I am working on a C++ simulation framework. In the current version, I identified two bottlenecks. The first one is Random number generating, and second one (which motivates this question) is the use of highly iterative processes (heavy for loops).
Does some of you could give me a feedback on gcc 4.8 concerning loops optimization.
I have read some of the few available literature on the subject but, may be you can make more explicit to me (with more simple words), what I can expect from this early version.
What do one call loop normalization, for example... Or what does it has to do with multi-core architecture...
Also, what is behind the optimization options :
-ftree-loop-optimize Perform loop optimizations on trees. This flag is enabled by default at -O and higher.
-ftree-loop-linear Perform loop interchange transformations on tree. Same as -floop-interchange. To use this code transformation, GCC has to be configured with --with-ppl and --with-cloog to enable the Graphite loop transformation infrastructure.
-floop-interchange Perform loop interchange transformations on loops. Interchanging two nested loops switches the inner and outer loops.
Any code optimization and/or compiler experts over there ? Thanks a lot.
NB : I currently work under Debian 4.7.2, gcc 4.7
