Items tagged with supercomputer

A few weeks ago, Azure received an NVIDIA A100 upgrade to its virtual machines. Announced at GTC, NVIDIA is now building its supercomputer with the A100 platform across the pond. This behemoth, dubbed Cambridge-1, will be the most powerful supercomputer operating within the United Kingdom. It will become available to U.K healthcare researchers using AI to solve complex issues, such as COVID-19, by the end of the year. The Cambridge-1 supercomputer will be an NVIDIA DGX SuperPOD supercomputer with over 400 petaflops of AI performance and eight petaflops of Linpack performance. This sort of horsepower would allow it to rank 29th on the TOP500 list of most powerful supercomputers in the world. Inside,... Read more...
NVIDIA has partnered with the University of Florida (UF) to build an artificial intelligence (AI) supercomputer that will be the fastest in all of academia. Or more specifically, it will be capable of a whopping 700 petaflops of AI performance, made possible by enhancing the university's existing HiPerGator supercomputer with NVIDIA's DGX SuperPod architecture. This essentially represents an infusion of NVIDIA's DGX A100 systems, each of which wields eight Tensor Core GPUs based on the company's 7-nanometer Ampere architecture. And each of those GPUs touts 54 billion transistors capable of delivering 5 petaflops of performance in a single node, along with the ability to handle 1.5TB of data per... Read more...
The 55th edition of the Top500 supercomputers list was published, and a supercomputer from Japan has taken the top spot on the list. The supercomputer is called Fugaku, and it turned in a High Performance Linpack (HPL) result of 415.5 petaflops beating the Summit supercomputer by a factor of 2.8 times to take the top spot. The Japanese supercomputer is powered by Fujitsu 48-core A64FX SoC and is the first machine to sit at the head of the Top500 supercomputer list that's powered by Arm processors. Fugaku delivered extreme performance in single or further reduced precision, which is a workload often used in machine learning and AI applications, with a peak performance of over 1000 petaflops. The... Read more...
A number of supercomputers across Europe have been targeted by malware that focuses on mining for cryptocurrency (Monero). The malware has forced supercomputers in the UK, Germany, and Switzerland to be shutdown as operators investigate the security incidents. The high-performance computing center in Spain was also reportedly targeted by a malware attack. The first reported attack surfaced last Monday and came from the University of Edinburg, home of the ARCHER supercomputer. The university reported that there was a "security exploitation on the ARCHER login nodes." ARCHER operators shutdown the system for an investigation, and all SSH passwords were reset to prevent further intrusions. In Germany,... Read more...
AMD just announced today a design win in conjunction with the Lawrence Livermore National Laboratory (LLNL) and Hewlett-Packard Enterprise (HPE), win which it will be providing the hardware powering the El Capitan exascale-class supercomputer. What caught our eye about this announcement was not the compute performance -- which will be enormous at 2 exaflops double precision -- but the CPUs and GPUs that will be inside. According to AMD, its next-generation Radeon Instinct GPU accelerators will be deployed featuring advanced high bandwidth memory (HBM). The company will also be using third-generation Infinity Fabric Architecture that provides a high-bandwidth, low-latency interconnect to string... Read more...
The United States Navy is enlisting a multi-million dollar Cray Shasta supercomputer powered by hundreds of thousands of AMD EPYC server processor cores and over 100 NVIDIA Volta GPUs. It ranks as the largest, most capable supercomputing system every deployed by the Department of Defense (DoD). Cray Shasta houses 290,304 EPYC 7002-series processors core and 112 Volta V100 general purpose graphics processing units (GPGPUS). The hardware is all interconnected by way of a 200 gigabit per second (Gbps) Cray Slingshot network. To round things out, Cray Shasta is outfitted with 590 total terabytes (TB) of memory and 14 petabytes (PB) of usable storage, including 1PB of NVMe-based solid date drive storage.... Read more...
In sticking with the plan that was announced last summer, Indiana University is now the owner of Cray's first "Shasta" supercomputer. As shown in the image above, it bears the university's color scheme and is called Big Red 200, which is a tribute to the 200th anniversary of IU. In addition to oozing school pride, Big Red 200 is notable because it will rock NVIDIA's next-generation Ampere GPUs. NVIDIA has not yet released Ampere, but is expected to soon. We have covered various leaks and rumors on the subject, as it pertains to NVIDIA's efforts in the gaming space, but Ampere will also find a home in datacenters and supercomputer applications. Or at least that is the expectation, based on what... Read more...
