I did all tests on my old computer which is equipped
with a PIII/1000 MHz, 512 MB SDRAM, Windows 2000 SP2
and an ELSA Gladiac 411, which is basically a NVidia
GeForce 2 MX with 64 MB SDRAM and the 23.11 detonator
driver. I did only the grafic tests and the values
are an average of 5 runs.
Now for the results:
Suite 1: Performance settings as demanded, large assembly
mode as SolidWorks default (treshold 700, remove details,
hide all planes etc, automatic activate: always), shaded
views
Result: 123
Suite 2: like suite 1, but also set remove details during
pan/zoom/rotate on the normal performance options
result: 81 (!)
Suite 3: like suite 1, but activated display HLR edges in
shaded mode (for the nice pictures)
Result: 211
Suite 4: like suite 3, but set large assembly mode to
never activate (so the engine is MUCH slower)
Result: 1029 (!!!!)
I think that's it, if you deactivate large assembly mode
or uncheck "remove details" for this the benchmark results
will grow like crazy. I haven't tested further on, but I'm
sure you can even get some worser results :-))
For the only idea behind a benchmark is to have a test
suite which gives repeatible and comparable results I
recommend that you activate large assembly mode and check
the "remove details" option for that. (like suite 1)
Why? In SolidWorks 2001Plus there are these very nice
settings and without these settings the results are
a) not as your working environment and b) very frustating.
BTW, the results on my result page at
http://solidworks.cad.de/int_swxbench10.htm
are all with the suite 1 settings.
HTH,
Stefan
--
unofficial german SolidWorks helpsite
http://solidworks.cad.de
tools and programs for SolidWorks
http://swtools.cad.de
Thanks Stefan,
Mike Wilson
-----------== Posted via Newsgroups.Com - Uncensored Usenet News ==----------
http://www.newsgroups.com The #1 Newsgroup Service in the World!
-----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =-----
> This sounds like good info for the SolidWorks FAQ.
> It would be nice to have a list of settings (or even
> screenshots) to make the test fair.
I should have thought of that on my own *sigh*
I'm out of business until June, but as soon as
I'm back I will set up a website with the screenshots
with the settings. I really expected the benchmark
itself to set these options as it would recommend
them or at least to have that included in the helppage.
BYe,
Stefan