Punch Holder Benchmark File
In your SolidWorks System Options under Performance, check to be sure Verification On Rebuild is Turned Off. (No check mark in box)
Open the punch holder file and do a Control-Q Rebuild. Post your results for Total Rebuild Time shown in SolidWorks under Tools > Feature Statistics after the Control-Q Rebuild of the model.
Download the files from the following link:
http://visualvault.auerprecision.com/ViewFile.aspx?DlID=9ace8149-f8d4-4f7f-8a1f-0f78bb0469b0
Files are available for SolidWorks 2007, 2008, 2009 and 2010.
To submit your rebuild results use the form found at this link:
http://visualvault.auerprecision.com/formviewer.aspx?FormID=9e042945-d89e-42b8-8c49-3b9aa2680603
The results are being posted in a Google Spreadsheet at:
http://spreadsheets.google.com/pub?key=pz7wTpIkC7LA28ybEyxyTPw
Ed,
What are your computers specs?
Sometimes the Feature Statistics does not refresh properly. You may have seen this on the x32 bit install and where getting the time from when I last rebuilt the file when setting up the benchmark files for upload.
Cheers,
Anna
Posted by: Anna Wood | December 28, 2009 at 03:27 PM
interestingly, I have just run the benchmark on the punch holder on my windows 7 (32bit system) and it performed well at 57seconds. I have now just installed (on a seperate partition) windows 7 x64. Ran the same test on the same version and it came out a 90seconds?!? The system has 8Gb RAM installed - which is obviously recognised on the x64 side - but the 32bot side performs better? Any ideas?
Posted by: Ed | December 22, 2009 at 07:36 AM
I tested a Dell Latitude to see how it compared with the others in the benchmark list. While the results are not all that great, they surprised me. I assumed the laptop would do horribly since the clock speed is only 1.8GHz, but I suppose the dual cores helped.
Total rebuild time in seconds: 223.15
SW Version: sw2009sp0
Computer: Dell Latitude D620 (Laptop), not modified in any way.
System Cost: 540$
OS: Windows XP Pro, 32bit, SP2
Number of Cores: 2
Clock Speed: 1.83GHz
RAM: 2GB
Posted by: Tobias | October 15, 2009 at 02:06 PM
Wayne,
Yes, I know that. :-) There are several other sins in the model. This model represents a worse case scenario on possible modeling techniques that can affect rebuild time.
I have not corrected the model as that would skew the intent of the model to test rebuild time on various systems with a consistent dataset.
The model was the subject of a SWW breakout session by Phil Sluder and Ed Eaton a couple years ago. They were able to get the rebuild time under about 20 seconds with different modeling techniques.
Cheers,
Anna
Posted by: Anna Wood | August 26, 2009 at 07:55 AM
Anna, I looked at the Punch Holder model and noticed the linear pattern that needed the most rebuild time did not have geometry pattern turned on. This option, when usable, greatly reduces the rebuild time. As an example, my system took 94 seconds with this turned off. With this turned on, it only took 47 seconds.
Posted by: Wayne | August 26, 2009 at 06:10 AM