>> This last part seems racy to me.  In principle, it is possible
>> that one thread terminated and simultaneously one new thread was
>> created, since the last time we stopped.  In this case, the thread
>> count would be the same, but the thread list still incorrect.
>>
>> Given the other test for single-stepping, is this particular
>> check even still resulting in a noticeable performance
>> difference?

Thanks Ulrich. In v3 version of the patch I have made sure we use only single step and no counts are used to save time. It is something I did not think of. 

>I would suggest making a patch with just the get_signaled_thread batch
>change first, since that one looks fairly safe and low hanging fruit.
>Then, measure again and then look for the next optimization.
>I would suggest using the "scoped_time_it" class to measure the
>execution time of various functions and scopes.  This object measures

Thanks Simon,

In 2^nd and 3^rd patch of the coming series I have done the same. Sending the patch with this data in the commit message. 

Have a nice day ahead.

Thanks and regards,

Aditya.