Problems with programs compiled with CVF 6.5 under windows XP

Problems with programs compiled with CVF 6.5 under windows XP

Hello all,

In all my programs (Quickwin Applications) I use an infinite loop to keep the program running without closing. The code of this infinite loop is:
do while (.true.)
end do
Well, I have three fortran compilers: fortran power station, digital visual fortran and compaq visual fortran 6.5 and i use windows XP. Well, when I compile a quickwin application that has an infinite loop using either Power Station or Digital Visual Fortran I get 100% CPU usage that is due to the infinite loop but the computer doesn't become slow. But when i compile at CVF 6.5 i get also 100% CPU but the difference is that the computer becomes really slow, making impossible to use other programs.
Do any of you guys know what is happening?

Greetings,
Ivomar

10 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

No I don't -- but if you put a CALL SLEEPQQ(1000) into that infinite loop your problems will vanish :-).

Jugoslav

Jugoslav
www.xeffort.com

I am having a similar problem. My program consists of an integration loop that integrates hydrodynamics and transport code over time. I have a new machine using XP and while in the debugger, the code executes quite fast from one iteration to the next until, at some random iteration but usually less than 5 iterations from the start, execution slows to a snails pace and basically brings my system to a halt. So much so that when I try to bring up the task manager to see what is happening with regards to resources being used, it will not come up after five minutes, thus requiring rebooting.

However, if I don't step through the program one iteration at a time and set a breakpoint that breaks when the number of iterations is > 2000, then there are no problems until I try to proceed one iteration at a time. I have no such problems with the same code on my laptop running Windows 2000. Any ideas?

I have a similar problem (halting the system in long minutes) while debugging in XP, which I did not have in W2K.

Sabalan.

Greetings,

According to Compaq (Click the Product Compatibilities link and scroll down a bit), "CVF 6.6 supports Windows XP." That implies not under earlier versions.

Hope this helps,
Mark Tompsett
Programmer/Analyst

I missed the version number in the thread title. I have 6.6B but it doesn't matter. The problem persists.

Sabalan.

I am using 6.6b and am having the problem. Anyone else?

I've just received an e-mail from C. Staudinger of Intel:

Saw the conversation in the Forum and just wanted to let you know that Steve
is on a well deserved vacation. I asked one of the CVF engineers for some
feedback and here is what he had to say:

I endorse the suggestion of "jugoslavdujic" to put a CALL SLEEPQQ in the
busy loop.
This should solve "ivomar"'s problem. As for "tmcole", I don't know if the
program has
the "busy loop in the main program with real work being done by other
threads" structure.
If CALL SLEEPQQ does not solve "tmcole"'s problem, I suggest he/she should
submit a
vf-support report with more information (if possible a program showing the
problem).
The same for "sabalan".

This particular problem is caused by the fact that Quickwin programs have
several
threads, one running at normal priority, and the other at above normal
priority.
This particular scheme was developed for WNT4 so that Quickwin programs
could
compete with other applications. The original setup had the Fortran thread
running
at below normal priority so it could get almost no time if another
application was
running.

My guess is that the WinXP scheduler uses a different algorithm from that of
Win2000
(or WNT4) which makes the problem more obvious. Hopefully, the CALL
SLEEPQQ will
solve the problem. If not, I need a sample program.

I hope this will be enough for you to provide an initial response. If
needed, I'll provide support through vf-support requests.

Jugoslav

Jugoslav
www.xeffort.com

Our statement that CVF 6.6 supports XP means that's the earliest version that we tested on XP. However, we aren't aware of any general problems with any previous version of Visual Fortran on XP. The only XP-specific issue we know of is an obscure one relating to GETCHARQQ in a QuickWin application with a key being held down - I think that one is fixed in 6.6B.

I concur that a QuickWin app that has an infinite loop at the end with no "sleep" in the loop, is going to cause problems for the rest of the system.

Steve

Steve - Intel Developer Support

My problem seems to has been IntelliMouse. I installed the latest driver from its home page and now W-XP, DevStudio and CVF-debug work as they should. I had the same problem even with W2K. Steve helped me to solve it. See here.

Sabalan.

Message Edited by intel.software.network.support on 12-09-2005 01:03 PM

Leave a Comment

Please sign in to add a comment. Not a member? Join today