which will be providing an HPE Cray EX supercomputer powered by the aforementioned AMD hardware. LUMI has an investment cost set at 200 million euros, for both hardware, installation, and the foreseeable lifetime of its operation. This design win by AMD marks another big contract for the company, which was all but absent from the supercomputing space until launch, and subsequent iterations, of its Zen architecture and latest generations of Instinct HPC accelerators.

Los Alamos National Laboratory has completed the installation of a next-generation high performance computing platform, with aim to enhance its ongoing R&D efforts in support of the nation's response to COVID-19. Named Chicoma, the new platform is poised to demonstrate Hewlett Packard Enterprise's new HPE Cray EX supercomputer architecture for solving complex scientific problems.

"As extensive social and economic impacts from COVID-19 continue to grip the nation, Los Alamos scientists are actively engaged in a number of critical research efforts ranging from therapeutics design to epidemiological modeling," said Irene Qualters, Associate Laboratory Director for Simulation and Computing at Los Alamos. "High Performance Computing is playing a critical role by allowing scientists to model the complex phenomena involved in viral evolution and propagation."

NVIDIA today announced that the Italian inter-university consortium CINECA—one of the world's most important supercomputing centers—will use the company's accelerated computing platform to build the world's fastest AI supercomputer.

The new "Leonardo" system, built with Atos, is expected to deliver 10 exaflops of FP16 AI performance to enable advanced AI and HPC converged application use cases. Featuring nearly 14,000 NVIDIA Ampere architecture-based GPUs and NVIDIA Mellanox HDR 200 Gb/s InfiniBand networking, Leonardo will propel Italy as the global leader in AI and high performance computing research and innovation.

Marvell (NASDAQ: MRVL) today introduced the industry's first native NVMe RAID 1 accelerator, a state-of-the-art technology for virtualized, multi-tenant cloud and enterprise data center environments which demand optimized reliability, efficiency, and performance. Hewlett Packard Enterprise (HPE) is the first of Marvell's partners to support the new accelerator in the HPE NS204i-p NVMe OS Boot Device offered on select HPE ProLiant servers and HPE Apollo systems.

As the industry transitions from legacy SAS and SATA to NVMe SSDs, Marvell's offering helps data centers fast-track the move to higher performance flash storage. The innovative accelerator lowers data center total cost of ownership (TCO) by offloading RAID 1 processing from costly and precious server CPU resources, maximizing application processing performance. IT organizations can now deploy a "plug-and-play," NVMe-based OS boot solution, like the HPE NS204i-p NVMe OS Boot Device, that protects the integrity of flash data storage while delivering an optimized, application-level user experience.

The Alliance for Computing at Extreme Scale (ACES), a partnership between Los Alamos National Laboratory and Sandia National Laboratories, announced the details of a $105 million contract awarded to Hewlett Packard Enterprise (HPE) to deliver Crossroads, a next-generation supercomputer to be sited at Los Alamos.

"This machine will advance our ability to study the most complex physical systems for science and national security. We look forward to its arrival and deployment," said Jason Pruet, Los Alamos' Program Director for the Advanced Simulating and Computing (ASC) Program.

Ten years ago, Arm set its sights on deploying its compute-efficient technology in the data center with a vision towards a changing landscape that would require a new approach to infrastructure compute.

That decade-long effort to lay the groundwork for a more efficient infrastructure was realized when we announced Arm Neoverse, a new compute platform that would deliver 30% year-over-year performance improvements through 2021. The unveiling of our first two platforms, Neoverse N1 and E1, was significant and important. Not only because Neoverse N1 shattered our performance target by nearly 2x to deliver 60% more performance when compared to Arm's Cortex-A72 CPU, but because we were beginning to see real demand for more choice and flexibility in this rapidly evolving space.

Direct-attached performance from network-attached devices is no longer a thing of storage architects' dreams. KIOXIA America, Inc. (formerly Toshiba Memory America, Inc.), is now sampling Ethernet SSDs to select partners and customers interested in validating the benefits of Ethernet attached storage to their existing Ethernet (RoCEv2) networks. KIOXIA has been working in collaboration with key industry players Marvell, Foxconn-Ingrasys and Accton to bring groundbreaking Ethernet Bunch of Flash (EBOF) technology solutions to market - and this announcement is pivotal to that endeavor.

