reef demo as a benchmark!

[TdB]

http://www.mwgl.org/articles.php?refid=EpuZVAlFEEMDgUgFMX

haven´t read it yet.

[GNEP]

They have to be careful - as it is cyclical you should report max/min/avg/st.dev for a "typical" complete cycle only - in excel from the FRAPS dump:

My thread on this with a Parhelia from ages ago

Oh, and it IS cpu dependent on my system.

Another nice picture:

[Jon P. Inghram]

What'd you use to make that chart GNEP?

[GNEP]

Excel. And the data from a fraps logfile, imported into excel, and split up manually (by eye) into each "cycle" of the demo.

But Openoffice.org's calc spreadsheet would work just as well.

The top line was either new drivers, or a new something or other in the machine - I did this a while back and have forgotten.

(OK, I was bored and this I could disguise as my job - which is generally farting around with spreadsheets anyway...)

[Jon P. Inghram]

Cool, I've give it a spin on my 8500 and see what I can come up with.

[GNEP]

Yup. Just have to be careful when doing an average that you take it from whole number multiples of a "cycle" - as you can see from the charts they are very repeatable (after the first one sometimes) and so you only need one really. But taking the average from any part-portion of the "cycle" will give you very different numbers.

Also useful, if charts can't be produced, to produce max, min and standard deviation numbers. Not that I'm anally retentive or anything...