|
[Sponsors] |
[Workbench] How do you increase the computational speed |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
November 27, 2017, 06:56 |
How do you increase the computational speed
|
#1 |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
Hi everyone;
I wonder there s some way to accelerate the computational speed.The computer work station has worked for one night,while the model process is 1%,It is too slow. Multiple nuclear acceleration is set up. I really appreciate your help! Thank you, zhizhuo |
|
November 27, 2017, 07:26 |
|
#2 |
Senior Member
Gwenael H.
Join Date: Mar 2011
Location: Switzerland
Posts: 392
Rep Power: 20 |
Hi,
It depends on a lot of factors like the kind of hardware you're using, do you output results at all time steps, writing them on a ssd or hhd, are you solving in core or out of core, nb of processors, etc... One thing that you can check is the nb of cores that you request, by default it's 2. You can change that either in Workbench, > tools > options > solution process > Default number of processors, and also under the Mechanical APDL tab > analysis > processors. Or directly in Mechanical > tools > solve process settings > advanced > max number of utilized cores. Have fun |
|
November 27, 2017, 07:27 |
|
#3 |
Senior Member
Gwenael H.
Join Date: Mar 2011
Location: Switzerland
Posts: 392
Rep Power: 20 |
BTW, just be careful where you post the thread, this should not be in the meshing forum
|
|
November 27, 2017, 07:43 |
|
#4 |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
Thank you very much for your suggestion and sorry for the mistake.you are so kind. I will pay attention to it next time.What you said setting nb of cores,I have set six .Maybe there are some other problem.
|
|
November 27, 2017, 08:07 |
|
#5 |
Senior Member
Gwenael H.
Join Date: Mar 2011
Location: Switzerland
Posts: 392
Rep Power: 20 |
You should also look at the solver file output > Solution branch > solution information > solver output. You'll find a lot of useful information there.
|
|
November 27, 2017, 08:12 |
|
#6 | |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
Quote:
|
||
November 27, 2017, 15:22 |
|
#7 |
New Member
LK_
Join Date: Oct 2017
Posts: 16
Rep Power: 8 |
Go through Explicit tutorial course from the Customer portal.
Your timestep is very small. You can increase your timestep by for example using mass scaling. |
|
November 27, 2017, 18:10 |
|
#8 |
Senior Member
Svetlana Tkachenko
Join Date: Oct 2013
Location: Australia, Sydney
Posts: 407
Rep Power: 14 |
Another possibility is using a more coarse mesh to get preliminary results. Once that works start refining the mesh. It is difficult to advise on this without knowing your mesh size or your geometry though. Mesh convergence needs to be established starting from the coarser mesh not the other way around.
Could also check whether your employer or university provides you with access to a supercomputer. |
|
November 27, 2017, 22:28 |
|
#9 |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
I donnot know how to increase the time step.could you tell me more details about the mass scaling.I am very appreciate your help.
|
|
November 27, 2017, 22:36 |
|
#10 | |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
Quote:
|
||
November 28, 2017, 02:22 |
|
#11 |
Senior Member
Gwenael H.
Join Date: Mar 2011
Location: Switzerland
Posts: 392
Rep Power: 20 |
As LK suggested I would also recommend that you go through the Explicit tutorials that you can find on the customer portal.
You can also have a look at Ansys Meshing classes to help you getting started |
|
November 28, 2017, 02:53 |
|
#12 |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
||
November 28, 2017, 05:11 |
|
#13 |
Senior Member
Stuart
Join Date: Jul 2009
Location: Portsmouth, England
Posts: 733
Rep Power: 25 |
As mentioned above this is an explicit dynamics issue and not a mesh generation issue, but it is caused due to the smallest mesh size, see Progressive Collapse of the structure with Ansys Autodyn.
Run your model for a few timesteps , e.g. 20 and locate the smallest timestep using a User Result, then you can proceed in fixing this. Apply mass scaling (you may be part and element scaling as well) or remake the mesh avoiding the small element size. |
|
November 28, 2017, 08:51 |
|
#14 | |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
Quote:
|
||
November 28, 2017, 08:53 |
|
#15 | |
New Member
Join Date: Nov 2017
Posts: 19
Rep Power: 8 |
Quote:
|
||
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Increase speed of parallel computation | Purushothama | Siemens | 2 | November 30, 2010 14:51 |
Simulation speed increase | Attesz | CFX | 24 | October 13, 2010 09:55 |
Does it metter to increase RAM for solver speed?? | raima | Main CFD Forum | 1 | February 28, 2008 10:47 |
free "Computational Mechanics" | noname | Main CFD Forum | 0 | March 2, 2007 07:03 |
How to Increase Speed? | steve podleski | Main CFD Forum | 23 | January 11, 2000 15:41 |