In an ongoing quest to contain explosive amounts of data, storage capacity and bandwidth must continue to grow while processing time must decrease. An EBOF system addresses these challenges through an Ethernet fabric that can scale flash and optimally disaggregate storage from compute. The EBOF storage solution bypasses the cost, complexity, and system limitations inherent with standard JBOF storage systems, which typically include a CPU, DRAM, HBA, and switch. This accelerates applications and workloads where disaggregated low-latency, high bandwidth, highly available storage is needed - bringing greatly improved performance and lower total cost of ownership to edge, enterprise and cloud data centers.

Rambus Inc. (NASDAQ: RMBS), a premier silicon IP and chip provider making data faster and safer, today announced it has achieved a record 4 Gbps performance with the Rambus HBM2E memory interface solution consisting of a fully-integrated PHY and controller. Paired with the industry's fastest HBM2E DRAM from SK hynix operating at 3.6 Gbps, the solution can deliver 460 GB/s of bandwidth from a single HBM2E device. This performance meets the terabyte-scale bandwidth needs of accelerators targeting the most demanding AI/ML training and high-performance computing (HPC) applications.

"With this achievement by Rambus, designers of AI and HPC systems can now implement systems using the world's fastest HBM2E DRAM running at 3.6 Gbps from SK hynix," said Uksong Kang, vice president of product planning at SK hynix. "In July, we announced full-scale mass-production of HBM2E for state-of-the-art computing applications demanding the highest bandwidth available."

GIGABYTE Technology, an industry leader in high-performance servers and workstations, today is announcing a partnership with Northern Data AG to create a HPC mega-project with computing power of around 3.1 exaflops. GIGABYTE will supply GPU-based server systems equipped with proven AMD EPYC processors and AMD Radeon Instinct accelerators from technology partner AMD, a leading provider of high performance computing and graphics technologies, to Northern Data.

Northern Data develops a distributed computing cluster based on the hardware at locations in Norway, Sweden and Canada, which in its final stage of deployment will provide FP32 computing power of around 3.1 exaflops (3.1 million teraflops and 274.54 petaflops FP64). The world's fastest supercomputer, the Japanese "Fukagu" (Fujitsu), has a calculation power of 1.07 exaflops FP32 and 415.3 petaflops FP64, whereas the second fastest, the US supercomputer "Summit" (IBM) has a calculation power of 0.414 exaflops FP32 and 148.0 petaflops FP64.

NVIDIA over the weekend formally renamed Mellanox Technologies to NVIDIA Networking. The graphics and scalar computing giant had acquired Mellanox in April 2020, in a deal valued at $7 billion. It is expected that the NVIDIA corporate identity will cover all Mellanox products, including NICs, switches, and interconnect solutions targeted at large-scale data-centers and HPC environments. Mellanox website now defaults to NVIDIA, with the announcement banner "Mellanox Technologies is now NVIDIA Networking." With the acquisition of Mellanox, a potential bid for Softbank's Arm Holdings, and market leadership in the scalar compute industry, NVIDIA moves close to becoming an end-to-end enterprise solution provider.

Here's the first picture of an NVIDIA "Ampere" GA102 GPU die. This is the largest client-segment implementation of the "Ampere" architecture by NVIDIA, targeting the gaming (GeForce) and professional-visualization (Quadro) market segments. The "Ampere" architecture itself debuted earlier this year with the A100 Tensor Core scalar processor that's winning hearts and minds in the HPC community faster than ice cream on a dog day afternoon. There's no indication of die-size, but considering how tiny the 10.3 billion-transistor AMD "Navi 10" die is, the GA102 could come with a massive transistor count if its die is as big as that of the TU102. The GPU in the picture is also a qualification sample, and was probably pictured off a prototype graphics card. Powering the GeForce RTX 3090, the GA102-300 is expected to feature a CUDA core count of 5,248. According to VideoCardz, there's a higher trim of this silicon, the GA102-400, which could make it to NVIDIA's next halo product under the TITAN brand.

