See below ...  :-)

> -----Original Message-----
> From: Bernie [SMTP:[EMAIL PROTECTED]]
> Sent: Friday, 11 August 2000 19:29
> To:   [EMAIL PROTECTED]
> Subject:      RE: Comm's performance & FIFO settings & epppd(d)
> 
        ----- snip -----

> >     OK ... provided the "pppd.cfg" file is still supported ... (I found
> >this file
> >     was the easiest way to try different options :-).
> 
> There will be one Arachne version and one normal version, the "Arachne
> version" will use pppdrc.cfg
> 
        [da Silva, Joe]  

        Hmmm. If there are to be _any_ user-selectable options for "apppd"
        (I just made up a name here ;-), how will the user select them
without
        a "pppd.cfg" file (Arachne writes "pppdrc.cfg" itself, so the user
can't
        add options to "pppdrc.cfg" ...)?

> >     Sometimes it takes me up to 8 minutes to load a 20-25K web page from
> >     the ISP (38400bps). With the FIFO change, it just takes a few
> >seconds!!!
> 
> Ok, I'll see what I can do about letting the user configure it (or is this
> located in miniterm?)
> 
        [da Silva, Joe]  

        Good question. However, the "pppd" documentation says that it
        sets the FIFO trigger level to 8, so I guess that's where it's done
        (any setting by "miniterm" would get over-written by "pppd" ...).

        Best regards,
        Joe.

        PS. BTW, the FIFO trigger level of 8 is theoretically optimal
        (for reasons which should be obvious). Unfortunately, it is
        problematic with Arachne when there is a fast response from
        the host. I suspect interrupts are being disabled for 2-3ms
        (or perhaps 2-3ms longer than they should be), with the
        result that the effective depth of the FIFO (8 bytes) is not
        sufficient ...

Reply via email to