[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Gabriel benchmarks, HP DN10000 and others

	From att!MCC.COM!ai.gooch Tue Mar  6 09:12 CST 1990
	In-Reply-To: The message of 5 Mar 90 17:38 CST from davel@whutt.att.com
	Message-Id: <19900306151216.7.GOOCH@KANGCHENJUNGA.ACA.MCC.COM>
	The Symbolics numbers you gave are all quite slow compared to what I've
	seen before for the same machines.  What was your testing procedure?
Freshly booted optimized full-gc'ed worlds 
(not necessarily distribution worlds)
in which there were no background processes running
and no intermachine communications. Loaded 
sys:examples;gabriel-benchmarks.bin from the file server
(local machine in the case of the 3675 and the DN10000).
Ran the entire suite using (run-series-to-file). 
I had compiled the gabriel-benchmarks for the Symbolics and TI with 
declarations off, and for the DN10000 both on and off.  In all cases,
EGC was on and DGC was off.

I was particularly perturbed by the results on the 3650, and I ran them
twice, once on a fully-gc'ed but old world and once on a clean world,
with essentially identical results.  The quoted results were for the
clean world.

The exception to the above was the 3675, for which I used an old world 
which had not been fully gc'ed, because I did not want to lose the state of
some development work on the system.  However, no other 
user processes were running during the benchmark.

As Barry Margolin has pointed out, the Symbolics system patch level
dates from 1988.  We do not have a system administrator, and frankly,
no one has bothered to request an ECO tape in the last year.

Disclaimer: I have no special expertise in benchmarking.  My department
wanted some internal measure of the DN10000, and some SLUG members 
expressed an interest in the results.  Your mileage may vary.

David Loewenstern
AT&T Bell Laboratories
Whippany, NJ 07981
email: davel@whutt.att.com 
at&t: 201-386-6516