XWindows ??

[Alexis Rosen]

In article <5268 at lanl.gov> chn at lanl.gov (Charles Neil) writes:
>In article <11116 at goofy.Apple.COM>, abm at alan.aux.apple.com (Alan Mimms) writes:
>> (Note that our X Window System product is NOT a
>> straight port; very significant enchancements -- especially with regard
>> to performance -- are built into the software, and an excellent set of
>> manuals above and beyond the normal MIT manual pages is included.)
>
>I have to respond to this performance statement.  Here on a MacIIfx with
>16 M memory, I have both Apple's commercial MacX for A/UX 2.0 and MIT's 
>standard X11R4 (with patches 1-19) distribution compiled under gcc
>1.37.91.  In doing large nos. of vector draws on a Tektronix xterm, I
>roughly gauged the MIT server to operate twice as fast as the MacX server.
>So I ran x11perf on each server in one-bit mode to check it out.
>Here are the results of the first few x11perf tests:
> [much substatiation of this claim deleted.]

You're misinterpreting Alan's claim. (Or perhaps he mistyped. But in a number
of discussions, nobody at Apple has ever made this claim, to my knowledge.)

What he and others have been saying is that the _native_ X server which is
included in MacX is faster than the native X from MIT. I haven't tested
this personally but just by eyeballing it I think they're right.

>These figures support the rule my fingers already knew:  don't do
>graphics under MacX; use X11R4 for that.  The beauty is we can have
>both; it takes me maybe 30 sec. to switch from X11R4 to the Macintosh
>desktop.

Which ever native X you use, this is certainly true.

BTW, Apple is including MacX with A/UX 2.0.1.

They have also stated that they'll be sending their changes for the native X
back to MIT for inclusion in R4. No time frame on this yet.

---
Alexis Rosen
Owner/Sysadmin, PANIX Public Access Unix, NY
{cmcl2,apple}!panix!alexis