![]() |
![]() |
#1 |
Oct 2016
5 Posts |
![]()
Quick observation. I am running Prime95 v28.9 build 2, two instances of Prime95- one for each of two Xeon E-5-2960 v4 CPUs, each with 32 Gb of quad channel RAM (SuperMicro motherboard). Each CPU has 4 clients running. There are currently no other applications running except an antivirus program.
Before Windows 10 reboot: -CPU0 workers are running at about 24 ms/iteration for a 78M exponent -CPU1 workers are running at about 13 ms/iteration for a 78M exponent After reboot due to Antivirus application update, all 8 workers were running at about 50 ms/iteration. I stopped and started the workers without another reboot, but there was no change in speed (this has worked for me in the past). I then rebooted without success. I rebooted again, and now the workers are again running at about 24 and 13 ms/iteration. Anyone know why this might be happening? It is quite mysterious why such differences in speed exist over time. |
![]() |
![]() |
![]() |
#2 |
P90 years forever!
Aug 2002
Yeehaw, FL
2×4,079 Posts |
![]()
I assume each worker is using the same number of cores?
How do you make sure each prime95 instance binds it's workers to a specific CPU? What problems did you have running one prime95 instance with 8 workers? |
![]() |
![]() |
![]() |
#3 |
Jun 2003
543610 Posts |
![]()
I don't see any E5-2960 v4 chips at ark.intel.com. The one I see is http://ark.intel.com/products/91770/...Cache-2_60-GHz. Assuming you made a typo, and it really is 2690, this is a 14 core CPU with 35MB L3 cache. Your optimal configuration will be to run 1 worker per CPU with 14 threads. And this can be easily achieved with one instance of P95 (instead of two).
|
![]() |
![]() |
![]() |
#4 |
Nov 2016
2·3 Posts |
![]()
What is the optimal setting for a I7-6700k for one worker? I have all 4 cores assigned and I can still run other programs easily.
|
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Speeding up (n mod m) for very large fixed n and varying smaller m | mickfrancis | Computer Science & Computational Number Theory | 9 | 2016-03-29 19:00 |
Hard reboot | dans | Hardware | 5 | 2009-12-30 08:22 |
Benchmarks varying FSB, Memory Latency and multiplier | S485122 | Software | 0 | 2006-11-08 20:21 |
Compaq ML370 G3 server 2x+2,4GHz reboot | piotrkonarski | Hardware | 11 | 2004-03-03 12:16 |
Reboot and Power Down | jwbento | Hardware | 5 | 2003-12-04 00:30 |