Message boards :
Number crunching :
quadro cards...
Message board moderation
Author | Message |
---|---|
Dragokatzov Send message Joined: 6 May 00 Posts: 45 Credit: 15,742,981 RAC: 5 |
I am curious as to what kind of crunchers these are. I ask because i see a few Quadro cards that use pci-e 1x on tiger direct. Would be a good way to squeeze another gpu into my computer. I am going to get rid of my ATI card eventually too and replace it with some type of geforce card, but what about running the quadro as well? I just don;t want to dump a few hundred bucks into one and have it only turn out 6 work units a day thanks! Steve http://www.tigerdirect.ca/applications/SearchTools/item-details.asp?EdpNo=3617652&CatId=3599 http://www.tigerdirect.ca/applications/SearchTools/item-details.asp?EdpNo=7210852&CatId=4044 |
Cruncher-American Send message Joined: 25 Mar 02 Posts: 1513 Credit: 370,893,186 RAC: 340 |
Try this Google search for starters: cuda quadro vs geforce |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
I can't speak to their performance, but I can offer a word of advice: forget the 256MB model. With the latest graphics drivers, you need at least 384-512MB of RAM at minimum. |
dan Send message Joined: 18 Oct 02 Posts: 392 Credit: 25,046,383 RAC: 0 |
I can't speak to their performance, but I can offer a word of advice: forget the 256MB model. With the latest graphics drivers, you need at least 384-512MB of RAM at minimum. I'm running a box with 2 Quadro FX570s 256 mb. Currently the box overall has a RAC at ~1800 with opti apps. I don't have the latest driver but they work fine. My next build will use GTX5xx cards though. I would suggest heading that way if possible as it seems to be more bang for the $. Dan |
Helli_retiered Send message Joined: 15 Dec 99 Posts: 707 Credit: 108,785,585 RAC: 0 |
Uh..10,000 sec. for a Workunit - with Cuda. That's on of the longest time i ever seen for a WU done by the GPU. Helli A loooong time ago: First Credits after SETI@home Restart |
dan Send message Joined: 18 Oct 02 Posts: 392 Credit: 25,046,383 RAC: 0 |
Uh..10,000 sec. for a Workunit - with Cuda. That's on of the Still getting science done and that is all that matters. Dan |
Helli_retiered Send message Joined: 15 Dec 99 Posts: 707 Credit: 108,785,585 RAC: 0 |
That's right. Do you know what this rascals need on power? Helli A loooong time ago: First Credits after SETI@home Restart |
dan Send message Joined: 18 Oct 02 Posts: 392 Credit: 25,046,383 RAC: 0 |
I think it is 38W. At least that is what this website says http://www.mv4t.com/Hardware_Graphics_CUDA-Capable-Cards.php My research group bought the box for me for 1500$ in 2008 and that was the best I could get from Dell with the big educational discount we have here at school. Dan |
Fred J. Verster Send message Joined: 21 Apr 04 Posts: 3252 Credit: 31,903,643 RAC: 0 |
Maybe this one, will suit your needs, but doesn't have a PCI-E x1, but x16 , :) Just kiddin, atleast with SETI MB WU's, you don't need a 16x PCI-E slot, I ran, first without noticing, but completion times were a bit longer, a GTX470 and 480 on an ASUS P5E mobo, which runs these cards at 1x and the other at 2x PCI-E. Times where only a few minutes longer, about 15%, with a 1x on a 480, doing 2 at a time and 470 at 2x, also doing 2 at a time, with less then 10% increase, on an 0.4AR WU. Depends on how much data has to be moved to GPU and from GPU to CPU. GPUgrid WU's have more data to move and do require atleast an 8x PCI-E slot. And EINSTEIN@home, also has a better CUDA app. and a lot more traffic from CPU to GPU and GPU to CPU, since their ABP3xxx WU, are split up in 8 or 10 parts, I'm not sure on this one. |
soft^spirit Send message Joined: 18 May 99 Posts: 6497 Credit: 34,134,168 RAC: 0 |
If I am not mistaken, my old GT 9600 is a PCI-E x1. 512MB, still cranking out units. It is approaching retirement age though. Janice |
©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.