[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: kernels and b0xen



On Thu, 23 Jan 2003 simon@nuit.ca wrote:
> On Wed, Jan 22, 2003 at 22:26:00 +0100, Michel Lanners composed:
> > On  22 Jan, this message from simon@nuit.ca echoed through cyberspace:
> > 
> > >> - check your hard disk speed: hdparm -tT /dev/<disk with kernel source>
> > >>   (is non-destructive; post the result to compare)
> > > 
> > > 1.  Timing buffer-cache reads:   128 MB in  3.73 seconds = 34.32 MB/sec
> > >  Timing buffered disk reads:  64 MB in 18.10 seconds =  3.54 MB/sec
> > > 
> > > 2. /dev/sda:
> > >  Timing buffer-cache reads:   128 MB in  3.68 seconds = 34.78 MB/sec
> > >   Timing buffered disk reads:  64 MB in 12.71 seconds =  5.04 MB/sec
> > 
> > 34 MB/sec for buffered reads s essentially a measure of memory
> > bandwidth. And 34 MB/sec seems low for a 7500-class machine. I get 45
> > MB/sec (OK, on an interleaved memory config).
> 
> that may be because of the smallish amount of RAM on this thing;
> interleaved i was getting pretty much the same ratings. it may have to
> do with the RAM sticks' actual ms speed rating - i have 4MB sticks @ 70
> ms IIRC, and some 8MB sticks @ 60 ms.

us (microseconds), I guess?

BTW, I get 45 MB/s for the buffer-cache on my 66 MHz SDRAM.

> > 3.54 MB/sec for disk is low; 5.04 is about acceptable for an older SCSI
> > disk. Unfortunately Apple fitted those systems with _horrible_ disks
> > (like the Quantum Fireball TM's).
> 
> this isn't the default drive. it's a seagate ST39173N (can't remember
> the name ATM). which is rated @ 7200 rpm. i think that's the bus, more
> than the actual disc.

And the MESH. I don't get much more from my UW-SCSI disk if I connect it to the
MESH instead of the Sym53c875.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds



Reply to: