Message boards :
Number crunching :
Whats the Best Platform to Crunch on?
Message board moderation
Author | Message |
---|---|
rob smith Send message Joined: 7 Mar 03 Posts: 22199 Credit: 416,307,556 RAC: 380 |
Having and AMD farm, I would say Intel I7, a pair (at least) of GTX980ti... That should get the debate going.... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
To answer that question, I'll post my own. How much are you willing to spend and what sort of RAC are you looking for? Depending on your answer will get some of ours |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Having and AMD farm, I would say Intel I7, a pair (at least) of GTX980ti... WUT, considering joining the darkside rob ? "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
I know the PS3's have been pretty much depreciated since the firm ware update. The Android Based Platform runs slow on mid range to low end devices, so with that in mind is the best plat form pretty much the Desktop CPU and GPU? Best in which aspect? -Best total output -Best Performance Per Watt -Best Performance Per Dollar/unit of currency One option not many consider is the MoDT(Mobile on Desktop) setup. A desktop MB that takes a mobile/notebook CPU & gives the ability to add standard desktop GPUs. Here are some examples of the current Socket G3 MBs from ASrock. A similar, & cheaper to MoDT, way to go would be with an Intel Atom based system & adding GPUs. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
spitfire_mk_2 Send message Joined: 14 Apr 00 Posts: 563 Credit: 27,306,885 RAC: 0 |
Having and AMD farm, I would say Intel I7, a pair (at least) of GTX980ti... We have cookies... |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
Hal, you forgot ... which setup gives the most reliable since results. From what I see in my validating of other's inconclusive results ... NVidia cards suck at being reliable Scientifically. But you will bang up your credits. Now that should start a debate! |
rob smith Send message Joined: 7 Mar 03 Posts: 22199 Credit: 416,307,556 RAC: 380 |
Jason - the dark side? I did consider it when plotting last year's upgrade of two of three crunchers. Cost deterred me. I suppose for the ultimate I would be looking for something that would support a dozen (or more) GPU, costing less than a hundred bucks...... The majority of my credit comes from the GPUs, so money spent there is more effective than money spent of CPUs and mother boards, where, in comparison, the gains are "marginal". The biggest problem I find with AMD boards is that finding one that will actually take 3 or more double width GPUs is rather hard... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
Hal, you forgot ... which setup gives the most reliable since results. Measuring reliability is a tricky thing to do. If you simply compare a hosts valid & invalid results the top 20 hosts run ~1-2% invalid. Which isn't to bad on most scales of reliability. Also if they were very unreliable then they would not be producing so many valid results. Granted there are a number of hosts out there producing 100% garbage. Like due to the person running them not ever monitoring the machine. There is the well known issue when a seemingly valid CPU result will get marked invalid against two GPU results. With the two GPU results are obviously trashed. However, That is more of an issue with limits the project, or BOINC, has on validating results. That can also happen with 3 CPU results, but is probably less likely due to the much more mature nature of CPU computing vs GPGPU computing. Error rates and issues are a given for GPGPU computing that we can only accept for the time being. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
I think, given the tens of thousands of WUs I process, my error rate is quite acceptable. All nVidia cards here. If I see something running amock, I tend to it. Usually just a reboot fixes it quickly. Have had one or two rouge cards that needed to be taken out of service, but not many. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.