Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 5
|
![]() |
Author |
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
After having compared in steady state two identical devices from my system over time, with the only difference being the Memory speed I have the following conclusion:
----------------------------------------With an increase in 25% of DDR3 ram speed (from 1'600 Mhz CL8 to 2'000 Mhz with low latency CL7) I got an increase of 7% in daily number of points produced. 7% is not negligible. Unfortunately the cost of such an increase is prohibitive. You nearly double the memory cost. To do this across all machines does not make any sense. ![]() |
||
|
darth_vader
Veteran Cruncher A galaxy far, far away... Joined: Jul 13, 2005 Post Count: 514 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
The results you saw are in line with many of the benchmarks that are publicly available. The memory speed, once it hits a certain point, does not make that much difference when the processor has fairly large L2 / L3 cache.
Do you know if the two machines had a similar distribution of WUs? Different projects have different memory usage patterns and could be more or less affected by the memory speed. - D |
||
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
The two machines did crunch exactely the same and unique project HCC. No project mix. I think the result is fairly reliable.
----------------------------------------![]() |
||
|
mikaok
Senior Cruncher Finland Joined: Aug 8, 2006 Post Count: 489 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
With an increase in 25% of DDR3 ram speed (from 1'600 Mhz CL8 to 2'000 Mhz with low latency CL7) I got an increase of 7% in daily number of points produced. The speed of the memory module makes difference only when you increase the FSB speed from your mobo. In case your CPU speed stays the same, only the latency of memory counts. You should be comparing the price between two DDR3 1600 MHz memory modules (low-latency and normal). cheers Mika
to infinity and beyond
![]() |
||
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Mika, promised I will do that test but in a distant cold future. I have not enough active machines for the moment.
----------------------------------------![]() ![]() |
||
|
|
![]() |