On Wed, Jan 07, 2004 at 11:30:12PM +0100, Hubert Feyrer wrote:
> In article <20040107011740.GA19050@olib.org> you wrote:
> > Thoughts?  Opinions?  Jeers?
> 
> I've played a bit with bochs too in the past few days, and have a 
> bochs-current package here, based on 20040106 sources. Drop me a mail if 
> you're interested.

If performance is much improved, I'd be interested.  (^&

On my 2GHz AMD64, I ran bochs with "pit: realtime" (or "pit: 1"?)
to get its clock synced with the real clock.  I was getting approx.
20MHz by NetBSD's installer boot messages.

Certain things (which I loosely correlate to emulated I/O) seem to
be much slower with "pit:".  Raw computation seems to run about the
same.  From that, my guess is that the clock gets resynced frequently
during I/O, but rarely (or never) during other instructions.

This makes some sense, since straight computation doesn't need the
time.


> I had trouble getting anything gfx'is going though.
> Both in-pkgsrc and bochs-current display a knoppix gfx logo when booting 

I got the logo.  After it finished uncompressing GNU/LINUX, the screen
updated.  (Without the "pit:" option, it *flashed* the update.  With
"pit:", I could see it clear and redraw from top-down.)  I gave up
waiting on Knoppix 3.1 at that point, so I don't know.

Also, I had an old partial attempt at installing BillOS 95 on my
virtual hard disk.  It put up the clouds-and-flag logo for BillOS
when I selected to boot from the virtual hard disk.  (But the disk
was essentially empty, so booting didn't do much once the logo
appeared.)


Just some random data points.


-- 
  "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/