fouts@bozeman.ingr.com (Martin Fouts) (06/01/90)
In article <10305@batcomputer.tn.cornell.edu> kahn@batcomputer.tn.cornell.edu (Shahin Kahn) writes:
From: kahn@batcomputer.tn.cornell.edu (Shahin Kahn)
A supercomputer is a machine whose purchase is a purely political decision!
(smiley here).
There's actually some truth to that. and it can be argued why that's the way
it ought to be, IMO.
There is a lot of truth in your statement, but it isn't restrictive
enough. A lot of nonsupercomputermachines (sorry) are also bought for
purely political reasons....
--
Martin Fouts
UUCP: ...!pyramid!garth!fouts ARPA: apd!fouts@ingr.com
PHONE: (415) 852-2310 FAX: (415) 856-9224
MAIL: 2400 Geng Road, Palo Alto, CA, 94303
If you can find an opinion in my posting, please let me know.
I don't have opinions, only misconceptions.hamrick@convex1.convex.com (Ed Hamrick) (06/04/90)
In article <395@garth.UUCP> fouts@bozeman.ingr.com (Martin Fouts) writes: >In article <10305@batcomputer.tn.cornell.edu> kahn@batcomputer.tn.cornell.edu (Shahin Kahn) writes: > > A supercomputer is a machine whose purchase is a purely political decision! > (smiley here). > >There is a lot of truth in your statement, but it isn't restrictive >enough. A lot of nonsupercomputermachines (sorry) are also bought for >purely political reasons.... A particularly common example of this phenomenon is the practice of procuring supercomputers based on single-job performance instead of throughput (based on representative workload). I'm sure that many of you are familiar with large Cray sites with hundreds of users with an average job turn-around time (for 1-hour jobs) of a week. Even though the workload at these sites is dominated by lots of relatively short jobs, hardware upgrades (and the justification for more money) is almost always based on single job performance (how many times can you say Grand Challenges? 8-). Of course, for any broad generalization there exist specific counterexamples. However, I challenge those readers of this newsgroup from the National Supercomputer Centers et. al. (e.g. NASA Ames NAS) to post computer utilization statistics such as median job turnaround time (by job run time), number of jobs (by job run time) and other similar data. I've noticed that many of you seem to have plenty of spare time (still waiting for your jobs to run? ;-). Regards, Ed Hamrick
toby@hobbes.nas.nasa.gov (Toby E. Harness) (06/05/90)
In article <102750@convex.convex.com> Ed Hamrick writes: >I'm sure that many of you are familiar with large Cray sites with hundreds >of users with an average job turn-around time (for 1-hour jobs) of a week. > .... >Of course, for any broad generalization there exist specific counterexamples. >However, I challenge those readers of this newsgroup from the National >Supercomputer Centers et. al. (e.g. NASA Ames NAS) to post computer >utilization statistics such as median job turnaround time (by job run >time), number of jobs (by job run time) and other similar data. I've >noticed that many of you seem to have plenty of spare time (still waiting >for your jobs to run? ;-). the following command was run on the NAS YMP-8/128 (reynolds); the data is for the previous 7 days. this may or may not have been a typical "week". only jobs that spent longer than 1 second waiting in a queue are counted. % qavg 1w QUEUE N MEAN MEDIAN MAX q4M_1h 164 02:23:09 01:32:41 10:14:48 q4M_4h 71 03:15:50 02:25:13 15:00:46 q10M_20m 1277 00:06:25 00:00:02 08:57:30 q10M_1h 154 01:17:21 00:03:53 08:56:10 q10M_4h 67 05:48:45 01:31:22 32:49:49 q16M_20m 73 00:22:01 00:00:02 08:21:52 q16M_1h 60 02:01:43 00:30:02 13:16:17 q16M_4h 50 05:40:59 04:20:14 17:08:15 q32M_20m 15 06:20:27 03:44:31 24:00:26 q32M_1h 23 09:37:34 11:35:17 15:53:11 q32M_4h 11 09:14:52 09:12:48 32:17:20 q64M_20m 5 04:06:29 04:51:00 07:58:34 q64M_1h 7 06:22:25 06:58:02 16:40:20 q64M_4h 7 11:47:17 12:30:39 22:00:28 [the leading 'q' means it's a production NQS queue (i've removed the others) the first number is the memory size, in megawords the second is the time limit the times are in hour:minute:second format] toby harness toby@nas.nasa.gov "not a cute sign-off" -- Toby Harness General Electric Corporation (415) 694-4626 NASA Ames Research Center toby@Ames-NAS.ARPA Mail Stop 258-6 ...!ames!amelia!toby Moffet Field, CA 94035