|
[Sponsors] |
December 29, 2015, 16:13 |
Evaluating machine performance
|
#1 |
Senior Member
Join Date: Nov 2010
Location: USA
Posts: 1,232
Rep Power: 24 |
Hello folks
I'm trying to see if it's worth it to add additional machines or just get a larger machine with more CPUs to my small cluster. Right now I have a little hodgepodge of: Machine 1: i7 920 @ 3.7GHz (4 cpus), 24GB of memory Machine 2: 3570K @ 3.4GHz (4 cpus), 12GB of memory Both machines are connected through a gigabit ethernet switch. Basically I'm looking to see how much would be gained over adding additional machines over ethernet. I would like to run these tests: 2 core job, 2 on machine 1 2 core job, 2 on machine 2 2 core job, 1 cpu on each machine 4 core job, 2 cpus on each machine 8 core job, 4 cpus on each machine OpenFOAM outputs the clockTime and executionTime at every timestep, but is there a way to have it report the time used during each part of the iteration? Also how do I evaluate network lag/latency/usage? Do these tests seem reasonable or is there a flaw someplace? |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
[TUTORIAL] Run fluent on distributed memory with 2 windows 7 64 bit machines | ghost82 | FLUENT | 54 | February 9, 2022 04:32 |
Evaluating power in a rotor-stator machine | ghost82 | FLUENT | 6 | November 26, 2014 12:45 |
First CFD Machine: Some questions | chimaera | Hardware | 3 | June 15, 2012 17:39 |
Distributed Parallel on dual core remote machine | Justin | CFX | 1 | February 3, 2008 18:23 |
PC vs. Workstation | Tim Franke | Main CFD Forum | 5 | September 29, 1999 16:01 |