home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
TeX 1995 July
/
TeX CD-ROM July 1995 (Disc 1)(Walnut Creek)(1995).ISO
/
tex-k
/
tex-k-archive
/
000049_manuel@gaston.CS.Berkeley.EDU_Wed Mar 8 05:36:51 1995.msg
< prev
next >
Wrap
Internet Message Format
|
1995-03-08
|
2KB
Received: from gaston.CS.Berkeley.EDU by cs.umb.edu with SMTP id AA02852
(5.65c/IDA-1.4.4 for <tex-k@cs.umb.edu>); Wed, 8 Mar 1995 16:36:56 -0500
Received: (from manuel@localhost) by gaston.CS.Berkeley.EDU (8.6.10/8.6.9) id NAA05337; Wed, 8 Mar 1995 13:36:51 -0800
Date: Wed, 8 Mar 1995 13:36:51 -0800
From: Manuel Fahndrich <manuel@CS.Berkeley.EDU>
Message-Id: <199503082136.NAA05337@gaston.CS.Berkeley.EDU>
To: tex-k@cs.umb.edu
Subject: slow path searching on HP 9000/700 series
Reply-To: manuel@CS.Berkeley.EDU
This is not a bug report, rather a question/comment.
I'm experiencing a big difference in path searching speeds between an
HP 9000/712 and a SUN4 sparc under SUNOS 5.3.
I'm using kpathsea 2.6 with the patch for the texmf
distribution. Using the debugging output, I made sure that both, the
SUN and the HP touch exactly the same files, have the same paths
etc...
Both the sun and the hp get the tex tree over NFS from the same
machine. The directory setup follows the one described in HIER.
Here are the measured real-time execution times (ps1 is a small
1 page document testing postscript fonts.)
Sun
w/o ls-R latex2e ps1 3.58 sec (best out of 3)
w/ ls-R latex2e ps1 1.75 sec (best out of 3)
HP
w/o ls-R latex2e ps1 16.89 sec (best out of 3)
w/ ls-R latex2e ps1 5.23 sec (best out of 3)
What surprises me is that the sun is faster w/o ls-R database than the
HP with ls-R.
I'm wondering whether there is a problem in the code for the HP, or if
this is a problem in the NFS setup for the HP. What is really
frustrating is that the old latex runs much faster on HP than latex2e,
(albeit with a different tex tree). Below is a comparison between
latex209 and latex2e on HP for a 7 page document. LaTeX2e was NOT
running in compatibility mode.
On HP
TeX, Version 3.1415 (C version 6.1) LaTeX2e <1994/12/01> patch level 1
latex2e paper2 5.79 sec (best out of 3, w/ ls-R)
TeX, C Version 3.14t3 LaTeX Version 2.09 <25 March 1992>
latex paper 2.74 sec (best out of 3)
I'm wondering if such behaviour has been observed by other people and
if someone found a reason for it.
-Manuel