[Beowulf] El Reg: AMD reveals potent parallel processing breakthrough
Lux, Jim (337C)
james.p.lux at jpl.nasa.gov
Fri May 10 17:17:23 PDT 2013
I agree with you..
But proving that you can build a HPC using whatever, be it PS/3s or Furbys or nVidia cards, if you can't find people to program it, that's a problem.
(As someone who proposed building a cluster with Arduinos, I recognize that I am eating my own dogfood now... But I DO happen to have 5 Arduino Uno Ethernets here in my desk drawer, and I have a 5 port hub/switch as well. The switch claims 1 Gbps, so I probably won't have to worry about saturating its bisection bandwidth. Or constructing some sort of toroidal fabric. Funny, I don't see an implementation of MPIduino anywhere though.)
From: Prentice Bisbal [mailto:prentice at rdi2.rutgers.edu] On Behalf Of Prentice Bisbal
Sent: Friday, May 10, 2013 7:25 AM
To: Lux, Jim (337C)
Cc: beowulf at beowulf.org
Subject: Re: [Beowulf] El Reg: AMD reveals potent parallel processing breakthrough
On 05/10/2013 12:04 AM, Lux, Jim (337C) wrote:
> On 5/8/13 6:41 PM, "Prentice Bisbal" <prentice.bisbal at rutgers.edu> wrote:
>> On 05/08/2013 09:41 AM, Lux, Jim (337C) wrote:
>>> The game console business is a strange one, and I don't know that it
>>> has much to bring to the HPC world (whoa, that will provoke some comment).
>> Roadrunner's body isn't even cold yet, and everyone's already
>> forgotten about it. :(
> I think roadrunner is an example of a one-off stunt..
> In the long run, "easy programming" is probably a bigger cost driver.
One man's "stunt" is another man's "experiment" or "proof-of-concept". I think that Roadrunner succeeded as a proof-of-concept for using accelerators. If it failed because of programming difficulty, that's a software/people issue. While still a significant problem that can't be ignored, I was thinking only about hardware, and your statement. The success of Roadrunner in breaking the petaflop barrier can't be ignored, and proves that the gaming console world has already brought something to the HPC world.
Now it's success as a USABLE computer, is a different story, and a valid point.
Roadrunner isn't the only example of game console technology in HPC.
Google for "PS3 cluster", and you'll find a multiple cases of people using PS3s for HPC clusters. Yes, most were small experiments, but it can be argued that the reason PS3 clusters weren't widely adopted was a people issue more than a technical hardware capabilities problem - I'm sure there wasn't a large ecosystem of software available for the PS3, and while it wasn't hard to create a proof-of-concept cluster, I imagine the resources it would take to port all the libraries/utilities to the
PS3 hardware scientist are used to having on an x86 system probably prohibited practical use and wide-spread adoption.
More information about the Beowulf