No, I doubt there are other bottlenecks involved. Although I still don't understand the use of a block size of 500K for the dd commands.
Anyway, using a SSD (OCZ Octane 128GB), I've also seen more than 60MB/s write speed, using SATA on PRO and/or Classic.
True, 48MB/s write speed is not bad at all and read speed is just fine.
But there is no denying, that the SATA performance of the old kernel sources is still better than that of the new sources.
Nobody says, though, that this can't change! I would be glad to see it happen. It just seems that we are not quite there,
yet.