[2480] in Release_Engineering
[jfc@ATHENA.MIT.EDU: RVD server]
daemon@ATHENA.MIT.EDU (Mark Rosenstein)
Sun Nov 18 18:37:01 1990
Date: Sun, 18 Nov 90 18:36:48 -0500
From: Mark Rosenstein <mar@MIT.EDU>
To: rel-eng@MIT.EDU
Has this makefile been fixed yet?
-Mark
To: mar@ATHENA.MIT.EDU, salemme@ATHENA.MIT.EDU
Subject: RVD server
Date: Wed, 15 Aug 90 17:04:04 EDT
From: John Carr <jfc@ATHENA.MIT.EDU>
I recommend that operations build and install its own copy of the RVD server
for release 7.1. The release version was built with -g instead of -O, and I
have been told this won't be fixed. This option on the RT reduces the speed
of a program that does a lot of memory access by about a half (the compiler,
to aid debugging, puts an instruction after each C statement that causes the
processor to wait for all outstanding memory references to complete [in
addition to not doing any optimizations]).
I think RVD performance is important enough that the RVD servers should get
non-release software.
