Pushing the limits in Computing. Like this. http://www.zdnet.com/article/obama-orders-creation-of-exascale-supercomputer/ President Obama is seeking the power of exascale computers to power the country's government departments in the future. In a new executive order, the president has issued demands for a new initiative which will focus exclusively on supercomputing research, released by the White House on Wednesday. Titled "Creating a national strategic computing initiative," the president's order outlines plans to create the world's first exascale computing system in order to establish the country's position in high-performance computing (HPC) research and development. An exascale system would be capable of performing 1 million trillion floating-point operations per second -- otherwise converted as 1,000 petaflops or 1 exaflops -- which is far quicker than today's supercomputers. However, such a system would need an architecture capable of combining thousands of high-power processors and CPUs, a complete overhaul of how today's computers operate, and would require a fortune in energy and building costs based upon today's system usage rates. As reported by IEEE Spectrum, an exascale computer could be built and this week, but current systems simply aren't tenable. Steve Scott, senior vice president and chief technology officer at Cray told the publication: "To some degree it depends on how much money a country is willing to spend. You could build an exaflop computer tomorrow, but it'd be a crazy thing to do because of the cost and energy required to run it." Despite this, HPC research has to start somewhere. Obama's order has led to the launch of the National Strategic Computing Initiative (NSCI), a "whole-of-government effort designed to create a cohesive, multi-agency strategic vision and Federal investment strategy, executed in collaboration with industry and academia, to maximize the benefits of HPC for the United States." The initiative will primarily be a partnership between the US Department of Energy (DOE), the Department of Defense (DOD), and the National Science Foundation (NSF), although the private sector will also be consulted. The US president believes that increased demand for computing power and maintaining economic growth requires the high performance of next-generation computers. As a result, the new federal strategy aims to propel research in this area forward through orders for the US to deploy and apply HPC technologies across the board, the fostering of growth for public-private collaborative research efforts and increased communication between different government departments, academic institutions and companies. In addition, Obama says the US must develop a "comprehensive technical and scientific approach" to extend HPC research into hardware, system software, development tools, and applications. Ultimately, the NSCI is tasked with the development of an exascale computing system which can deliver approximately 100 times the performance of current 10 petaflop systems across a range of applications for government use. The initiative is also expected to provide, "over the next 15 years, a viable path forward for future HPC systems even after the limits of current semiconductor technology are reached (the "post- Moore's Law era")," according to the order. If exascale supercomputers become reality through the initiative, NASA, the Federal Bureau of Investigation (FBI), the National Institutes of Health, the Department of Homeland Security, and the National Oceanic and Atmospheric Administration (NOAA) will benefit the most -- and have the option to pitch in their own ideas, requirements and applications through the research process. While there is no definitive timeline for the NSCI to pull an exascale computer out of the hat, the group's council is required to submit a plan for the development of exascale computers within 90 days of the order being issued and check-in with updates annually.
What they should do is get a competitor to Intel. They might actually try building high performance chips. When AMD was relevant at least they tried making faster chips.
No, this is to go beyond the current computers. NSA computer are hardly legacy. Their IT might be, but they have to computers that none of us civilian have. This is more of, breaking out of the pack and leaping forward. Every country is now capable of the same massive super computer with China being the lead several times. You want to leap frog everyone, you have to do something impossible.
AMD still make fast chips. They just don't have any market appeal. Their problem is more leadership, internal political fights than technology. They have the technology and the know-how, but is a horribly managed company with so many bad directions over the past few years. The once pretty good ATI brand is also being dragged down with them.
We're nearing the transistor size barrier. Moore's law is dead unless there is some massive unforeseen breakthrough in the coming years.
In terms of HPC, Nvidia is a major player and competitor. The current top supercomputer in US based on the Top 500 (Titan) relies on AMD host processors and Nvidia GPU.
It would be like that end scene in 2001 space odyssey. Exascale is worth at least wiki read. The challenges at that level goes beyond computer chips. We're talking about series of tubes on the thousands thousands scale with what we think we have in laptops and cables. That itself takes energy and cooling on another level besides the problem of getting it that fast and uniformly in the first place.
