GMP, MPFR and VS version compatibility

Michael Abshoff Michael.Abshoff at
Fri Jul 13 11:11:07 CEST 2007

Jim White wrote:
> Michael

Hello Jim,

> Do I understand correctly, your problem is that there
> is no gcc for Win64?

I wouldn't call it a problem, but the lack of 64 bit support for gcc on
Windows is certainly an issue for the code I have to maintain. Other might
be forced to use the MS toolchain and some people even enjoy using it
looking with confusion at people who tell them that vim, make and gcc
ought to be enough for everybody. I certainly come from the vim, make and
gcc camp, but ultimately open source is also about choice. You can do it
your way and if that includes using the MS toolchain: more power to you,
but don't bother me with your bugs.

And by the way: feel free to insert you favourite editor in there, there
is no need to start a flamewar a la "emacs vs vi", the Microsoft/Windows
sucks debate is already a great source for conflict.

> While it might not help you in the short term, the key
> development would seem to be filling that hole.

Yes, fortunately there is some movement in getting this fixed. It has been
easily 5 years since 64 bit betas of Windows for x86-64 became available
and while 64 bit wasn't that big of a deal in 2002 (who had 2/4G+ memory
back then in their desktop) it is an issue today.

> I keep forgetting the MinGW compiler is (again, if I
> am correct) independently "maintained" - it still
> seems to be stuck at version 3.3 or thereabouts.
> What's needed is for MicroSquash to provide it - it
> would make good business sense, surely, to have a
> fully-functional gcc compiler,

They have one: SFU or whatever it is called today contains gcc and various
gnu utils. But is isn't free as in freedom (the whole package which
contains more than GPLed components) and last time I checked 32 bits only.

> so they would have more
> chance to flog their (nefarious) wares in the
> "computational" marketplace, being able to swear with
> hand on heart that any open-source math package (GMP,
> MPFR, CLN, Pari-GP, etc etc) should be able to run on
> any Win32/64 box.

I agree, but I doubt that this is going to happen for political reasons. I
am certain that the software engineers at Microsoft would certainly be
capable to achieve such a thing, but management sees choice and freedom as
in free of a thread to their empire. I don't want to sound like a raving
madman bend on destroying the commercial software industry, but
competition on technical merit would certainly do some companies good. The
enemy was IBM up to the 90s at which point Microsoft replaced IBM. And
look at IBM now: open standards and a good memeber of the open source
community. The young people probably have a hard time understanding why
IBM was the way it was back then. Fast forward 10-15 years from now and
maybe Microsoft will have learned to play nice with others and the big
enemy is somebody else.

> Not that I'd ever buy another Win box myself, mind
> you!   There are limits ....          ;-)

I agree 100%

> Cheers
> Jim White
> ANU, Canberra



More information about the gmp-discuss mailing list