Hi
>Overall performance degregation is embrassing these days.  My plan
>under desk is, 1) /sys/dev/rasops/ console screen rendering should be
>revisited.  Partial word load/store brings read-modify-write memory
>traffic.  That hurts speeds visibly.  2) TLB miss handler should be
>retooled.  Indeed, pmap.c will be replaced with the new one.
>
But this doesn't address  what changed between 1.4.1 and 1.5 to make
 all these other things issues again. I've never had a problem of this
 sort with 1.4.1.  Until I moved my http service from the 5000/33 I'm
 talking about to my 240 it (the 33) too ran NetBSD 1.4.1 and whilst I
 noticed an improvement after the move to the 240, it was nowhere near 
the difference I'm noticing now.
So the question still stands, what changed ?
However, I am glad to see at least one major influential person in NetBSD/pmax acknowledge useability is an issue.

Charles


------------------------------------------------------------
--== Sent via Deja.com ==--
http://www.deja.com/