yao.qi at linaro.org
Wed Aug 18 05:00:03 BST 2010
Michael Hope wrote:
> I know that Dhrystone isn't a very good benchmark, but it's still interesting:
We can use 'opannotate --assembly' to get profiling results in details,
like how many events hit on a certain instruction. As showed here,
There are some other events in oprofile on ARMv7, that we can have a try
to get more information on time usage of memcpy() and strcpy().
I am not familiar with ARM arch and these hardware events on ARM, but I
think some of them should be useful for us.
> If we can get twice the performance out of strcpy() and memcpy() then
> the Dhrystone result should go up by almost 30 %. It would make a
> nice headline at least :)
More information about the linaro-toolchain