Can someone (a moderator?) tell me what exactly is different about the QNX-compatible Intel C++ compiler and the standard Intel C++ compiler? It would seem that both generate Intel x86 32-bit code - perhaps the object formats are different? What else is different?
Many thanks for your time,
The QNX targeted compiler is different than our Linux targeted compiler in the following ways:
1. QNX targeted compiler is hosted on Windows and targets QNX Neutrino RTOS. Our Linux compiler is Linux hosted.
2. QNX targeted compiler integrates into Momentics, QNX's eclipse-based IDE and qcc.
3. The object format question is really an OS question. I believe Linux & Neutrino are both elf/dwarf, but that doesn't mean you would be able to take an obj compiled by our Linux compiler and link it and have it work on Neutrino.
Both compilers target IA32.
Hope this helps.
I appreciate the reply - but now that I have an eval copy of the QNX Intel compiler to play with, I'm seeing one other big difference: no mention of OpenMP, and no OpenMP associated command-line options (e.g. -openmp). What's the story there?
The QNX-targeted compiler currently does not support OpenMP. The requisite libraries are not part of the compiler. In addition, the libraries haven't been interfaced to the QNX threading libraries so there is no easy path to make it happen.
Can I ask what sort of work were you looking at doing using OpenMP on QNX?