Subject: Re: FreeBSD Bus DMA
To: None <perry@piermont.com>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: tech-kern
Date: 06/11/1998 22:21:22
On Thu, 11 Jun 1998 23:45:35 -0400
"Perry E. Metzger" <perry@piermont.com> wrote:
> I don't care much about the VAX. What I care about is next year's
> system that requires flexibility in ways that having the VAX work
> taught us to accomodate. As it stands, we work on virtually everything
> that can be thrown about us -- and these days, building a new port is
> becoming a cookie cutter operation, with massive code sharing.
Running on insanely slow hardware also motivates you to improve the
performance of your kernel :-)
I recall that Ragge (the VAX port gooroo) posted that his uVAX-II ran
/etc/rc something like 10 or 15 seconds faster w/ UVM :-)
Also, I recall that the Arm32 guys, Mark and Neil (who John indirectly
referred to earlier, in case you didn't know) observed quite a noticeable
performance boost w/ UVM, and after making some changes to their pmap
suggested by Chris Demetriou and Chuck Cranor (IIRC - I was put in the loop
on the mail exchange, and I'm pretty sure he posted to port-arm32 about it),
got additional performance gains near the _50%_ range!
Can't wait to run that code on my Shark :-)
Jason R. Thorpe thorpej@nas.nasa.gov
NASA Ames Research Center Home: +1 408 866 1912
NAS: M/S 258-5 Work: +1 650 604 0935
Moffett Field, CA 94035 Pager: +1 650 428 6939