This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Executable built using gcc runs slower than one built using CCS



Hi ,

I have noticed considerable difference in the performance of a StarterWare USB application when I building with StarterWare GCC tool chain and when built using the CCS compiler .

In order to set the context , I am using the usb_host_msc example which essentially supports reads and writes to devices like hard drives. My test setup is as follows

Platform : AM335x  - only data caches are on.

Code base : StarterWare release 7 package  modified to do RAW reads and writes. By RAM i mean bypassing the file system

Device : WD external hard drive 

Using GCC I obtanied a READ speed of 4MB/s and the same code base when compiler with CCS compiler returned 16MB/s.With GCC it is set optimization level 2 . In case 

anyone needs a list of compiler settings , i would gladly post them.

Initially it was reasoned that the memcopy or similar library supplied  functions would be affecting the performance of GCC executable. But this reason does not seem to hold as the write speed obtained using both the tool chains was very similar (close to 16 MB/s) .  

Has any body seen something similar ?

Thanks in advance

Vineeth