Slashdot Mirror
Cray Unveils Its First GPU Supercomputer
An anonymous reader writes "Supercomputer giant Cray has lifted the lid on its first GPU offering, bringing it into the realm of top supers like the Chinese Tianhe-1A"
The machine consists of racks of blades, each with eight GPU and CPU pairs (that can even be installed into older machines). It looks like Cray delayed the release of hardware using GPUs to work on a higher level programming environment than is available from other vendors.
...will promptly be used for mining BitCoins.
A Beowulf cluster of these!
...can it run Metro 2033 on High?
My single GTX580 can.
"I zero-index my hamsters" - Willtor (147206)
The fact that writing C and Fortran code using a message passing library constitutes a high level programming environment is a complete indictment of the sad state of parallel programming today. Seriously, do you want to be programming complex parallel algorithms on HPC machines using Soviet Era technology? I've tried that and it made me want to jump out a window. It's about as easy to program in this type of an environment as it is to program an FPGA (hint: it's a pain in the ass).
There is still a lot of HPC applications written in Fortran with this run them?
Also how hard if any of a porting will be needed to get good results from this.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
You would be right except that their are applications that do require the performance.
You can never have too much computing power for some applications like climate modeling. So what is your point?
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
Uh, no you couldn't. The rate of bitcoin creation is fixed (it's about 50 BTCs / 10 mins, for now). If you add more computational time the system will adjust and it'd become proportionally harder to generate them, so the global rate would keep stable.
So despite the 100 thousand-fold increase in mining difficulty in the past 15 months, the network continuously self-adjusts itself to issue one block of Bitcoins about every 10 minutes. The difficulty increase is entirely caused by users competing between themselves to acquire these blocks.
http://blog.zorinaq.com/?e=49
Dilbert RSS feed
bringing it into the realm of top supers like the Chinese Tianhe-1A
Uh, Cray already has machines in service that blow Tianhe-1A out of the water on real science. Tianhe-1A doesn't even exist anymore. It was a publicity stunt. Cray is already making the top supers. It's others that have to catch up.
Really, you've tried it and it made you want to jump out of a window? OpenMP is an extremely simple, easy to use add-on to the C language. It is one of the two current standards used for parallelized scientific computing, and although it will eventually be succeeded by a language with more features, it will be difficult for its successor to match its ease and workmanlike grace.
I honestly have trouble believing someone could have much difficulty with it. If you want to have the work in a "for" loop parallelized the extremely mentally challenging thing to do is write
#pragma omp for
just before your for loop. Look at all that difficult message passing you have to contend with! And since you're writing scientific calculations, C generally lends itself to good clarity in the code. I worked with a math major who had only done the smallest bits of C and Java beforehand, and he picked up OpenMP immediately. If you can't handle it, you really should reconsider whether your talents lie in software development. If you want to see what a truly awful parallelizing language or API is, look at UPC - unified parallel C.
The point is not for the job to be easy for your lazy ass, the point is for the code to execute as quickly as possible.
Physics simulations involving discretized partial differential equations can make any machine less powerful than a Matroshka Brain cry uncle. If some of my optimizations work out, I'll be able to get near-slideshow framerates on a 512x256 2D simulation of a single, ideal, purely hydrodynamic fluid using nVidia's top of the line C2050 GPU.
:(
Now consider that I'd prefer to have at least a thousand cells per side in all 3 dimensions, which makes the problem ten thousand times larger, preferably several thousand which would make it nearly a million times larger. Then add magnetism, resistance, viscosity and ExB drift, and move to at least a two or three fluid model, which would make for about 30 times the work per cell over what I've currently got running.
So before even beginning to consider atomic physics, radiation and non-ideal equations of state, I've made the problem about 20-30 million times more difficult than one which utilizes a GPU to the limit. Even if everything scaled perfectly across a thousand GPUs, major problems would take days or weeks to run