From: witr@rwwa.COM (Robert Withrow) Subject: Re: X and math co-processor; A Speed up? Date: 20 Mar 1993 17:02:07 GMT
In article <1993Mar19.200955.23095@cbnewsj.cb.att.com>,
dwex@cbnewsj.cb.att.com (david.e.wexelblat) writes:
| An FP coprocessor may well speed up CLIENT processing, but not the server,
| by any noticable factor.
As a practical rather than theoretical (if that is what this discussion has
been) datapoint, my experience running X11R5 both ways is that having a
FP speeded up SVGA X (as a whole, not just the server, and not any 1 application,
but a non-scientific weighted average of all of the applications I use, which
basically means emacs, xman, xterm, xfig, xrn, and the like) some, but not
by any breathtaking amount. For comparision, the amount of speedup seemed to
be about 0.5 times the speedup I get by running Xfree rather than
stock X11R5. For further comparision, switching from a 33 MHz 386 *with*
fp to a 33MHz 486 seems to give about 1.5 times the X11R5 to Xfree speedup.
I have 386/33 nofp, 386/33 fp, and 486/33 systes here in the office that
I can make direct comparisons with. But I don't do xperf or whatever...
I just use them...
-- Robert Withrow, Tel: +1 617 598 4480, Fax: +1 617 598 4430, Net: witr@rwwa.COM R.W. Withrow Associates, 21 Railroad Ave, Swampscott MA 01907-1821 USA