If you just built a high-end desktop (HEDT) configuration based on a Threadripper or Core-X processor, you may want to avert your eyeballs somewhere else. At the very least, brace yourself for some numbers to follow. Your system is still fast, mind you, just not as fast as a pair of EPYC 7H12 processors in a Cray Shasta supercomputer performing a Geekbench run. Before we get to the results, let's talk specs. AMD offers three EPYC processors with 64 physical cores and 128 threads of computing muscle, the EPYC 7H12 is the fastest of the bunch—it has a 2.6GHz base clock and 3.3GHz max boost clock, to go along with 256MB of L3 cache. Here's how it stacks up with a few other EPYC 7002 processors... Read more...
AMD has definitely been on an serious upswing in recent years thanks to its Zen CPU microarchitecture, and the company has gotten the attention from some big names in the computing industry. Now, AMD is definitely showing its computing muscle thanks to a collaboration with the Cray, the Oak Ridge National Laboratory (ORNL) and the U.S. Department of Energy (DOE).  It has been announced that AMD's EPYC processors will power what will be the world's fastest exascale supercomputer. Called Frontier, this new supercomputer will deliver in excess of 1.5 exaflops of compute performance and will be delivered to ORNL rough two years from now. Although AMD is currently being a bit light... Read more...
The hardware may still be in development, but Intel just landed a major design win nonetheless, one that will reportedly culminate in the world’s first Exascale Supercomputer powered partially by Intel’s upcoming Xe graphics technology. It was just announced that Intel and the U.S. Department of Energy (DOE) will deliver the first supercomputer capable of one exaFLOP of compute performance, or a “quintillion” floating point operations per second. The system, named “Aurora”, is being developed at the DOE’s Argonne National Laboratory in Chicago. It is being designed for High-Performance Computing and Artificial Intelligence applications and will be used... Read more...
The brain is the most complex organ in the body and the most difficult to unravel. Scientists have developed a variety of ways to better understand the brain, including the use of supercomputers. The world’s largest neuromorphic supercomputer, the Spiking Neural Network Architecture (SpiNNaker), was just switched on for the first time yesterday. The mammoth array of server cabinets boasts one million processor cores and is able to perform 200 trillion actions per second. SpiNNaker has been twenty years and £15 million (nearly 19.5 million USD) in the making. The project was originally supported by the Engineering and Physical Sciences Research Council (EPSRC), but has been most recently... Read more...
Samsung is reportedly wrapping up development of an all-new in-house GPU design, which has reportedly been dubbed SGPU. SGPU is a brand-new GPU design that potentially will be able to scale from devices like smartphones to applications in the supercomputing realm. Dr. Chien-Ping Lu, known as CP, has been brought in by Samsung to help put the finishing touches on the product and to ensure that it has a successful launch. CP is a known quantity in the tech space, having been involved with the development of NVIDIA's ahead-of-its-time nForce IGP along with his subsequent work at MediaTek. According to Graphic Speak, CP is leading Samsung engineering teams based in Austin and San Jose dedicated... Read more...
NVIDIA and IBM have helped the United States lay claim to the fastest supercomputer ever made. Called Summit, the remarkable machine sits at the US Department of Energy's Oak Ridge National Laboratory (ORNL) where it is capable of crushing workloads via a staggering 200 petaflops of high-precision computation, making it eight times more powerful than Titan, previously the fastest supercomputer in the US. "Summit is fast, but what Summit means is even more remarkable," NVIDIA CEO Jensen Huang said at Summit’s debut. "Summit is a magnificent scientific instrument that will attract the world’s great scientists...Summit is a milestone in a global race. Not a race to space, but a race... Read more...
Researchers from Oregon State University (OSU) have announced that they have created a new material that is a key step in creating the next generation of supercomputers. These next-generation quantum computers will be far faster than current supercomputers while consuming much less energy. The scientists on the project work in the OSU College of Science and the compound they have developed is an inorganic material that adopts a crystal structure. The material is capable of sustaining a new state of matter that is known as quantum spin liquid. That new compound is lithium osmium oxide and it has osmium atoms that form a honeycomb-like lattice, and can enforce a phenomenon called "magnetic... Read more...
The Top500 list of supercomputers globally was published again this month, and one of the most interesting tidbits is that every single one of them runs Linux. This list comes out every six months and shows the top supercomputers in the world along with the countries with the most supercomputers. This fiftieth edition shows that there are some changes afoot in global supercomputing power. China has traditionally held the top spot for the most powerful supercomputer in the world, and it continues to hold that spot on the most current list. However, the U.S. has traditionally had the most supercomputers on the list on a per country basis, but that has changed: China now has the most supercomputers... Read more...
