Intel Composer XE 2011

[schotti]

Hi all,

I am frustrated. I installed ComposerXE-2011 as update for my Visual FTN 11.1.067.
Using DEBUG leads to catastrophic crashes of well-designed and proven software.
Using the "RELEASE" option the code is properly compiled and linked.
So - how do I DEBUG in future ?

I lost last weekend until today of useless work.
Why dont INTEL check CAREFULLY new updates BEFORE distributing?
I have no time to play the tester free of charge.
Is there somebody out there who can help me?

Kind regards

[anthonyrichards]

You have my sympathy. Nothing I have read in these threads since its release induces me to consider installing the Composer update until occurrences such as yours no longer happen.

My feelings can be summed up in the following mangling of a quote from Hamlet:

Shakespeare:
"Who would fardels{heavy sacks} bear, to grunt and sweat under a weary life,
but that the dread of something after death,
that undiscovered country from whose bourn no traveller returns,
troubles the soul, and makes us rather bear the ills we have,
than fly to others that we know not of?"

Me:
"Who would keep 11.1.067, to grunt and sweat under a weary life,
but that the dread of something after Composer XE 2011,
that undiscovered update from whose install no programmer returns unscathed,
troubles the soul, and makes us rather bear the version we have,
than fly to others that we know not of?"

[rase]

I also had a lot of problems with a large Fortran project after installation of Composer XE 2011.

But: all the problems were home-made, because previous versions of CVF and Intel Fortran were not able to detect certain kinds of flaws in my programs. I took the time to correct the errors/warnings generated by V12, and now everything is fine (so far). It was worth the time, because with V12 the execution times were reduced much more than I expected.

[Steven_L_Intel1]

schotti,

I'm sorry to hear that you are having problems. We did test extensively for six months, with hundreds of beta testers around the world. But sometimes problems remain.

It is not yet clear that it is a fault of the compiler. Even "well-designed and proven software" often has bugs that are not revealed until a compiler challenges an invalid assumption of the programmer. We'll be glad to help you sort this out - if you can provide us a test case, either here or through Intel Premier Support, we'll take a look.