|
[Sponsors] |
June 28, 2017, 13:45 |
Computing time/effort for model validation
|
#1 |
New Member
Jacob Mannhardt
Join Date: Jun 2017
Posts: 2
Rep Power: 0 |
Hey folks,
I am comparing different modelling approaches of heat exchangers for my thesis and want to plot accuracy against computing time to find the most efficient model approach. But I am using different computers with different computing powers and thus a "time-based" computing effort might lead to wrong or computer-dependent results. Is there a more suitable and computer-independent way to measure the effort ion CFX? Cheers Jacob |
|
June 28, 2017, 19:41 |
|
#2 |
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,854
Rep Power: 144 |
Baseline all the machine first with a small benchmark simulation so you know the relative speeds of the machine. Then you can compare across machines accounting for speed variations. In /examples there is a file Benchmark.def which is ideal for this.
Noe this benchmark is only accurate providing your machines are CPU limited. If you have memory or hard disk speed limitations this method will not work well. |
|
June 30, 2017, 09:54 |
|
#3 |
Senior Member
Erik
Join Date: Feb 2011
Location: Earth (Land portion)
Posts: 1,188
Rep Power: 23 |
Good idea Glenn,
Also I would think all you should care about is relative time to each other, not absolute computation time? |
|
Tags |
accuracy, cfx, computing time, efficiency, effort |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
CFX problem in ubuntu (linux) | Vigneshramaero | CFX | 0 | July 13, 2012 11:22 |
CFX-Pre problem, pls help!!! | cth_yao | CFX | 0 | February 17, 2012 01:52 |
CFD algorithm design with reconfigurable computing | Beatríz Navarro | CFX | 4 | June 27, 2006 05:39 |
CFD algorithm design with reconfigurable computing | Beatríz Navarro | Fidelity CFD | 0 | June 24, 2006 19:23 |
CFD algorithm design with reconfigurable computing | Beatríz Navarro | FLUENT | 0 | June 24, 2006 19:22 |