W
HITE
P
APER
(cont.)
30
3.
To measure workstation performance, run the benchmark locally. That is with local parts, and
Pro/E installed locally. Remove the network connection if possible. This will remove the
possibility of network data transfer from causing performance variability.
4.
Make sure you have t he latest BIOS and drivers before starting.
5.
Be consistent when comparing different workstations:
§
same memory amount
§
same screen settings (resolution, frequency, vsync on/off)
§
same Pro/Engineer version and datecode
Web Sites of PTC benchmarks
The following web site is from the Pro/USER community and represents a variety of user activities:
http://utopia.knoware.nl/~ocorten/
What kind of actions does it test?
§
Wireframe redraws
§
Shaded mouse spins
§
Shaded redraws
§
Shade calculations
§
Regenerations
§
Menu redraws
§
Saves
§
Dialogue box redraws
§
Mouse zooms
§
Patterned X-section creation
§
Family Table Verify
§
Assembly creation
§
Explode translations
§
Screen updates
§
Model Tree expansions
§
Automatic Regenerations
§
Perspective views
The SPEC APC for Pro/E 2000i website, which uses a very large model, is a benchmark to compare
graphics cards:
http://www.spec.org/gpc/