🚀 go-pugleaf

Rowan D. Hughes wrote:
> 
> On Thu, 25 Sep 1997 15:46:25 +1000, Brian Miller <bmiller@telstra.com.au> wrote:
> >Alan Palmer wrote:
> 
> The speed of ISA ether cards is very dependent on the type of CPU, since
> ISA cards generate a lot of interrupts and have fairly slow I/O. So we
> see above the pentiums do much better than a 486DX2. (Bus masters, such as
> the Tulip, depend less on the CPU speed.) These comments are also true
> for PLIP using the // port. 486's get up to 20kb/s while pentiums get
> over 50kb/s.  
Referring back to the original post it's on a DX2/50.

A DX/2 50 means the raw external bus speed is about 25 MHz,  if you look in
your BIOS you should be able to specify the ISA bus speed as a division
of BUSCLK or CPUCLK.  Although this should not make that much of a 
difference since the ISA is slow anyway.  

On a clear ethernet you would expect to peak (AT best) around 1 MB, but that
is pushing the limit of the medium, (you can't get 100% utilisation).  Then
you have to take into account all the headers (and trailers) on the ethernet.
Then there will also be collisions (aren't TCP ACKs great).

I would expect around 400-600 kB/s in real terms.

> Also, my guess is that 10base2 is a good bit slower
> than 10baseT.  

There shouldn't be much of a difference between 10base2 and 10baseT,
they are both shared medium, operating at 10 Mb/s.  If you have
a switched hub with the 10baseT you might find it quicker, but that
is another story.  For those who are interested the connection to the
medium is mainly electrical, no different processing is done on either.


> The only real performance tuning is done by changing the MTU size
> with ifconfig. Generally it's best between 1500 and 2000 (max).
> There are other tuning parameters for NFS, etc.

Also if you do an ifconfig on the Linux box and see if there are any
errors in the packets with /sbin/ifconfig.  That would indicate something
screwy.

Matt