| Be sure you're measuring your computer and not your network... [message #819] |
Wed, 12 May 1993 16:50 |
deutsch
Messages: 19
Registered: February 1992
|
Junior Member |
|
|
In article <1993May12.125311.1@univrs.decnet.lockheed.com>, sextonm@univrs.decnet.lockheed.com writes:
> I'm posting this for Mr Gurman
>
> MK Sexton
> ______________________________________
> Multitasking OS's seem to give a spread of results; here are some of
> mine for a VAXstation 4000 Model 90, also running OpenVMS 5.5-2, with 64
> Mbyte of memory (although the user's working set extent is more
> relevant: mine is 60K pages):
>
> Comp. Geo. Avg. Graphics Geo.
> 47.3 0.79 2.36 0.40
> 34.1 0.70 2.75 0.53
> 34.2 0.70 2.14 0.31
>
> Guess it's the old process initialization lethargy of VMS showing.
>
Be careful! Especially when just posting your totals and not every
individual test, the last test, which is the intensive disk test, can
totally destroy your results if the current default (working) directory
is not physically attached to the computer running IDL. A local disk will
yield a time on the order of .5 - 3 seconds while a network disk can be on the
order of 20 seconds, which will make any machine look like a sloth!
For example on a VAXstation 4000 M90 (maybe 60?) on local disk, I get
Comp Total 17.9 seconds and on network disk, 26.9 seconds! This machine
is also running OpenVMS 5.5-2.
I think, in general, it is true that disk access under VMS is slower than
under e.g. Unix, which is important to know, but be sure to run all tests
with a local disk, so at least the comparison is fair.
This note of caution reponds Joe's message, but there are several
postings in this rash of benchmark fever which are guilty of using net disks.
-Eric
> For comparison, the DEC 3000 Model 400 AXP (Alpha workstation),
> OpenVMS AXP V1.0, 64 Mbyte numbers are:
>
> Comp. Geo. Avg. Graphics Geo.
> 18.7 0.43 1.80 0.32
>
> You might want to post these results, since my news reader is
> gagging on attempts to post today.
>
> Joe Gurman
>
> --
> J.B. Gurman / Solar Physics Branch / NASA Goddard Space Flight Center /
> Greenbelt MD 20771 USA / gurman@uvsp.gsfc.nasa.gov
> | Federal employees are prohibited from holding opinions under the Hatch Act.|
> | Therefore, any opinions expressed herein are somebody else's. |
Eric Deutsch
University of Washington
Department of Astronomy FM-20
Seattle, WA 98195
deutsch@astro.washington.edu or UWSPAN::JANUS::DEUTSCH
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
