CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > General Forums > Hardware

Comparing CPUs beyond 2x GHz => 1/2 Time

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   March 14, 2019, 08:09
Default Comparing CPUs beyond 2x GHz => 1/2 Time
  #1
New Member
 
Mike
Join Date: Feb 2018
Posts: 12
Rep Power: 8
Superdude_123 is on a distinguished road
So I've been scratching my head on this. I have one Openfoam case that I ran on 3 PCs and recorded the time it took for each PC to reach the same amount of steady state iterations. The versions of Openfoam, Ubuntu, and all other parameters were the same, the only difference between the PCs is their hardware setup. Here's the setups:

PC #1
CPU = Ryzen 1950x (3.4GHz/16 cores)
64 GB of ram (4x 16GB)
4x 8 TB with 256 MB of cash on RAID0 (via the motherboard)

PC#2
CPU = Intel i5-2400 CPU (3.10GHz/4 cores)
8 GB of ram (2x 4GB)
2x 500 GB + 1 TB USB HD on a software RAID 0

PC#3
CPU = Intel i7-47900 CPU (3.60GHz/4 cores)
16 GB of ram (2x 8 GB)
2x 500 GB on RAID 0 (via the motherboard)

Time wise, I used PC#1 as my baseline, so 100%, yet PC#2 was 41% and PC#3 was 58%.

So what I'm scratching my head with is, if I use the example of #1 vs #3, which have similar GHz/core, why wasn't PC#3 closer to 25% and PC#2 less than 25%?

I'm trying to estimate how much of a performance gain a future system would have, and I would like to compare options between dual Epyc and dual Xeons, however, my half baked time study seems to tell me that comparing GHz/core between Intel CPUs and AMDs isn't like comparing apples to apples.

Any help is appreciated.
Superdude_123 is offline   Reply With Quote

Old   March 14, 2019, 13:33
Default
  #2
Super Moderator
 
flotus1's Avatar
 
Alex
Join Date: Jun 2012
Location: Germany
Posts: 3,406
Rep Power: 48
flotus1 has a spectacular aura aboutflotus1 has a spectacular aura about
Quote:
however, my half baked time study seems to tell me that comparing GHz/core between Intel CPUs and AMDs isn't like comparing apples to apples
True, but only part of the issue here.
The same applies to comparing a 2nd-gen Intel Core CPU (probably with slower RAM) to 4th gen. They are not exactly close.
But the main point is: you are trying to work out CFD performance by using aggregate CPU frequency aka Frequency x number of cores. It is a useless metric for many applications: bad scaling or other bottlenecks like memory bandwidth will throw off any estimate.

There are quite a few performance numbers for different platforms in the pinned thread of this subforum. They will give you a better idea of how a platform could perform.
flotus1 is offline   Reply With Quote

Reply

Tags
amd epyc, cpu time, intel processor, openfoam 4.1

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Elastic or Plastic Deformations in OpenFOAM NickolasPl OpenFOAM Pre-Processing 36 September 23, 2023 08:22
pimpleDyMFoam computation randomly stops babapeti OpenFOAM Running, Solving & CFD 5 January 24, 2018 05:28
pressure in incompressible solvers e.g. simpleFoam chrizzl OpenFOAM Running, Solving & CFD 13 March 28, 2017 05:49
AMI interDyMFoam for mixer nu problem danny123 OpenFOAM Programming & Development 8 September 6, 2013 02:34
mixerVesselAMI2D's mass is not balancing sharonyue OpenFOAM Running, Solving & CFD 6 June 10, 2013 09:34


All times are GMT -4. The time now is 06:15.