- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi there,
I am developing a 64-bit numerical modeling program and looking for a right compiler. Intel C++ Compiler 9.1 seems ideal for me. Here are my requirements
Host System
Windows Server 2003 x64
Intel Dual Core Xeon 7120 (multiple processors)
Targe Systems:
Multiple Intel dual-core 64-bit processors
Any 64-bit Windows (XP, Vista, Server 2003)
OpenMP support
Possible backward compatiblity for IA-32 systems.
Portability to Linux Clusters (MPI)
It seems to me that both Intel C++ and Fortran compilers are great to meet my goal. I do not know which one to choose. Will Intel Math Kernel Library (MKL) support both Intel C++ and Intel Fortran compilers?
The performace is one of my top concerns. I have some legacy code in C++ which may take one man-month to port it into Fortran.
Any advice will be appreciated. Thank you.
Jim
Link Copied
2 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Yes, MKL works fine with both Intel C++ and Fortran. Your license for an Intel compiler covers use of the 32- and 64-bit versions, and any new releases during the year.
When you say legacy C++, does that mean it won't run, or at least not efficiently, on standard compilers? Intel C++ is pretty much equal to Fortran in terms of ability to optimize with vectorized SSE and OpenMP (if you use all the restrict keyword and pragma extensions), so it's your call.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I have a Borland C++ program with some VCL GUI. Now I am developing a GUI-less application based on my existing C++ code.
I learned that Fortran code often runs faster than C++ code for number crunching. I was also told that C++ code can run as fast as Fortran code. I am an intermediate level programmer in both Fortran and C++. I am kind of lost and do not know which way to go (which Intel compiler to choose). Thank you.
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page