j
k
j a
j l
Did you measure the performance difference? Or checked if generated code is any different, especially for the second patch. I looks like forced inlining might have the same effect, if it's not inlined already automatically.
Back to the thread
Back to the list