First, I'm sooo glad I don't have to justify anything to anyone else but
myself. 🙂
Now, I never saw a ROI which covered dual processors - you can do this
yourself by calculating the time you wait for views to become precise - then
compare that to near zero time spent waiting for them because you'll be
working on another part or drawing using the other processor.
Autodesk's very nice benchmarks of video cards is helpful but without adding
tests with multiple windows open at the same time, the bean counters would
buy everyone GeForce 4 cards! In reality, as you may know, there are worlds
of difference between a GeForce 4 and a 900XGL or 700XGL. Your bean
counters are just going to have to learn to have faith in their own
employees.
"GCooper" wrote in message
news:FDE1B53F354A098ED1B43879CB4FF0C1@in.WebX.maYIadrTaRb...
> Sorry to bring this up again, since it has been nearly discussed to death
> here and in the other NGs.
>
> But, Where would I find and "official" Autodesk document that I could
> reference in my cost justification for a new machine with dual processors.
> The guy writing the ROI wants an "official" document that sites dual
> processors as beneficial. I have seen this before but just can't seem to
> find it.
>
> I also need some ammo for justifying RAID0 w/ three drives (rumor has it
> this is the stuff!!!??) and a 900XGL? Does anyone know of a really good
> article, especially with benchmarks that used Inventor for some level of
> testing. Adesk's video card bench marks and certifications really don't
seem
> to guide you toward any sort of a performance workstation or video card. I
> still can't see why anyone would buy a 900xgl if a (550xgl or DCC or GF4)
> performs as well
> (http://support.autodesk.com/inventor-graphic-cards/C_P4_2K_files/7.htm)
and
> I really want a 900XGL!!!!? I need some sort of hard evidence. From
adesk's
> testing we should all be running the GF4 or Quadro 2 Pro or 200 NVS? And
> Cadence and Cadalyst never seem to use a bench mark that includes
Inventor?
>
> And if I use Sean's data an M50 or single processor 340 (thankfully with a
> 900XGL ) would be the way to go? Sean, when is the benchmark going to
> include an *.idw with 10 views of a 15K part assembly, with dimensioning
> while the views are going precise in the background????? Just jerking your
> chain, thanks for giving us something! I really appreciate your efforts!!!
>
> gcooper
>