Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!decvax!decwrl!ucbvax!GVAX.CS.CORNELL.EDU!jqj
From: jqj@GVAX.CS.CORNELL.EDU.UUCP
Newsgroups: mod.computers.workstations
Subject: Vaxstation/GPX performance
Message-ID: <8605111017.AA03005@gvax.cs.cornell.edu>
Date: Sun, 11-May-86 06:17:38 EDT
Article-I.D.: gvax.8605111017.AA03005
Posted: Sun May 11 06:17:38 1986
Date-Received: Tue, 13-May-86 18:13:52 EDT
References: <8605021645.AA09943@CHARON>,
Sender: daemon@ucbvax.BERKELEY.EDU
Reply-To: gvax!jqj (J Q Johnson)
Organization: Cornell Univ. CS Dept, Ithaca NY
Lines: 18
Approved: works@red.rutgers.edu

In article <8605051509.AA09244@CHARON> meltsner@ATHENA.MIT.EDU
writes: 
>...  I've seen X on the QDSS (II/GPX display) and it is fast!
>Numbers I got at a DEC road show indicate the GPX can be 2x the
>speed of a Sun 3, even under X.

DEC has made the same claim to us.  The traditional (originally SUN)
bitmap-access demos do seem to run very fast on the GPX.  Does anyone
have concrete performance data comparing the GPX with the SUN-3/160C
for graphics (especially color) operations?  I would be particularly
interested in data including software overhead at higher levels than
raw bitmap access, e.g. GKS polygon fill.

It is clear that the GPX will win for 2-d bitmap rotation (SUN
doesn't support it in hardware unless you buy their coprocessor).  I
woulg guess that the SUN would win big for cpu-intensive graphics
algorithms; does anyone have a hidden-surface-elimination benchmark
or ray casting or something?