| ram disk [message #15971] |
Thu, 24 June 1999 00:00  |
Grady Daub
Messages: 22
Registered: June 1999
|
Junior Member |
|
|
I'm reading in some data that is in a weird format and I've devised a
technique which involves reading the one original data file, making a
temporary file for each different type of data, doing some stuff with
the data through reading the temp files, then, creating a new single
data file.
Does IDL have a feature where, instead of using the hard drive for the
temporary files, the RAM is used? (Us amiga users call it a RAM disk.)
No, I'm not talking about making a huge (huge) array
|
|
|
|
| Re: ram disk [message #16100 is a reply to message #15971] |
Mon, 28 June 1999 00:00  |
J.D. Smith
Messages: 214
Registered: August 1996
|
Senior Member |
|
|
Grady Daub wrote:
>
> I'm reading in some data that is in a weird format and I've devised a
> technique which involves reading the one original data file, making a
> temporary file for each different type of data, doing some stuff with
> the data through reading the temp files, then, creating a new single
> data file.
>
> Does IDL have a feature where, instead of using the hard drive for the
> temporary files, the RAM is used? (Us amiga users call it a RAM disk.)
> No, I'm not talking about making a huge (huge) array
You could use... a RAM disk. Joking aside, most OS's provide this
feature. If you're using a *good* OS, however, you'll really never need
this, since it will provide aggressive memory buffering of your file IO
by default, if you have the memory, and performance will be the same
either way.
Having said that, however, as far as I can see, there is no advantage to
trying to stuff everything in a RAM disk file... The huge (huge) array
you mention would not be any more huge (huge) than the file(s) you
created in RAM. So long as you use a compact representation within IDL
itself, you have nothing to lose and everthing to gain by working
entirely in memory. It's not hard to keep the data compact in IDL,
which only adds a little bookeeping to each variable, but otherwise is
quite close to machine format. If your data cannot possibly all fit in
memory, you could consider breaking it on input into chunks and
outputting the final data file one chunk at a time.
Good Luck,
JD
--
J.D. Smith |*| WORK: (607) 255-5842
Cornell University Dept. of Astronomy |*| (607) 255-6263
304 Space Sciences Bldg. |*| FAX: (607) 255-5875
Ithaca, NY 14853 |*|
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
