Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!ukma!gatech!rayssd!d2b
From: d2b@rayssd.RAY.COM (Donald A. Borsay)
Newsgroups: comp.os.vms
Subject: Re: VAX 750/780 comparison
Message-ID: <1511@rayssd.RAY.COM>
Date: Wed, 16-Sep-87 12:15:42 EDT
Article-I.D.: rayssd.1511
Posted: Wed Sep 16 12:15:42 1987
Date-Received: Sat, 19-Sep-87 01:02:12 EDT
References: <12331690952.196.OC.CIVIL@CU20B.COLUMBIA.EDU>
Sender: d2b@rayssd.RAY.COM (Donald A. Borsay @ Raytheon Company, Portsmouth RI)
Reply-To: d2b@rayssd.RAY.COM (Donald A. Borsay)
Organization: Raytheon Company, Portsmouth RI
Lines: 21
Keywords: FPA paging good_deal
Summary: Capacity for another 6 to 10 users would be reasonable

In article <12331690952.196.OC.CIVIL@CU20B.COLUMBIA.EDU> 
OC.CIVIL@CU20B.COLUMBIA.EDU (Holt Farley) questions the performance
gains of upgrading from a 750 to a 780, when operating in a engineering
environment.  Holt points out that the current 750 has 8mb of memory
and the proposed 780 would have 16-20mb.

In a program development environment, memory is the highest demand.  The linker,
for one, is a big hog.  The run-time environment for some of the applications
you describe (finite element and symbolic langauges) most likely contribute
to your lack of CPU time.  Doubling your memory will help alot (need hard
parameter values and fault rates to guess at how much).  Since a 750 is
60% of a 780, the CPU upgrade should give you a net increace of 66% in CPU
resouces.

The one configuration option you do not mention is the Floating Point 
Accelerator.  In your environment, I would strongly urge you to have one.
Hope this helps.
-- 
Don    |Raytheon Company, Submarine Signal Division, Portsmouth, RI
Borsay |ARPAnet: d2b%rayssd.RAY.COM@a.cs.uiuc.edu
       |UUCPmail: {allegra, decvax!brunix, linus!raybed2}!rayssd!d2b