tulip V.90f

Peter Green pcg@gospelcom.net
Tue Dec 22 14:37:36 1998

On Tue, 22 Dec 1998, Nils Rennebarth wrote:
> > Anyway, .90f solves the media detection problems we were having with the
> > 21143 chip as used on the API/Samsung Alpha motherboards.
> Yeah, and it solves my problems with the Macronix 98715 card too.
> That's really nice because these cards are quite cheap here, i.e. $24 each.

It still doesn't work well with the Lite-On Netgear cards (or the card is
really a piece of garbage). Here's a little test:

file = ungzipped 2.1.127 linux kernel (still tarred)  = SMC Etherpower II = SMC Etherpower II  = SMC Etherpower II = Netgear FA310TX w/ Lite-On ->
time scp file
file              |      50670 KB | 904.8 kB/s | ETA: 00:00:00 | 100%
real    0m55.569s
user    0m0.130s
sys     0m3.320s ->
time scp file
file              |      50670 KB | 1034.1 kB/s | ETA: 00:00:00 | 100%
real    0m48.888s
user    0m0.100s
sys     0m1.570s ->
time scp file
file              |      50670 KB | 830.7 kB/s | ETA: 00:00:00 | 100%
0.060u 1.800s 1:00.45 3.0%      0+0k 0+0io 12783pf+0w ->
time scp file
file              |      50670 KB | 164.0 kB/s | ETA: 00:00:00 | 100%
0.090u 0.780s 5:09.08 0.2%      0+0k 0+0io 800pf+0w

Am I just missing something, or is this device receiving at about 1/5 the
speed of the Etherpower II?! I mean, the lower system CPU numbers are
great, but not if I'm only getting 164 kBps over *ethernet*. (BTW, there
are no firewalls involved in any of these cases.) Is this just a matter of
time before the driver gets better performance than this? Or is this card
just fatally flawed? Or is my *test* flawed somehow?

> > Thank you very much Donald, once again you have been very helpful, and we
> > do appreciate it!
> > Merry Christmas and Happy New Year to you!
> Frome here too :-)

I agree wholeheartedly! Keep up the great work! (And death to
manufacturers that try to sneakily switch chips and call it a new
revision! DEATH!)

Peter Green
Gospel Communications Network, SysAdmin