Posted by Ilya Zakharevich on 01/18/08 23:28

[A complimentary Cc of this posting was sent to
John Navas
<spamfilter1@navasgroup.com>], who wrote in article <eu22p3lk7aqqdu36a4k3pumqqd8iili4ar@4ax.com>:
> It's probably more the issue of Firewire being designed for high-speed
> independent transfers, whereas USB is controlled by host polling. See
> my prior post to this thread.

The conjecture I've heard on why USB2 is so much slower (in practice)
than Firewire 400 is that current drivers/adapters do not use maximal
possible buffer/window sizes.

[AFAIK, maximal practically measured transfer rate of USB2 is
33MB/sec. This (was claimed to) coincides with the maximal
theoretical transfer rate with some particlar small size of
buffer/window.]

E.g, see
http://groups.google.com/group/comp.os.os2.setup.storage/msg/0d761a99fc3c6a0d
and the thread.

Hope this helps,
Ilya

[Back to original message]