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

CPUs vs GPUs for CFD?

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

Like Tree17Likes

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   May 14, 2024, 01:40
Default
  #21
Member
 
EM
Join Date: Sep 2019
Posts: 59
Rep Power: 7
gnwt4a is on a distinguished road
never compare per-iteration timings in black-box codes. u dont know if the algorithm has been changed, the precision lowered or the arithmetic (usually divs/logs/trigs) done at reduced (but faster) accuracy. always compare timings of what is considered a converged solution at (very) high resolutions - (eg) 5 digits in the mean velocity profile.


Edit: Furthermore, do not use simple geometries (e.g. no plane channels nor straight pipes) which prove nothing and are likely to use specially simplified algorithms, and never uniform grids that converge fast and can be done more efficiently with trig expansions.

Last edited by gnwt4a; May 14, 2024 at 06:18. Reason: stuff added
gnwt4a is offline   Reply With Quote

Old   May 14, 2024, 09:00
Default
  #22
Senior Member
 
Arjun
Join Date: Mar 2009
Location: Nurenberg, Germany
Posts: 1,285
Rep Power: 34
arjun will become famous soon enougharjun will become famous soon enough
Quote:
Originally Posted by CFDfan View Post

I observed the same weird behavior (different answers, or solver divergence when using the GPU) with FEA codes for electromagnetic simulations as well. Obviously not all bugs are cleared and "fast" doesn't necessarily mean "accurate"





Solver can diverge in GPU because the AMG is hugely affected by it. Both the coarsening and the smoothing part.





PS: I have recently developed a smoother that shall not be affected by this issue but still i do not have it in GPU version. At some point of time, will get it. In other solvers, they suffer.
arjun is offline   Reply With Quote

Old   July 23, 2024, 16:59
Default
  #23
Senior Member
 
Tom-Robin Teschner
Join Date: Dec 2011
Location: Cranfield, UK
Posts: 211
Rep Power: 16
t.teschner is on a distinguished road
Just one more point to add, power consumption scales quadratically with the processors' clock frequencies. CPUs have higher clock-frequencies and thus are more power hungry than GPUs (although, as pointed out earlier, an apples-to-apples comparison here is difficult as GPUs and CPUs have vastly different number of cores).

The way that GPUs are designed from a hardware level also makes them quite well suited for parallel processing (in fact, this is what they have been designed to do) and fits many application types in CFD.

There are a ton more reasons why pretty much all general purpose CFD solvers are switching to GPUs these days, I have recently written about that in case it is of interest (where I look at the two points above in more detail).
__________________
cfd.university
t.teschner is offline   Reply With Quote

Reply

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
OpenFOAM benchmarks on various hardware eric Hardware 817 Yesterday 23:07
General recommendations for CFD hardware [WIP] flotus1 Hardware 19 June 23, 2024 19:02
AMD Epyc 9004 "Genoa" buyers guide for CFD flotus1 Hardware 8 January 16, 2023 06:23
CPU for Flow3d mik_urb Hardware 4 December 4, 2022 23:06
Parallel speedup scales better than number of CPUs MikeWorth OpenFOAM Running, Solving & CFD 5 August 21, 2020 18:30


All times are GMT -4. The time now is 12:22.