[general] #598 Removed all usages of asmlib, because it is worse performing than...

[general] #598 Removed all usages of asmlib, because it is worse performing than std::memcpy on both Linux and Windows. Also removed the multicore copy because it actually seems to lower the performance. Also when memcpy performance is most crucial is when they are naturally parallelized like when having many channels and using channel_grid or many channel_producer instances