processor optimisations

[forall]

I will be running floating-point-intensive CVF code (with average runtime up to several days per process) on several processors, including a Pentium-M 1.7GHz, an (dual)Opteron 2Ghz and a (dual)Xeon 3.2GHz. Is it worthwhile fiddling with the processor-optimisation options, given that none of these processors existed when CVF was written. All the compilation will be done on the P-M and I will be transferring the exes to the other machines? I am trying to determine whether to embark on a potentially lengthy benchmarking exercise.

Is the situation different for the Intel Compiler?

many thanks!

[Steven_L_Intel1]

For the compiler abort problem, please report that to Premier Support and supply a sample.

/QxN does, I am told, enable some additional optimizations over /QxW, but not enough to be worth discarding non-Intel processor compatibility.

[Intel_C_Intel]

Dear Forall,
Note the difference between vectorization for the multimedia extensions SSE/SSE2/SSE3 (enabled by default with switches /QxKWNP) and parallelization for e.g. dual cores and/or the HT technology (enabled with switches /Qparallel [implicit parallelism] or /Qopenmp [explicit parallelism]). I initially recommended the former, and then answered your question on the latter assuming this is what you referred to.
First, I would just see how well automatic vectorization works for your applications (viz. vectorizing an EXP function has great performance potential when occurring in a hot spot!). If you are interested in more background on vectorization for multimedia extensions (including programming guidelines on how to make the compiler more effective in extracting vector instructions from your code), see the Software Vectorization Handbook at http://www.intel.com/intelpress/sum_vmmx.htm (using the C programming language for most examples, but similar concepts apply to Fortran).
Aart

PS. I hope we are notoverwhelming you withtoo muchinformation