China successfully developed "Milky Way One" Supercomputer

[Mungri]

But can it run crysis?

Why dont they run 3D mark on that beast and hold the No.1 spot forever and ever?

I want benchmarks dammit!

And it had better be folding.

And can I go use it to play a game of overclock please? I'll try my hardest not to break it, promise.

[gumballguy]

But can it run crysis?

WTF is with the resurrection of that comment?

It's become stupidly popular again.

TIME. TO. DIE.

Not funny. Give it up. I'm looking at you *glares at all*

[lkiller123]

why the hell would they use radeons? i think its fake. no one would deploy a massive scale of gpu's and not use the professional series. their is just too much risk.

Well, they are more than 50% cheaper and calculates the same (in terms of flops). Why not?

[Origin_Unknown]

the real questions are;

1) can it run WCG? and why isnt it?
2) is movieman's personal server fam more powerful? i think so!

[Power5]

Would be sweet to see a 3d06 or vantage score, but I would guess it is not possible to run on this thing without some serious rework to the program code.

Crysis saying may be dead to you, but I still think that exact same thing every time I hear about some new super tech.

[Chumbucket843]

1.000.000.000.000.000 so in other words 1 peta flop... impressive...

so while nvidia talks gpgpu, ati makes it happen?
im curious though, how are they going to use the gpus?
open cl?

the gpu's are ~3 petaflops peak, probably getting ~450 teraflops running linpack b/c double precision. this would be implemented through ATi stream or whatever they are calling it now. nvidia tesla is in supercomputers too. they actually designed server racks with 4 gpu's. here is a video of tsubame.
http://www.computerworld.com.au/video/740796303

[revenant]

soooo... I guess they're going to get right to the task of generating the largest prime number... weeee!

[lkiller123]

Well, actually I think they should've used Nvidia GPUs instead of Radeons for the GPGPU purpose. I've heard that the CUDA technology had been way more mature than ATo Stream.

Was it probably because ATi sponsered this project?

[sniper_sung]

i'm questioning if this guy is really qualified to be a professor. my netbook can do this as long as i have enough external storage drives, you don't need a super computer. he doesn't even know the difference between processing power and storage space

I think you probably didn't calculate the storage space required for these photos. Suppose each photo occupies 1 MB, it takes 1.5 PB to store all these. Typically it takes a classroom to hold hard drives adding up to 1.5 PB today.

Secondary-storage space is another important performance index of a supercomputer, further to computation power. For example, to simulate ion diffusion in a single human heart, for 100 mm spacing we have 4.23x10^8 computational points. Storing 1 minute of simulation at 1 ms time steps requires 5.4 PB disk space.

I think it's pointless to mention your netbook here, as it's not relevant.

[cegras]

They might be giving numbers so people can put things into perspective, although not the most relevant ones.

[LordEC911]

Well, actually I think they should've used Nvidia GPUs instead of Radeons for the GPGPU purpose. I've heard that the CUDA technology had been way more mature than ATo Stream.

Was it probably because ATi sponsered this project?

Because if you actually code for AMD/ATi's architecture their cards are MUCH faster than Nvidia cards.

[MirageSys]

This is suspect.

"Milky-way". You know Chinese milk is tainted with Melamine.....
/sarcasm


Last I checked, ATI architecture lays the beat down on current NVIDIA architecture when it comes to DP calculations. Why are people so surprised its being used?