Revit general purpose performance benchmark

[Wes Macaulay]

Steven Campbell's rendering benchmark results have been revolutionary, but one thing I think we'd all benefit from is a benchmark testing day to day tasks in Revit: opening views, snapping, joining geometry, deleting objects, adding windows, etc.

People ask me all the time what a good Revit box is: what kind of RAM should I use? What sort of video card? How does processor speed help? Or bus speed on the mainboard?

A benchmark like this might be more illuminating than the rendering benchmark (oh wow - that pun was not intended), and I wonder if the fastest rendering box might or might not be the fastest at the modelling and drafting tasks we do every day in Revit.

[Scott Hopkins]

Wes - agreed. It didn't quite make sense to me that the Revit Benchmark only tested ray tracing and radiosity. While it is important, rendering may be less then one percent of my Revit work. I think you are right on the money. A really meaty Revit file with a complex journal of tasks would be an excellent benchmark study. At the very least it would wipe the smiles off the faces of all those hotshots with dual processors.

[D_Driver]

has anyone followed up on this, posting a journal file for testing?

[Roger Evans]

No matter how fast the machine is I'm sure my brain & skills would slow it down

[Steven Campbell]

Steven Campbell's rendering benchmark results have been revolutionary, but one thing I think we'd all benefit from is a benchmark testing day to day tasks in Revit: opening views, snapping, joining geometry, deleting objects, adding windows, etc.

People ask me all the time what a good Revit box is: what kind of RAM should I use? What sort of video card? How does processor speed help? Or bus speed on the mainboard?

A benchmark like this might be more illuminating than the rendering benchmark (oh wow - that pun was not intended), and I wonder if the fastest rendering box might or might not be the fastest at the modelling and drafting tasks we do every day in Revit.
Wes,
If you want to put together a runnable journal, I will turn it into a benchmark test when I get some free time at home. If you start with the Hotel sample and hit every area you think should be included, I will add the timers and output info to generate it.

Steve

[Wes Macaulay]

Wes,
If you want to put together a runnable journal, I will turn it into a benchmark test when I get some free time at home. If you start with the Hotel sample and hit every area you think should be included, I will add the timers and output info to generate it.

SteveI most certainly will, Steven! I appreciate your interest in this... and I'm sure everyone else will, too...

[cadpoobah]

Wes,
If you want to put together a runnable journal, I will turn it into a benchmark test when I get some free time at home. If you start with the Hotel sample and hit every area you think should be included, I will add the timers and output info to generate it.

I'm obviously late to the party :), but just curious if there has been any progress on this?

[Steven Campbell]

I think Wes got a bit busy... I am still open to it when we can get a good journal.

[Wes Macaulay]

Hey - good point - forgot about this. I'll put this back on the burner again, folks!

[dhurtubise]

Guys need help to put that together ?

[Wes Macaulay]

We do! If you've got something in the can and ready to go, let's get 'er rolling

[dhurtubise]

You mean a model ?

[DaveP]

I tried to put something together when we were doing a lot of network changes & found the Journal File route to be extremely - I don't know it fussy is the right word - precise, maybe. If ANYTHING is different, the Journal file would blow up. If it needs any user prompts, or if any object is not where the Journal expects it, -boom!

Unfortunately, this kind of precludes using a Journal benchmark on a Worksetted file, which is, of course, the thing you want most to test. I suppose you could manually make a Local copy, Open it & re-Save it, but I seemed to always get something wrong in the process. If you could just suppress the "This used to be a Central File...." blather, that would help a ton, but I couldn't figure out a good way around it.

Still, a Benchmark even without a Save To Central could test a lot of other stuff & be a great way to compare machines.

[Wes Macaulay]

I've got some monster files that are not worksetted that we could try... I could hand the file over to someone who would like to make a test out of it?