Benchmark test - World car

[avillanueva]

In reading the test documentation, it appears like I have to set a bunch
of IBAs in my production system and add them to the Cad Doc type. Does
this mean when a user creates a cad doc in the future, they will be
presented with 10 bogus attributes that are only used for testing? I
understand that it affects performance but is it really necessary to add
these in? Appendix B shows how you can map them to existing IBAs but I
just assume not clutter my system.

Am I correct in say that attribute mapping has nothing to do with the
attributes added to the template?

[MikeLockwood]

Totally agree. It's of great value to have this data and to run the test in the production system - but nobody in their right mind will add garbage IBA's to production just to be able to run the benchmark.

We need to have PTC update the test to exclude IBA's applied to CAD Doc's (until / unless there would be a way to isolate their use to the test data), maybe thru a second organization or something). Don't know how much the test results are skewed by not having the attributes - would be interesting to see what the difference is.

[avillanueva]

Well, they do have a section concerning the additional IBAs. If you use
the mapping technique in Appendix B, they do not need to be added to the
type. So long as you have 3 Strings, 3 Booleans, 3 Integers and 1 Float
already as attributes, it should be no problem. If you are using the
auto-create WTPart and have it dropping extensions for the number, you
will error on check in since there are a number of items that have the
same filename with different extension.

[avillanueva]

Another note. The only task user #2 does is rename the assembly. That
can be done via the browser. I am not sure what doing it via Wildfire
gains us. It sure would make testing easier since all you need to do is
open Firefox or another IE session and log in as a different user. I
wish this was an automated test. I plan on reworking this script a bit
to see if I can make an XML script based on the Single user testing doc.
I just don't know if that eliminates display time to the equation (time
IE takes to render results). At least it would be simpler to run.

[rjessen]

Just thought I would mention that by using linkTuner from Fishbowl Solutions you can automate performance testing of the World Car.

I've attached a bitmap showing the results I measured for the World Car on Intralink 3.4 vs. PDMLink on the same server (an underpowered monolithic server running on VMWare ;>)

Here's a link to the website if you're interested in finding out more:

http://www.fishbowlsolutions.com/PTCSolutions/FishbowlPTCAutomationandIntegrationSolutions/SolutionsForPDMLink/linkTuner/fs-mkt-linktuner_webcopy#Product%20Specs?WT.mc_id=L_MCAD_Software_raj