Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!apple!rutgers!mit-eddie!apollo!rehrauer
From: rehrauer@apollo.hp.com (Steve Rehrauer)
Newsgroups: comp.sys.apollo
Subject: Re: postscript preview
Message-ID: <1991Jun3.133702.7508@apollo.hp.com>
Date: 3 Jun 91 13:37:02 GMT
References: <1991May31.155935.6043@cs.cmu.edu> <51e5d5ff.1bc5b@pisa.citi.umich.edu>
Sender: netnews@apollo.hp.com (UUCP For Sam Murphy/HPnews)
Organization: Hewlett-Packard Company, Apollo Division - Chelmsford, MA
Lines: 19
Nntp-Posting-Host: copper.ch.apollo.hp.com

In article <51e5d5ff.1bc5b@pisa.citi.umich.edu> rees@citi.umich.edu (Jim Rees) writes:
>Funny you should ask.  I just finished compiling ghostscript on my Apollo,
>after getting disgusted with xps.  I made no changes to the source, only
>configuration changes to the makefile.  I suggest using:
>
[ ... ]
>-A cpu,3000 for the usual reasons (I wish they'd make this the default, or
>at least let me change the default on a per-node basis).

We did make it the default for the *.8 compilers.

(Okay, I lied.  We made "mathlib_sr10" the default.  It's almost the same
as "3000"; you get inline floating-point code for anything that an '040
supports directly.  E.g.: you get lib-calls for some of the transcendentals.
Still, it's (default performance) far, far better than "-A cpu,any".  Note
that if you're using a FPA board, or have a PEB node or some such early
hardware, then you should override the default -cpu.  Refer to the release
notes (you know: those papers no one reads :-) on the *.8 compilers for full
details.)