Intel has been focusing its resources on data center and high-performance computing lately and the company has made some interesting products. Today, Intel has released its latest creation - mOS operating system. Created as a research project, Intel has made an OS made for some extreme-scale HPC systems, meaning that the OS is created for hyper scalers and ones alike. The goal of mOS is to deliver a high-performance environment for software with low-noise, scalability, and the concept of lightweight kernels (LWK) that manage the system.. Being based on the Linux kernel, the OS is essentially another distribution, however, it has been modified so it fits the HPC ecosystem the best way. The mOS is a product in the pre-alpha phase, however, it can already be used in supercomputers like ASCI Red, IBM Blue Gene, and others. Intel is aiming to develop a stable release by the time the Aurora exascale system is ready so it can deploy mOS there.

GIGABYTE Technology, an industry leader in high-performance servers and workstations, today is announcing a partnership with German based Northern Data AG, one of the world's largest providers of High-Performance Computing (HPC) solutions, to develop distributed computing in Northern Data's data centers. These data centers with GIGABYTE built HPC servers will be located in the Nordics and managed by Northern Data.

GIGABYTE has more than 30 years of engineering expertise and success stories in the development and production of server solutions that cover a myriad of uses from AI servers to GPU dense servers to scientific computing servers and more. The success that GIGABYTE's servers and workstations have had attributes to the quality and resiliency inherent in the development of manufacturing processes, while also listening to customer needs to produce best-fit solutions.

Intel Optane persistent memory (PMem), in combination with Intel's open-source distributed asynchronous object storage (DAOS) solution, sets a new world record, soaring to the top of the Virtual Institute for I/O IO-500 list. With just 30 servers of Intel Optane PMem, Intel's DAOS solution defeated today's best supercomputers and now ranks No. 1 for file system performance worldwide. These results validate the solution's delivery as having the most performance of any distributed storage today. They also demonstrate how Intel is truly changing the storage paradigm by providing customers the persistence of disk storage with the fine-grained and low-latency data access of memory in its Intel Optane PMem product.

"The recent IO-500 results for DAOS demonstrate the continuing maturity of the software's functionality enabled by a well-managed code development and testing process. The collaborative development program will continue to deliver additional capabilities for DAOS in support of Argonne's upcoming exascale system, Aurora," said Gordon McPheeters, HPC systems administration specialist at the Argonne Leadership Computing Facility.

GIGABYTE Technology,, an industry leader in high-performance servers and workstations, today announced the launch of the GIGABYTE G242-Z11 with PCIe 4.0, which adds to an already extensive line of G242 series servers, designed for AI, deep learning, data analytics, and scientific computing. High-speed interfaces such as Ethernet, Infiniband, and PCI Express rely on fast data transfer, and PCIe 3.0 can pose a bottleneck in some servers. With the expansion of the AMD EPYC family of processors comes PCIe Gen 4.0, which is valuable to servers so as not to bottleneck high bandwidth applications. The 2nd Gen AMD EPYC 7002 processors have added PCIe Gen 4.0, and GIGABYTE has included an ever-evolving line of servers to accommodate the latest technology.

The G242-Z11 caters to the capabilities of 2nd Gen AMD EPYC 7002 series processors. The G242-Z11 is built around a single AMD EPYC processor, and this even includes the new 280 W 64-core (128 threads) AMD EPYC 7H12. Besides a high core count, the 7002 series has 128 PCIe lanes and natively supports PCIe Gen 4.0. It offers double the speed and bandwidth when compared to PCIe 3.0. Having PCIe 4.0 allows for 16GT/s per lane and a total bandwidth of 64 GB/s. As far as memory support, the G242-Z11 has support for 8-channel DDR4 with room for up to 8 DIMMs. In this 1 DIMM per channel configuration, it can support up to 2 TB of memory and speeds up to 3200 MHz.

Chenbro has launched the RB23708, a Level 6, 2U rackmount server barebone designed for mission-critical, storage-focused applications in Data Center and HPC Enterprise. The RB23708 is pre-integrated with an Intel Server Board S2600WFTR that supports up to two 2nd Generation Intel Xeon Scalable "Cascade Lake" Processors.

