--- Henning Meier-Geinitz <[email protected]> wrote: > > A glance at the kernel source (g_NCR5380.c) tells a lot : > > USLEEP_POLL = 1 > > USLEEP_SLEEP = 20 > > USLEEP_WAITLONG = 500 !!!! > > > > I'll try to play with them a bit. > > Be carefull. I had kenel lockups sometimes while playing with these > values.
I tried to play with these, first of all using the suggestion for Mustek from the link you told me, then I played around with other values as well, also I commeted out USLEEP. The results yielded a small increase in performance, but far from satisfactory. Now the scanner sensor moves continuosly for some X cm dependent on USLEEP_POLL and USLEEP_SLEEP, then waits a certain time, longer (1.5-3 times longer) then the time it took for the motion. With the settings suggested for Mustek, it moves about 5cm first, waits, the moves about 2 cm, waits again and repeat in 2cm increments. With USLEEP commented, the behaviour is similar. Playing around with these values I could only increse or decrease a little the distance of one continous move, with a similar increase of the waiting time. The machine was never dedicated completely to the scanner, not even with USLEEP commented. I started xscanimage from a remote X server and a ping in another window. It never lost packets. top command showed on average >80% idle Overall, there is a slight improvement of speed (at least for the preview), but still more the 3-4 times then under windows. Looks like I'm stuck with M$Win for using this scanner. Thanks for the suggestions. ===== ing. Andrei Boros Centrul pt. Tehnologia Informatiei Societatea Romana de Radiodifuziune __________________________________________________ Do You Yahoo!? Great stuff seeking new owners in Yahoo! Auctions! http://auctions.yahoo.com
