University of Texas is Getting a $60 Million Supercomputer (cnet.com) 88
The University of Texas at Austin, will soon be home to one of the most powerful supercomputers in the world. From a report: The National Science Foundation awarded a $60 million grant to the school's Texas Advanced Computing Center, UT Austin and NSF said Wednesday. The supercomputer, named Frontera, is set to become operational roughly a year from now in 2019, and will be "among the most powerful in the world," according to a statement. To be exact, it will be the fifth most powerful in the world, third most powerful in the US, and the most powerful at a university.
Dang (Score:3, Funny)
I can't wait to play Quake on that thing.
Re: (Score:3)
I heard it's fast enough that it can run Crysis.
Re: (Score:2)
"r_smp=one billion?"
But... (Score:3)
Re: (Score:1)
will they name it "HAL"?
I'm afraid they can't do that.
Supercomputers don't excite me anymore (Score:5, Interesting)
Back in the day, supercomputers used to be about cutting edge system architecture, making CPUs as absolutely fast as possible, and even shortening connecting wires in the system to squeeze every last bit of performance out of a system. Think back to the Cray systems and such.
These days, supercomputers are just about who can spend the most money to build the biggest data center and buy the largest number of generic blade servers. It's just not interesting anymore; whoever can spend the most money will have the fastest system simply because they can buy the most blades.
Re: (Score:2, Interesting)
Interconnect still matters.
Anyone can do it with generic ethernet.
Hell, I have a "supercomputer" with 8x 4core raspberry pi's. It's faster than the early crays, and certainly has more memory.
Then we go up a few notches, and go into fiberchannel, inifniband, and other high performance stuff.
it's just not about blades, it's about interconnect, memory utilization, data locality, algorithmic complexity, etc.
Heck, even CERN is known to use a "small" raspberry pi cluster to tune algorithms.
Re: (Score:2)
It's just not interesting anymore; whoever can spend the most money will have the fastest system simply because they can buy the most blades.
I was more excited back when System X was built (aka "Big Mac"). It was able to make #3 on the list for less than $6,000,000.
https://en.wikipedia.org/wiki/... [wikipedia.org]
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Agreed, which I fear makes me sound like a "back in my day" curmudgeon. The liquid cooled Crays were the coolest thing and way cooler than sci-fi stuff. But I was a kid, dreaming of the day I'd program a Cray, not knowing I'd have more than Cray-1 power under my fingers in a laptop. But with far more code for it to wade through...
I don't see it in TFA, but a quick search reveals it appears to be a huge pile of Dell blades, which makes sense they'd buy from Dell. It'd be nice to see some specs: Intel or
Re: (Score:3)
The liquid cooled Crays were the coolest thing
Cool for their day, but my iPhone has way more compute capacity today. Custom CPUs can't compete with a 14nm fab, and never will again.
it appears to be a huge pile of Dell blades
It is more than that. What makes it a "supercomputer" is the fast interconnects between the blades.
It'd be nice to see some specs: Intel or AMD CPUs?
Who cares? Most of the compute capacity is in the GPU, not the CPU. The press release mentions Nvidia.
This is just a funding announcement. It is likely light on tech details because the details haven't actually been ironed out yet.
Re: (Score:1)
Re: (Score:2)
whoever can spend the most money will have the fastest system simply because they can buy the most blades
My best Speed Racer voice:
And you will see that I will spend the most money and have the fastest system because I have the most blades because of the most money and therefore I have the fastest system and you did not spend the most money and therefore I did and I have the fastest system you will see ha ha.
Fastest compuer (Score:1)
Is in China. Because China has eclipsed the US in every conceivable way possible. In 50 years the US will be speaking Mandarin and there's nothing you can do about it.
Re: Fastest compuer (Score:2)
Most people in USA can't even speak English properly. Good luck teaching them Chinese.
Re: (Score:2)
Most people in USA can't even speak English properly. Good luck teaching them Chinese.
If you learn Chinese as a child, it is easier than English. The grammar is simpler, there are no irregular verbs, and the pronouns are drop-dead simple. You don't need regionalisms like "y'all" to make up for a lack of a second person plural, or singular "their" to make up for the lack of a gender neutral third person pronoun. There is no difference between subjective pronouns (I, we, they, who) and objective (me, us, them, whom).
I speak both. Chinese is better for haggling and insults. English is bette
Re: (Score:2)
Stupid question for us plebes in the US... where is a good place to get started learning Chinese, both written and conversational?
