Intel® Moderncode for Parallel Architectures
Support for developing parallel programming applications on Intel® Architecture.

What are the recommended gcc optimization flags?

Wongster
Beginner
283 Views
I am working with a developer that asked the following question:

1. What are the recommended gcc optimization flags for a concurrency analysis? We generally I use -pg -p -O2, and sometimes -finline-functions. For a concurrency test, I can see the CPU usage for different libraries, but it doesn't have the symbol info for each function, etc. Those same options work fine for a hotspot analysis (though the -finline-functions removes some symbols, but not all).

2. Is there some way to annotate each test result so it can be remembered what was done to them? The test names are written with folder names like r00hs or things like that.

0 Kudos
1 Reply
Aubrey_W_
New Contributor I
283 Views
Hello,

I have discussed this some internally with the engineers. Those flags should be OK for the Intel Concurrency Checker. But for testing the mechanics of his parallelization, optimizaiton (-O2) should be disabled.

Here are some links to resources that may help him: http://software.intel.com/en-us/blogs/2011/06/30/new-to-parallel-programming/

Concurrency Improvement Center: http://software.intel.com/en-us/articles/concurrency-improvement-center/

Although, from your description it sounds like hemay beasking aboutIntel VTune Amplifier XE. I believe that is what "hotspot analysis" refers to. If that's what he is using, it's best if he post questions to the forum at http://software.intel.com/en-us/forums/intel-vtune-performance-analyzer/where the dedicated support engineers can assist him.

I hope that helps.

Best regards,

==
Aubrey W.
Intel Software Network Support
0 Kudos
Reply