Simpler Navigation for Servers and Operating Systems - Please Update Your Bookmarks
Completed: a much simpler Servers and Operating Systems section of the Community. We combined many of the older boards, so you won't have to click through so many levels to get at the information you need. Check the consolidated boards here as many sub-forums are now single boards.
If you have bookmarked forums or discussion boards in Servers and Operating Systems, we suggest you check and update them as needed.
Languages and Scripting
cancel
Showing results for 
Search instead for 
Did you mean: 

Fortran90 with +Ofltacc flag on Itanium2 and PA-RISC

Vefa Kucukboyaci
Occasional Contributor

Fortran90 with +Ofltacc flag on Itanium2 and PA-RISC

Hi,
Few months ago, I compiled a large Fortran code on an Itanium2 machine with HP-UX11.23 / F90 V2.8 using +O3 optimization and +Ofltacc flag, which yielded the exact same numerical results as the version compiled with the -debug version. This was very good news, since we then had a fast running code which produced consistent results as the debug version.

Then, we moved to a new PA-RISC machine running HP-UX11.11 / F90 V2.8.2. The same code compiled with the same options (+O3 +Ofltacc) does not show the same behavior now.

I was wondering if this is an expected behavior, if there is any way to duplicate the same kind of floating-point consistency between the debug and optimized versions of the code on the PA-RISC platform running HP-UX11.11 / F90 V2.8.2.

I'd aprreciate your help/suggestions.

Sincerely,
Vefa Kucukboyaci
1 REPLY
Highlighted
Zygmunt Krawczyk
Honored Contributor

Re: Fortran90 with +Ofltacc flag on Itanium2 and PA-RISC

Hi,

the accuracy of Itanium2 is better than PA-RISC. Using +Ofltacc on both processors leads to different results. Read about the accuracy of libm on HP-UX/Itanium2:

"The Libm Library and Floating-Point Arithmetic
in HP-UX for Itanium®-Based Systems"
http://h21007.www2.hp.com/dspp/files/unprotected/Itanium/FP_White_Paper_v3.pdf

Regards,
Zygmu