If you are an adult English speaker, learning Chinese is going to be REALLY hard. Three reasons:
1. The characters. You need to memorize 1500 for basic literacy, and 3000 to match a college educated Chinese citizen.
2. The tones. Inflection changes the meaning of words, and it is really hard for an adult brain to adapt to this.
3. The idioms. English has idioms like "raining cats and dogs" that you just have to memorize. Chinese has WAY more. Personally, I think the idioms are fun to learn, but there
outdated stereotype (Score:2)
Per the June 2018 Top 500 List, [wikipedia.org] the US owns the title of fastest supercomputer. In fact, 6 of the top 10 on the list are installed in the United States. And for those opining the days of Cray's dominance in this space, I'll point out that several of the systems in the top 10 are identified with Cray as the vendor. Just two computers in the top 10 are hosted in Jackie Chan's home country.
Re: (Score:2)
Yeah, everyone knows about Trump, okay?
pointless purchase (Score:2, Funny)
Look UT Austin, just because you bought yourself a supercomputer doesn't mean its going to be enough to help the Longhorns beat the Sooners no matter how many xFLOPS it can do, or if it can run Witcher 3 in 4k smoothly.
Re: (Score:2)
A 60 million dollars supercomputer? (Score:2)
How big is a beowulf cluster of twelve million Raspberry Pi Zero?
What can a user access? (Score:5, Interesting)
Re:What can a user access? (Score:5, Insightful)
Well more to the question, What is the university using the supercomputer for? Is it just 60 million dollar bragging rights, or did they get some grants for research project(s) that can cover the cost that could have effects to make it worth the cost?
Using a Supercomputer for a Shared system is general a waste of money, and you are better off with just a server farm, or (gasp) a cloud service (which is a server farm hosted remotely). However if there is a project that really is utilizing the full computer then a Super Computer is needed.
Re: (Score:3)
Blue Waters is dead! Long live... Frontera (Score:5, Interesting)
I'm an atmospheric scientist who has been using federal supercomputing hardware to better understand thunderstorms [orf.media] for years. Blue Waters is the current "Leadership Class" NSF-sponsored supercomputer. My Blue Waters allocation is currently winding down, and I can speak to how great it has been as a machine that has enabled (I know it's a cliche, but it's true) breakthrough science. A typical Blue Waters node contains 16 floating point AMD cores and 64 GB of memory. Many of the Blue Waters nodes contain a GPU, but it's miles behind the times since the machine was created about 7 years ago.
Frontera (for some reason the Canyonera theme song plays in my head) is the Phase 1 machine for the next Leadership Class supercomputer. The Phase 1 machine is supposed to come on line in 2019 and hold us over until 2024 when the next machine will come on line. When you look at how much money is being spent on Fronterra, and you compare it to Blue Waters, you realize that the vendor is being asked to create a much more powerful machine for a fraction of the price. What this will mean in practice, and what most of the scientific computing world is not ready for, is that a large bulk of the FLOPS on this new machine will be GPU flops. GPUs are not easy to use for doing heavy lifting (say, fluid dynamics solvers) using existing code. So a lot of people are going to have to decide whether to try to shoehorn their current MPI only (or MPI + some OpenMP) code to MPI, OpenMP + OpenACL (or nvidia CUDA), or to start from scratch (nobody wants to start from scratch). You have to remember that the vast majority of us scientists are NOT trained computer scientists, and most of us code for shit. I am off to a hackathon at NCSA in a couple weeks with some students to optimize some radiation code for GPUS... I spend half of my time doing computer stuff, and the other half doing science (and the other other half writing proposals, etc.).
So for those of you who aren't excited about new supercomputers, or don't understand their true power, I'm here to say that it's currently a very exciting time to be a numerical modeler, if you're willing to learn a bit on how to best wrestle these supercomputers into submission. I've spend over a decade just figuring out the most efficient way to write, organize, and analyze the TB-PB of data that a high resolution model can produce, and trying to make sense out of the firehose of data that these things can make. The hard-won benefits are crystal clear to me, but as always, tech is a moving target, so what works today might not work tomorrow...
The nice thing about supercomputers is they serve as a virtual lab for just about any field you can imagine. There are people in the humanities using supercomputers to do interesting things, beyond all the usual astrophysics, chemistry, and geophysical modeling.
Yay supercomputers, and yay NSF.
Re: (Score:2)
Thanks for that - I am somewhat pleasantly surprised. I think the general scientific community that uses these things will be by and large happier - they seem to ramping up the clock speed on the CPUs (good ol' Moore) to get a bunch of their performance, and stuffing more cores on a node, etc... performance that will show itself w/out rewriting code. The accelerators will be there and I will be focusing on figuring out ways to exploit them in our fluid code. I suspect this will really be the last NSF machin
An iPad (Score:2)
and spare batteries.
Precious few technical details (Score:2)
So they got $60 million. What was the proposal, just "give us $60 million and we'll think about how to spend it?". Seems reasonable.
The last one, Stampede2, was Xeons + NVidia. Will this one be Ryzen + Radeon? I expect there are a number of Intel and NVidia salesmen now stalking their prey on campus.
Re: (Score:2)
Thanks. IMHO they should tear up the plans based on current realities and take another look at how to buy the most throughput for the fewest dollars.