When it comes today's high-end processors and graphics cards, it's a well-known fact that cooling is crucial to maintaining performance and preventing your hardware from having a premature meltdown. In the case of modern chips, that cooling comes in the form of heatsinks or water-cooling blocks that are directly attached to the wick away heat. However, Purdue University researchers have developed a rather novel way to cool high-performance chips more efficiently. The researchers incorporated tiny microchannels directly into the chip to enhance cooling performance. These microchannels are just a few microns wide and help to circulate liquid coolant to more efficiently regulate chip temperatures.... Read more...
Cray and Microsoft have teamed up to make supercomputing performance readily available for enterprise customers. Cray and Microsoft are offering customers dedicated Cray supercomputing systems in Microsoft Azure datacenters to enable users to run AI, advanced analytics, and modeling and simulation workloads while seamlessly connected to the Azure cloud. Cray wrote, "Cray’s tightly coupled system architecture and Aries interconnect addresses the exponential demand for compute capability, real-time insights, and scalable performance needed by enterprises today. With this new partnership, Cray and Microsoft have made it easier for cloud customers to harness the power of supercomputing and multiply... Read more...
Many computer hardware vendors have tossed around the expression "a supercomputer in your home!", and every time it's used, it's really incorrect - unless we're talking about comparing current tech to the supercomputers of olde. Well, if anyone ever said, "a supercomputer in space!", they can soon congratulate themselves, thanks to a joint partnership between HP Enterprise and SpaceX. On Monday, August 14th, at 12:31 EST, a SpaceX CRS-12 mission will take off from Cape Canaveral and send a Dragon resupply capsule to the International Space Station (ISS) to drop off some usual supplies - laptops, freeze-dried ice cream, and oh, a supercomputer. In fact, it's one of HP's aptly-named Apollo 40 supercomputers.... Read more...
It was bad enough when the Chinese leapfrogged the United States in the supercomputer field, but now it looks as though Switzerland has come in and poured salt on America’s wounds. This morning, the TOP500 list was refreshed and it’s not good news for the U.S. There were no changes the very top of the rankings, with China taking the number 1 and number 2 positions.  Sunway TaihuLight, which was commissioned by China’s National Research Center of Parallel Computer Engineering & Technology (NRCPC) still tops the list with Linpack performance coming in at 93 petaflops. Second place is still held by another Chinese resident, the Tianhe-2, which comes in at 33.9 petaflops. The shakeup comes... Read more...
With the Chinese dominating the charts when it comes to the world’s most powerful supercomputers, the United States is looking to leverage its homegrown talent to usher in next-generation exascale computing. The Department of Energy (DOE) Exascale Computing Project (ECP) is funding six of the country’s top chip firms in order to ensure U.S. leadership in the field. Those companies include AMD, Intel, NVIDIA, Hewlett-Packard Enterprise, IBM and Cray. The six companies will have access to $258 million in funds that will be dispersed over a three-year-period. According to DOE, the funds will be used by the six companies to boost research and development in hardware technology, software technology... Read more...
The supercomputer race has only just begun. Japan is currently planning to build the world’s fastest supercomputer. The machine, nicknamed AI Bridging Cloud Infrastructure (ABCI), would ideally be able to process 130 quadrillion calculations per second or 130 petaflops. Japanese Prime Minister Shinzo Abe has called for scientists, companies, and bureaucrats to work together in order to dominate robotics, renewable energy, and other up and coming industries. The Ministry of Economy, Trade and Industry will spend 19.5 billion yen ($173 million) on the supercomputer project and should commence in early 2017. IBM's supercomputer WatsonJapan is particularly interested in artificial intelligence such... Read more...
Artificial intelligence is at the center of several evolving and next-generation products and services, everything from video games and smartphones (virtual digital assistants and messaging) to self-driving cars and web searchers. It has the attention of companies small and large, including NVIDIA and Microsoft, the two of which are partnering to push AI technologies more aggressively into the enterprise. What this essentially boils down to is pairing NVIDIA's DGX-1 supercomputer with Microsoft's Azure cloud platform. The idea is to "help companies join the AI revolution" and this partnership means enterprises now have an AI platform that spans from their data center to Microsoft's cloud. "Every... Read more...
1 2 3 Next