are kernels <<<2*N,32>>> and <<<N,64>>> the same for speed ?

are kernels <<<2*N,32>>> and <<<N,64>>> the same for speed ?

Typically, they are very close. Sometimes, they will be ~10-15% different. See http://forums.nvidia.com/index.php?s=&…st&p=500384 for more extensive posts on this subject.