![]() |
|
|
#1 |
|
Mar 2010
1100112 Posts |
Has there been any research put into time that different CUDA gpu's take for polynomial selection? Like 8800 GT vs 465 GT? Does the number of cores affect the timing significantly?
Is there any place I can read into this topic that you all know of? Thanks -- |
|
|
|
|
|
#2 |
|
Dec 2010
Monticello
5×359 Posts |
http://mersenneforum.org/forumdisplay.php?f=92
may interest you for general comparisons of GPUs..... |
|
|
|
|
|
#3 |
|
Mar 2010
19B16 Posts |
Basically, the more cores(excluding gpus of sm_21 architecture) at higher clock(s), the faster the polyselection.
|
|
|
|
|
|
#4 |
|
Tribal Bullet
Oct 2004
1101110101012 Posts |
Recent work on the poly selection code has reduced the GPU side of things to a curiosity and not a critical advance; the GPU does run faster on smaller problems but smaller problems (< 130-140 digits) don't need that much help selecting polynomials.
|
|
|
|
![]() |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Comparing Intel and AMD | JHagerson | Hardware | 14 | 2005-08-21 16:41 |