The RB23708 is an easy-to-use barebones server solution that pre-integrates a 2-socket Intel Server Board to ensure a flexible, scalable design with mission-critical reliability. Notably, it offers Apache Pass, IPMI 2.0 & Redfish compliance, and includes Intel RSTe/Intel VROC options, providing an ideal solution for hosting Video, IMS, SaaS and similar storage-focused applications.

Asetek today announced a collaboration with Hewlett Packard Enterprise (HPE) to deliver its premium data center liquid cooling solutions in HPE Apollo Systems, which are high-performing and density-optimized to target high-performance computing (HPC) and Artificial Intelligence (AI) needs. The integration enables deployment of high wattage processors in high density configurations to support compute-intense workloads.

When developing its next-generation HPC server solutions, HPE worked closely with Asetek to define a plug and play HPC system that is integrated, installed, and serviced by HPE that serves as the ideal complement to HPE's Gen10 Plus platform. With the resulting solution, HPE is able to maximize processor and interconnect performance by efficiently cooling high density computing clusters. HPE will be deploying these DLC systems, which support warm water cooling, this calendar year.

Today, The Khronos Group, an open consortium of industry-leading companies creating graphics and compute interoperability standards, announced its SYCL 2020 Provisional Specification, for which Intel has made significant contributions through new programming abstractions. These new capabilities accelerate heterogeneous parallel programming for high-performance computing (HPC), machine learning and compute-intensive applications.

"The SYCL 2020 Provisional Specification marks a significant milestone helping improve time-to-performance in programming heterogeneous computing systems through more productive and familiar C++ programming constructs," said Jeff McVeigh, vice president of Datacenter XPU Products and Solutions at Intel Corporation. "Through active collaboration with The Khronos Group, the new specification includes significant features pioneered in oneAPI's Data Parallel C++, such as unified shared memory, group algorithms and sub-groups that were up-streamed to SYCL 2020. Moving forward, Intel's oneAPI toolkits, which include the SYCL-based Intel oneAPI DPC++ Compiler, will deliver productivity and performance for open, cross-architecture programming."

Western Digital today announced new solutions that provide the foundation for next-generation data infrastructures designed around Ultrastar NVMe SSDs and supercharged by NVMe-oF. Building upon the company's innovative design and integration capabilities - from NAND Flash to storage platforms - the new dual-port, performance Ultrastar DC SN840 NVMe SSDs and in-house RapidFlex NVMe-oF controllers are standalone solutions that combine to create the new OpenFlex Data24 NVMe-oF Storage Platform, a new shared storage JBOF (Just a Bunch of Flash) enclosure that extends the value of NVMe to multiple hosts over a low-latency Ethernet fabric network. These new solutions further expand Western Digital's data center portfolio to help customers transition to higher efficiency NVMe SSDs and more advanced shared storage architectures to meet the evolving demands of performance-driven applications and workloads.

In today's global digital economy, microseconds count. As hyperscale cloud and enterprise data centers constantly work to remove bottlenecks to ensure uncompromised performance and availability of essential applications, while also keeping pace with unprecedented data growth, customer adoption of NVMe and NVMe-oF solutions continues to accelerate. Industry analyst firm IDC expects hyperscalers, OEMs and end-user IT organizations to continue to transition away from legacy SATA and SAS interfaces, with NVMe on track to reach more than 55 percent of total enterprise SSD units shipped in 2020 and grow at a 2018-2023 CAGR of 38 percent.

ASUSTek, the leading IT Company in server systems, server motherboards and workstations today announced the new NVIDIA A100-powered server - ESC4000A E10 to accelerate and optimize data centers for high utilization and low total cost of ownership with the PCIe Gen 4 expansions, OCP 3.0 networking, faster compute and better GPU performance. ASUS continues building a strong partnership with NVIDIA to deliver unprecedented acceleration and flexibility to power the world's highest-performing elastic data centers for AI, data analytics, and HPC applications.

