Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site watmath.UUCP
Path: utzoo!watmath!bstempleton
From: bstempleton@watmath.UUCP (Brad Templeton)
Newsgroups: net.unix-wizards,net.works,net.micro.68k
Subject: A real compile benchmark to try
Message-ID: <5555@watmath.UUCP>
Date: Thu, 21-Jul-83 00:46:34 EDT
Article-I.D.: watmath.5555
Posted: Thu Jul 21 00:46:34 1983
Date-Received: Thu, 21-Jul-83 01:42:25 EDT
Organization: U of Waterloo, Ontario
Lines: 23

Many people talk about the time for compiling hello.c
When you are doing this test, I think it is also important to do the
following, perhaps in /tmp

cp hello.c hello2.c
time cc hello.c -o 1 & time cc hello2.c -o 2&

so that both run at the same time.
This tests how badly a machine bogs down when it has something
in the background, plus code scheduling, swapping and disk
scheduling.

Bad systems take more than twice as long.
The good systems like an Onyx take only a little bit more than twice
as long to do both compiles.

Let's see how systems score on this benchmark.

My results are that onyx wins, fortune is ok, and unisoft is a dog.
(Sorry guys at unisoft)
Xenix is about half and half.
-- 
	Brad Templeton - Waterloo, Ont. (519) 886-7304