Robot Structural Analysis Forum
Welcome to Autodesk’s Robot Structural Analysis Forums. Share your knowledge, ask questions, and explore popular Robot Structural Analysis topics.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Improving the speed ofmy computer (hp Z800 workstation)

45 REPLIES 45
Reply
Message 1 of 46
Refaat
2674 Views, 45 Replies

Improving the speed ofmy computer (hp Z800 workstation)

Dear Sir

 

Can you help me to improve the performance of my computer(the speed), where I am doing dynamic analysis for concrete tower using (200 mode) , the finite element size(1.00m) on the version (ARSApro2013), whereas the program takes more than (60 hours) until I can get the analysis results.

Knowing that my compture(hp Z800 workstation) specifications as attached.Although , I've developed  it from (6 GB RAM) to (48 GB RAM) and don’t note any improvement in performance (increased speed).And the graphics card specifications as following :

GV-R795WF3-3GD

AMD Radeon HD 7950/PCI-E 3.0

3072 MB GDDR5/384 bit

Could  you guide me how to increase the efficiency of the computer (increase speed) or what way I can get the results of the analysis faster than (60 hours)?

 

Thanks

Refaat

45 REPLIES 45
Message 2 of 46
Rafal.Gaweda
in reply to: Refaat

1. Send us model

2. Turn off saving quadratic combinations results in job preferences and check.



Rafal Gaweda
Message 3 of 46
Refaat
in reply to: Rafal.Gaweda

Dear Rafal

 

Attached is my model

 

Thanks in advance

 

Refaat

http://www.2shared.com/file/uM-iS2RQ/Refaat_05112012.html

Message 4 of 46
Refaat
in reply to: Rafal.Gaweda

Dear Rafal ? Attached is my model ? Thanks in advance Refaat ? http://www.2shared.com/file/uM-iS2RQ/Refaat_05112012.html
Message 5 of 46
Rafal.Gaweda
in reply to: Refaat

whereas the program takes more than (60 hours) until I can get the analysis results.

 

~5h25min

settings below

 

rcalc.jpg



Rafal Gaweda
Message 6 of 46
Refaat
in reply to: Rafal.Gaweda

Dear Rafal

 

Even , I turn off saving quadratic option as you suggested but still my compture running since  more than 16 h ago.

 

So, what can I do ?

 

Thanks alot

 

Refaat

Message 7 of 46
Rafal.Gaweda
in reply to: Refaat

 

Log:

 

17:48:03 - Structure preparation for analysis #0
18:14:48 - End of preparation for analysis #0
18:14:48 - Start of analysis #0
18:14:48 - It is suggested to use 'DSC algorithm' in case of dynamic and non-linear analysis of structure containing released directions.
The DSC algorithm option can be switched on in the 'Job preferences' dialog box (the 'Structure analysis' tab).
Do you want to continue? #0
18:44:19 - For 8 modes the mass participation factor is equal 70.19
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #0
18:49:56 - For 17 modes the mass participation factor is equal 76.66
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #0
18:56:46 - For 26 modes the mass participation factor is equal 76.70
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #0
19:04:45 - For 29 modes the mass participation factor is equal 86.65
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #0
19:13:57 - For 44 modes the mass participation factor is equal 87.23
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #0
23:38:36 - End of analysis #0

 

My comp:  HP z600

syst.jpg



Rafal Gaweda
Message 8 of 46
Refaat
in reply to: Refaat

Dear Rafal

Do you have any other advice to overcome this problem ?

attached is basic information about my computer.

Thanks

Refaat

 

Message 9 of 46
Rafal.Gaweda
in reply to: Refaat

PLease attache log file 

"c:\Users\username\AppData\Roaming\Autodesk\Autodesk Robot Structural Analysis 2013\Tmp\robwin.log" 

 

What happens after changing solver onto multithreaded ?



Rafal Gaweda
Message 10 of 46
Refaat
in reply to: Rafal.Gaweda

Dear Rafal

 

 

My computer is still runnig , once it finished I will send to u the (Log file ) and also I will try with multithreaded solver .

 

Also , I have noted that your computer has two processors while mine has only one.

Does this affect ?