I left out the geeky stats in the links, but these numbers are getting cuckoo insane. They aren't even going wild on spending like they would during a war or something. An exaflop is 1,000,000,000,000,000,000 https://www.tomshardware.com/news/a...the-exascale-barrier-now-fastest-in-the-world Frontier also now ranks as the fastest AI system on the planet, dishing out 6.88 ExaFlops of mixed-precision performance in the HPL-AI benchmark. That equates to 68 million instructions per second for each of the 86 billion neurons in the brain, highlighting the sheer computational horsepower. https://www.nextplatform.com/2022/05/30/frontier-step-by-step-over-decades-to-exascale/ Any time you build anything with more than 60 million parts, it is going to be a headache. And if you have to create a space in a datacenter and build an exascale system with all of those parts, each of which is crucial, during a global pandemic, it gets incredibly tougher. And then the panic sets in because there is only one way to test an exascale system because no one has ever built one before. And that is throw that big red switch, wince, and find out. And that is what Oak Ridge National Laboratory did when it fired up the “Frontier” exascale system – or more precisely, most of what it will eventually scale to – as it geared up to run the High Performance Linpack benchmark test to prove that it had broken the exascale barrier for the June 2022 edition of the Top 500 supercomputer rankings. The Oak Ridge Leadership Computing Facility outside of Knoxville, Tennessee is a massive datacenter that has over 100 megawatts of juice running into it. The lab has been a direct electricity customer of the Tennessee Valley Authority, which has nuclear, coal, and hydro plants to supply its power, since it was established in 1942 to do the uranium enrichment and plutonium processing for the Fat Man and Little Boy atomic weapons used to conclude World War II. While Oak Ridge can deploy up to 100 megawatts for its computing, it costs roughly a dollar per watt per year to do this – so $100 million – and that sure does add up fast. Which is why it is significant that Frontier is hugely more energy efficient than the “Titan” hybrid CPU-GPU supercomputer that it is being measured against. ... As configured for the Linpack run, the Frontier system had a power draw of 21.1 megawatts, but what you may not know is that when the Frontier machine is first set loose on the Linpack, according to Zacharia, it draws an extra 15 megawatts of power as it is starts the job. That’s enough to power a small city in its own right, although it is a far cry from the 3,000 megawatts – that’s three whole nuclear plants of juice – that Zacharia said it would have taken to build an exascale system using the Cray XK7 system with its 16-core Opteron 6274 CPUs and Nvidia K20X GPUs back in 2012. By the way, what Zacharia did not say is that would have meant that the exascale Titan would have cost $4.22 billion in 2021 US dollars to build an another $3 billion a year to operate, which explains why no one built an exascale system back in 2012. Or 2015, or even 2018. We had to wait for the technology to advance to get the time machine that lets us predict into the future. . . . ... “I have always looked at supercomputing at this scale as a time machine, allowing you to go forward in time 20 years into the future, not to just see the future, but to influence and shape how the future evolves,” explained Zacharia. “In the end, that is what this all means. Yes, it’s nice to celebrate exascale – that is worthy of celebrating. But it is really about peeking into the future and seeing how we can shape and influence the future for a better tomorrow.” Can you imagine the next console will meet or exceed the 2004 offering? Xbox Series S is benched around 12 Tflops.
The Frontier system is crazy because it not only destroyed the top supercomputer -- but it did so efficiently. It's also possibly the most efficient supercomputer. Of course, China probably already has exascale systems up and running. Maybe not as efficient, but they just throw more hardware at it and get it there. Much of the HPC community used to openly use TOP500.org as the barometer of the most powerful HPC's on the planet, but nowadays there's belief that secrecy is more important to some nations.
While it is amazing that these types of supercomputers can be made, what exactly do these get used for once the shine of "We did it!" wears off?
It's used for scientific applications, like physics modeling and simulations. The higher the flops, the higher the accuracy when running those simulations for a given period of time. They are also designed to support many concurrent users who submit their jobs through a workload manager.
They're pretty important in many science fields nowadays and by no means a "we did it" product. https://www.frontier-enterprise.com/what-are-the-worlds-10-fastest-supercomputers-being-used-for/
