Minor update or useless upgrade?‎

[sabalan]

I have Visual Fortran Composer XE 2013 SP1 Update 2 which makes a Visual Basic caller ‎interface crash on exit. I have searched here and found that update 5 of that product solved ‎the problem. But my support license has expired and Intel’s license renewal suggests that I ‎should purchase Parallel Studio XE Cluster Edition (2017).‎

I downloaded the trial version of this 2017 and tried to install it, without success, because my ‎computer doesn’t even have a parallel CPU! So, this version 2017 is going to be useless for me ‎and my question is then if I still need to purchase a (for me) useless upgrade to gain access to the latest ‎update of version 2013?‎

[andrew_4619]

If you buy support for your existing product it gives access to all older version and versions up 12 months in the future. There is clearly no need or requirement to 'upgrade' the product to cluster edition.

[Steven_L_Intel1]

The Cluster Edition will install fine even on a single-core system, but as Andrew says, you don't need that. "Composer Edition for Fortran Windows" is all you need. If you want to test, customize the Cluster Edition trial install and set everything except the Fortran compiler to "Will not be installed".

[sabalan]

Thanks Andrew and Steve for your answers.‎

Steve, I do not come so far to see those options. Running ‎parallel_studio_xe_2017_update1_setup, both online or complete setup file, after extracting ‎and license agreement, it begins to check for prerequisites and then says “Machine CPU ‎architecture IA-32 is not supported. Installation will end because this architecture is not ‎supported.”‎

Thereby I can not be sure if updating my existing version or upgrading to 2017 will solve my ‎problem. A Windows hotfix “HotFix_HangOnClose_426940_intl_i386” (KB2493869) didn’t ‎solve it.‎

I have “Intel® Core™ i5 CPU 660 @ 3.33 GHz 3.47 GHz” (and I thought that I had a single ‎‎–core system!).‎

[mecej4]

One more item to check: does your BIOS support the "advanced" capabilities of your CPU? I am only suggesting that you check, not that you change any BIOS settings without a thorough investigation of the consequences of doing so.

The i5-660 CPU has two cores and can run four threads, but BIOS support may be needed to enable these and other capabilities. Check your system/motherboard/BIOS documentation.

[Steven_L_Intel1]

Even if you have a 64-bit capable CPU, if you have a 32-bit Windows installed you can't run 64-bit applications.  It's also not something you can change yourself without buying a new Windows license.

It is true that the 2017 products require 64-bit support for installation. If your Windows is 32-bit, then the 2016 product is the latest you can install. I would recommend that over the 2013 SP1 version.

[sabalan]

I go round and round on Intel's website and can not find the 2016 version. All roads lead to Rome == 2017!

[Steven_L_Intel1]

https://software.intel.com/en-us/articles/older-version-installation/

[sabalan]

With the serial number of trial version of 2017 I was permitted to try updates 5 and 6 for ‎my existing version 2013, which I downloaded and installed. Unfortunately these updates ‎didn’t solve my problem.‎

The problem is deep inside Intel Array Visualizer libraries. After viewing some 2D graphs ‎produced by a Fortran DLL and then exiting the Visual Basic caller application, in release ‎version I get “vshost32-clr2.exe has stopped working”, and in debug version I get a break ‎inside crt0dat.c on line “onexitbegin_new = (_PVFV *) DecodePointer(__onexitbegin)”.‎

Unchecking Visual Studio Hosting Process and/or enabling Unmanaged Code Debugging makes the ‎VB caller hang on exit with computer restart as the only way out.‎

I have traced the problem up to calls to avStartWatch and avCreateGraph2D but I am unable to go ‎further.‎

I understand that this is a wrong thread and wrong topic to discuss this problem but just wanted to tell ‎how and why updates didn’t help me. AND I know that Intel Array Viewer is declared dead. But is there ‎any replacement for it?‎

[mecej4]

If the VB program expects the graph display utility to be terminated in an orderly fashion and, perhaps, receive return a status code that signifies orderly termination, aborting the display utility (by closing the graph window, for example) may cause control to follow an unplanned path and cause a crash.

I do not use the Array Visualizer, so I cannot tell you what to fix, but the Array Visualizer documentation should help you with how you are expected to close graph display windows.

It appears that you found some report of a compiler update fixing some problem with VB calling AV routines, but you have now found that your problem is only superficially similar to the one in that report.

[sabalan]

The report I found didn’t deal with some problem with VB calling AV, but just crash on exit ‎from VB (or was it C#?) AND breaking on the same line in crt0dat.c. That seemed to me ‎similar to my problem. Unfortunately I can not find that report again.‎

I follow all procedures needed to initialize and end/exit AV. I have been working with both ‎Compaq Array visualizer and Intel Array Visualizer since they were introduced. I have several ‎applications calling and using IAV compiled with XE 2013 and functioning without any ‎problems. But those are QuickWin applications. The problem appears when the caller is an ‎application made with VS 2012.‎