What performance improvement do you achieve? I’d be surprised if you
reach 1%, and even so that wouldn’t be worth the duplication of the
test, along with the decreased maintainability.
Surely there are parts of the code architecture that can be improved and
lead to better savings. This one looks a
One more point, which is "probably inlined" is not a very good guarantee,
since that is strictly an implementation option. Better to error on the
side of assuming it won't inline, and at least optimize the hot spots that
would be affect if inlining doesn't occur.
On Mon, Jan 21, 2013 at 12:12 PM,
I agree that when JIT inlines this it will not have much affect. However,
when profiling, this was identified as a hot spot.
On Mon, Jan 21, 2013 at 11:52 AM, Vincent Hennebert wrote:
> Hi Glenn,
>
> does that significantly improve performance? I would be really surprised
> if it would, given tha
Hi Glenn,
does that significantly improve performance? I would be really surprised
if it would, given that the method call is probably inlined by the JIT
compiler anyway. Do you have figures to share?
Thanks,
Vincent
On 18/01/13 22:13, gadams wrote:
> Author: gadams
> Date: Fri Jan 18 21:13:42