We have a linux cluster with athlon processors. I am trying to compile the NASA overflow code and would like to get as much optimization as possible. Are there any compiler flags to improve the performance of the code other than -O3.
There are exceptions, but for Athlons which have P-III compatibility, the P-III options should work well (/QxK). You would need to profile, to determine where /O3 might be faster than /O2.
Sorry, the linux forum title was off my screen. The linux version of that option is -xK. A minute spent reading 'info ifc' or 'ifc -help' would be a valuable investment.
