- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
So, here's my problem:
Due to change in jobs, I have to start using MS Windows.
I think I will be OK with it, but I have heard it has a bug in it somewhere.
And I will need an IDE to work on my F90 programs. Now, the version of the Intel F compiler for MS windows says that I have to buy something from MS called visual studio or visual c++, but when checking the local dealer, I note that he has *many* boxes with similar names on the shelf.
The cheapest is the "visual c++ .net standard edition". Would that work for me?
I just want the IDE to help me use F90. My applications are pretty much all calculations, and the idea of "managed code" scares me, so am not going to use the C++ net compiler ever, but I actually find the visual basic languae intriguing for QoD applications.
Thanks for any help!
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
MS Windows has LOTS of bugs in it. Try not to think about it.
Mike D.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
No unsigned integers in the forseeable future.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Dear Benta,
Please note that the Intel compilers for both C/C++ and Fortran and for the various platforms more or less share the same source base between front-end (language specific) and back-end (target specific). And since C/C++ happily supports unsigned integers, nothing in the compiler itselfinherently prohibits unsigned integer support. In a few rare cases, you can actually observe some of this support even in a Fortran fragment. For example, compiling the following fragment
program sat
integer*1 a(256)
integer*2 b(256)
integer x
do i = 1, 256
x = b(i)
if (x.lt.0) x = 0
if (x.gt.255) x = 255
a(i) = x
enddo
end
for IA32 as follows
sat.f(6) : (col. 9) remark: LOOP WAS VECTORIZED.
yields vectorized code that maps the whole loop onto an unsigned saturating conversion idiom:
$B1$3:
movdqa xmm0, XMMWORD PTR SAT$B$0$0[0+eax*2]
packuswb xmm0, XMMWORD PTR SAT$B$0$0[0+eax*2+16]
movdqa XMMWORD PTR SAT$A$0$0[eax], xmm0
add eax, 16
cmp eax, 256
jb $B1$3
But such cases are, alas,rare in Fortran, becauseall following code that uses this array must be aware of the unsigned interpretation of its contents (e.g. naivelyusing the 255 ==0xff value in "normal" integer expressions would be interpreted as signed -1).....
Aart Bik
http://www.aartbik.com/
Message Edited by abik on 05-17-2005 06:54 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page