 

Thanks alot

Refaat

Message 11 of 46
tony.ridley
in reply to: Refaat

If your computer is still too slow we can organise a swap!.  

 

 

Message 12 of 46
Refaat
in reply to: tony.ridley

Hi Tony

 

Thank you for your interest , I agree but how ?

 

Regards

Refaat

Message 13 of 46
Rafal.Gaweda
in reply to: Refaat


 

Also , I have noted that your computer has two processors while mine has only one.

Does this affect ?

 

I used Automatic solver = SparseM,so it does not matter.

Do not you have any "power saver" option on?



Rafal Gaweda
Message 14 of 46
Refaat
in reply to: Rafal.Gaweda

Dear Rafal

 

Sorry , What do you mean by " power saver "?

 

Thanks

Refaat

Message 15 of 46
Rafal.Gaweda
in reply to: Refaat

Something like the option in notebooks : in case of no activity slowing down hard disk or diming screen....


 

Sorry , What do you mean by " power saver "?

 

 

 



Rafal Gaweda
Message 16 of 46
Rafal.Gaweda
in reply to: Rafal.Gaweda

MUltithreaded ~1h

 

14:44:28 - Start of analysis #1
14:44:28 - It is suggested to use 'DSC algorithm' in case of dynamic and non-linear analysis of structure containing released directions.
The DSC algorithm option can be switched on in the 'Job preferences' dialog box (the 'Structure analysis' tab).
Do you want to continue? #1
15:08:43 - For 8 modes the mass participation factor is equal 70.19
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #1
15:11:54 - For 17 modes the mass participation factor is equal 76.66
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #1
15:15:27 - For 26 modes the mass participation factor is equal 76.70
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #1
15:19:20 - For 29 modes the mass participation factor is equal 86.65
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #1
15:23:37 - For 44 modes the mass participation factor is equal 87.23
Do you want to continue the increasing of modes?
[Press ESC to ignore all warnings] #1
15:41:05 - End of analysis #1



Rafal Gaweda
Message 17 of 46
Refaat
in reply to: Rafal.Gaweda

Dear Rafal

 

Good morning

 

First Trial

 

It has been done under the following conditions :

 -  turn off the display (10 min) .

-  Method of solving (Automatic) .

- Save results of quadratic combination (Active)

I got the results after  77h16min , and I got the percent of mass participation more than (90%) after 168 mode .

 

  Second Trial (pls ,sse shot 02) 

 

It has been done under the following conditions:

- Turn off the display (never)

- Method of solving (Multi- threaded )

- Save results of quadratic combination (Not active)

I got the results after 2h28min , and I got the percent of mass participation more than (90%) after 127 mode.

 

 

 

 

My Inquiries are :

 

 

1 . Do you agree with me that the processor affect  the calculation speed as it appears after comparing the time calculation  between us using multi-threaded solver?

 

 

 2. why  does the number of modes decreased in multi-threaded?

 

Thanks alot

 

Refaat

 

 

Message 18 of 46
Rafal.Gaweda
in reply to: Refaat

 

1 . Do you agree with me that the processor affect  the calculation speed as it appears after comparing the time calculation  between us using multi-threaded solver?

 

IN case of multithreded solver- yes - you can compare yours and mine times.

IN auto \ SparseM sover the times should be comparable

 

 2. why  does the number of modes decreased in multi-threaded?

 

I will check

 



Rafal Gaweda
Message 19 of 46
Rafal.Gaweda
in reply to: Rafal.Gaweda

I think you are working on different models.

It is enough to run 50 modes on mine : http://www.4shared.com/zip/r9UILvSV/Strucutral_meshed1.html

 

MOdalsettings.jpg



Rafal Gaweda
Message 20 of 46
Refaat
in reply to: Rafal.Gaweda

 

Dear Rafal

 

Actually , I am working in the same model but the settings are different.

I have noted that when I activated (inactive) in (model analysis parameter ) then I need to use more number of modes than if I activated (percent of mass participation ) to get (90%)

 

Do you agree with me or there is any another opinion

 

I appreciate your good cooperation

Refaat

Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums  

Autodesk Design & Make Report