ASUS ESC4000A-E10 is a 2U server powered by the AMD EPYC 7002 series processors that deliver up to 2x the performance and 4x the floating point capability in a single socket versus the previous 7001 generation. Targeted for AI, HPC and VDI applications in data center or enterprise environments which require powerful CPU cores, more GPUs support, and faster transmission speed, ESC4000A E10 focuses on delivering GPU-optimized performance with support for up to four double-deck high performance or eight single-deck GPUs including the latest NVIDIA Ampere-architecture V100, Tesla, and Quadro. This also benefits on virtualization to consolidate GPU resources in to shared pool for users to utilize resources in more efficient ways.

Samsung Electronics Co., Ltd., a world leader in advanced semiconductor technology, today announced the launch of 'Samsung Advanced Foundry Ecosystem (SAFE ) Cloud Design Platform (CDP)' for fabless customers, in collaboration with Rescale, a leader in high performance computing (HPC) applications in the cloud. The key highlight feature of Samsung foundry's first SAFE Cloud Design Platform is that it provides a virtual environment to design chips in the cloud. By accessing this platform through the cloud, customers can immediately start designing at anytime and anywhere.

To maximize customers' design convenience, SAFE CDP supports a very secure design condition that has verified with cloud companies. In addition, customers can utilize various Electronic Design Automation (EDA) tools offered by multiple vendors such as Ansys, Cadence, Mentor, a Siemens business and Synopsys. Gaonchips, one of Samsung Foundry's Design Solution Partners, has already tested the SAFE CDP on its 14 nm automotive project using Cadence's Innovus Implementation System and has successfully reduced its design run-time by 30 percent compared to current on-premise execution.

Mark Papermaster, chief technology officer and executive vice president of Technology and Engineering at AMD, today confirmed that CDNA is on-track for release in 2H2020 for HPC computing. The confirmation was (adequately) given during Dell's EMC High-Performance Computing Online event. This confirms that AMD is looking at a busy 2nd half of the year, with both Zen 3, RDNA 2 and CDNA product lines being pushed to market.

CDNA is AMD's next push into the highly-lucrative HPC market, and will see the company differentiating their GPU architectures through market-based product differentiation. CDNA will see raster graphics hardware, display and multimedia engines, and other associated components being removed from the chip design in a bid to recoup die area for both increased processing units as well as fixed-function tensor compute hardware. CNDA-based Radeon Instinct MI100 will be fabricated under TSMC's 7 nm node, and will be the first AMD architecture featuring shared memory pools between CPUs and GPUs via the 2nd gen Infinity Fabric, which should bring about both throughput and power consumption improvements to the platform.

Kingston Digital, Inc., the Flash memory affiliate of Kingston Technology Company, Inc., a world leader in memory products and technology solutions, today began shipping the 2 TB KC2500 M.2 NVMe PCIe SSD for desktop, workstations and high-performance computing (HPC) systems. The new, higher capacity gives customers the flexibility to keep up with increasing storage needs and improves workflow.

KC2500 combines outstanding performance with endurance using the latest Gen 3.0 x 4 controller and 96-layer 3D TLC NAND, with speeds up to 3,500 MB/s read and up to 2,900 MB/s write. The self-encrypting SSD supports a full-security suite for end-to-end data protection using AES-XTS 256-bit hardware-based encryption and allows the usage of independent software vendors with TCG Opal 2.0 security management solutions such as Symantec, McAfee, WinMagic and others. KC2500 also has built-in Microsoft eDrive support, a security storage specification for use with BitLocker.

AMD and technology partner Penguin Computing Inc., a division of SMART Global Holdings, Inc, today announced that New York University (NYU), Massachusetts Institute of Technology (MIT) and Rice University are the first universities named to receive complete AMD-powered, high-performance computing systems from the AMD HPC Fund for COVID-19 research. AMD also announced it will contribute a cloud-based system powered by AMD EPYC and AMD Radeon Instinct processors located on-site at Penguin Computing, providing remote supercomputing capabilities for selected researchers around the world. Combined, the donated systems will collectively provide researchers with more than seven petaflops of compute power that can be applied to fight COVID-19.

"High performance computing technology plays a critical role in modern viral research, deepening our understanding of how specific viruses work and ultimately accelerating the development of potential therapeutics and vaccines," said Lisa Su, president and CEO, AMD. "AMD and our technology partners are proud to provide researchers around the world with these new systems that will increase the computing capability available to fight COVID-19 and support future medical research."