This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Compressed MOD files?


Steve Kargl wrote:
On Mon, Oct 30, 2006 at 09:06:24PM +0200, Erik Edelmann wrote:
On Mon, 30 Oct 2006, FX Coudert wrote:

ssed or not when reading, which is pretty easy) It increases CPU
usage, but it gets the disk space used by module files by a factor 5
to 6 on my own codes.
This sounds interesting, but how much real space usage reduction do you see?
a factor of 5 to 6? e.g., from 27 megabytes to 5.2 megabytes on the real-life F95 code I currently use for my work.
My first reaction when I saw your first message in this thread was "Does it
really matter?".  But if there are real life code that prdoces 27 MB of
*.mod files, I guess compression would be nice.  Not of huge importance
perhaps, with the hard disk sizes of today, but nice.  Since it shouldn't
add much complexity to the compiler, my vote is "yes".



gfortran doesn't cache a previously read module.


module foo
contains
subroutine A
use B
end subroutine A
subroutine C
use B
end subroutine C
end module


Module B will be read twice, and hence, it would be decompressed
twice.  How much overhead is added by the decompression phase?

I think it would be better to fix this multiple reading first before bothering with compression.


Jerry


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]