[1692] in linux-scsi channel archive
Re: TCP/IP over SCS
daemon@ATHENA.MIT.EDU (Cameron MacKinnon)
Wed Apr 9 15:24:52 1997
Date: Wed, 09 Apr 1997 15:11:57 -0400
From: Cameron MacKinnon <mackin@interlog.com>
To: yuri mironoff <yuri@rgti.com>
CC: linux-scsi@vger.rutgers.edu
yuri mironoff wrote:
>
> If you look at:
>
> http://www.bork.com/~scottr/sd/design/index.htm
>
> * Test Pan and Results
> (bottom of file)
>
> Max TCP bandwidth seems to be around 430 KB/sec. I may be missing
> something but isnt that SLOWER than 10BASE-T???
>
> Can someone possibly comment on the throughput - or lack thereof?
> Is it possible that one of the two cards used in the benchmark
> (the Adaptec 1520) slows it down that much?
I think the project above was meant to be a proof of concept rather than
a solution in itself. The 152x driver, as patched for target mode, uses
PIO for the data phase.
Using the above as a guide, I have started work on target mode for the
NCR 53C8xx driver.
It is possible to do target mode on the aic7xxx series, but driver gurus
have expressed doubts as to whether initiator and target mode sequencer
code will both fit into onboard RAM. I will probably investigate this
after the NCR driver is working.
I also hope to improve the 152x driver, but I haven't got my programming
manual from Adaptec yet. If someone wants to send me an am974, I'll look
at that, too. I don't know much about the 974, but realistically, if
you're looking for fast data transfer, a bus mastering PCI card with an
independent processor is the only way to go - preferably a family that
will go to 40 or 80MB/s.
