Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!bloom-beacon!mit-eddie!ll-xn!ames!elroy!peregrine!ccicpg!cci632!ritcv!rochester!ken
From: ken@cs.rochester.edu (Ken Yap(probably))
Newsgroups: comp.text
Subject: Re: tex82 or ctex2.9?
Message-ID: <1988Jul7.001521.16184@cs.rochester.edu>
Date: 7 Jul 88 04:15:21 GMT
References: <17268@gatech.edu> <11071@sol.ARPA> <2375@ucdavis.ucdavis.edu>
Reply-To: ken@cs.rochester.edu (Ken Yap)
Organization: U of Rochester, CS Dept, Rochester, NY
Lines: 27

|I compile common tex 2.9 using "gcc -O". That is 40% faster and
|20% smaller than the PASCAL version of tex. I didn't try it on
|the WEB2C version. But I think the result would be similar.

I think I would take correctness over any small gain in speed.

|Another advantages of using common TeX 2.9 is that they have a
|BIG and BIGG version. BIGG version has a memory size about five
|times of TeX.

I have a changefile (originally from Pierre Mackay) for a big version
of TeX in C, with 256kb of memorywords and it can easily be edited for
more. Without preloading, which wins very little in speed, according to
my tests, the big-tex text (code) size is actually *smaller* than the
normal version because most short/long conversions have been
eliminated.  The virtex binary is only 180k. The bss area is allocated
from VM anyway.  It runs just as fast as the normal version.

Tests indicate no reason not to install just one, non-preloaded, big
version of TeX.  No more out of memory messages (until the users get
overambitious again, anyway).  No need for separate tex, latex and
slitex binaries either. There is just initex and virtex now.

	Ken

PS: Don't ask me for the changefile. It is (will be) in the
distribution.