[ALLY] Re: View Speed Test
From: larry:ecotope at: 8-May-2001 9:51
Hi PeO
> Wasn't the speed test's initial goal to compare the REBOL/View speed
between
> different platforms and graphic adapters ? So, removing the extra info
will
> give a very unfair comparision between the CPUs..
I think Carl's original intent was simply to get a quick indicator of the
speed of View in each of the three categories of the test. Like many
benchmarks, the test is far too simple and short to really evaluate CPUs,
but the results are interesting none the less.
The email pop-up will take any line of text so folks can also enter cpu
make, video card, video memory, etc., if they wish. I meant to say only that
I would discard any results lacking cpu speed.
There are several important variables not generally present in the data
received so far, for instance screen display resolution and color depth,
system bus speed, ram chip speeds, cpu cache specs, video memory and display
speed, etc. While these factors can be useful in understanding particularly
high or low times for a given cpu speed, they introduce so many variables
that it would require hundreds of test results overall in order to be able
to at least partially quantify the influence of each factor.
> Would also be interesting if everyone could add a RC5 benchmark to their
> results, since this test seems to give very small differences between OSes
> on the same system.
>
I am not familiar with RC5. I would guess that scripts which don't use
view/vid (like my loop-timer.r script) would run at comparable speeds with
different OS on same box.
-Larry