| Re: IDL and Dual Processor PC's [message #15628] |
Thu, 03 June 1999 00:00 |
Bruce L. Gotwols
Messages: 2
Registered: June 1999
|
Junior Member |
|
|
I am running dual processors under Linux kernel versions 2.0.36 (patched
to support symmetric multi processing), as well as the new Linux kernel
2.2 which supports dual processors without a patch. In both cases I
have been very happy with the performance. Note however, that to obtain
the benefit of dual processing you have to run two independent copies of
IDL. (Or IDL and a C++ routine for example.) When I run a single copy
of time_test2 (using the /nofileio keyword) it takes about 7
seconds to complete. Running a second copy of time_test2 increases the
time to complete to 8 - 10 sec. I have not yet figured out why
sometimes the performance hit is only 1 additional seconds, and other
times it is 3 seconds. Probably something to do with caching. I am
currently searching for a way to tell IDL which processor to use and any
other tuning parameters which would make it more consistent.
Even in the worst case the second processor is doing us a lot of good.
In practice we won't be using two copies of IDL, but rather a C rouutine
that pulls high speed data from a 100 base-T network socket, stores it
on disk, and simulatenously writes some of data into shared memory where
IDL can grab it and do Quick-Look proicessing and display on it. So yes
I'm very happy with dual processing in our work. (But not at all happy
with RSI's support of Linux...see my related post in this newsgroup.)
Cheers, Bruce Gotwols
----
Bruce L. Gotwols
Johns Hopkins University, Applied Physics Lab., Laurel MD 20723
Internet: gotwols@tesla.jhuapl.edu
Phone: 240-228-4543 FAX: 240-228-5548
Space Oceanography Group Home Page -- http://fermi.jhuapl.edu
Tanya Lancaster wrote:
>
> We are looking into purchasing a dual processor pc. I was wondering if
> there would be a notable increase in speed for running IDL programs. Right
> now we handle large medical image data sets and operations on the data sets
> take up to several hours on a PII -400, 256 mb memory.
>
> -Tanya Lancaster
|
|
|
|
| Re: IDL and Dual Processor PC's [message #15633 is a reply to message #15628] |
Thu, 03 June 1999 00:00  |
mgs
Messages: 144
Registered: March 1995
|
Senior Member |
|
|
In article <7j62f0$krm$1@pale-rider.INS.CWRU.Edu>, "Tanya Lancaster"
<lancaste@morph.ebme.cwru.edu> wrote:
> We are looking into purchasing a dual processor pc. I was wondering if
> there would be a notable increase in speed for running IDL programs.
Nope. IDL is single-threaded, implying it cannot divvy it's processing up
among multiple processors. Once you start IDL it stays on the same CPU.
You can try some interesting things like spawning addition copies of IDL
onto other processors. I tried that about 7 years ago on a Sun MP/670 with
4 processors. Not worth the effort was my conclusion.
> Right
> now we handle large medical image data sets and operations on the data sets
> take up to several hours on a PII -400, 256 mb memory.
If you're running other tasks on the same system the additional CPU would
be helpful, but not for IDL alone.
I would love to hear that I'm wrong about this.
--
Mike Schienle Interactive Visuals, Inc.
mgs@ivsoftware.com Remote Sensing and Image Processing
http://www.ivsoftware.com/ Analysis and Application Development
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
