>1. Hm, seems that different people get different results. Not uncommon thing with Windows hosts... Probably things are not that bad.

seeing the feedback in this thread, it looks like all people
who have done speed tests get pretty bad results. there was one reporting telnetd.device got some higher cps,
but telnet was never intended as a binary transfer thing, and he didnt say how much he got and on what hardware.

Mosnet seems faster in responsetime and requesting all the urls for images on a www site seems faster,this is very good.

but in filetransfer on a local network it seems to perform 10-15 times slower that the hardware allows.

in my case on a pc and pegasos 50cm from eachother
both connected to a switch with around 1 meters of
cable, both have 100mbit nics, i get around 500k/s
copying from pc to morphos
(this is the same morphos_powerup.iso 27megs i've
been always speedtesting eztcp with)

i think it was v0.12 that copyed the file over in 35 secs,
thats around 8-900k/s but every version after needs
60-65 secs.

rebooting and launching miamidx on the same setup, i copy the file with 3000-3200k/s.

so something must be limiting the speed inside mosnet,
as the hardware,cables, windows host performs so'
much better with emulated 68k stack from the last
century.

Have you done any speed tests yourself ? i find it wierd
to start porting and enhancing this stack without
knowing its performance is reached first, isnt it
harder to track down its slowness after you have implemented feature after feature after feature ?

i doubt ferrari builders start polishing chrome and other eyecandy before they know the frame and engine and
everything technical having to do with its performance is 100% tuned first.

im not trying to be too negative here, but this is a very
important part of software for Morphos (a tcp stack)
and the bounty amount reflects this too, and
it would be a shame if its performance is this poor in the final build.