The results I get are stable, ie 0 other running processes on an otherwise
quiescent system. I consistently get 120MB/s from the buffer-cache.
I now know this is due to the smaller page-cache index (or something
similar).
BTW: I get 20.25MB/s from my IDE Maxtor DiamondMax+ 5120 10GB harddisk. I
think that's quite fast, but there's lots of new faster SCSI disks
emerging.....
My processor is a Celeron 300A oc'd to 464Mhz, with a memory bus speed of
103Mhz. Since the level 2 cache of the celeron runs at full core speed (ie
464Mhz) memory performance rocks 8o) . Ideal for small (efficient)
algorithms. My memory is a unified (2 banks on 1 DIMM) 128MB 8ns DIMM,
working at CAS, RAS to CAS and RAS of 2. It's all tuned for speed, etc.
Email me for more info....
Dan
[snip]
> Indeed, I was wondering if this was reproducible, i.e. the exact same
> everything in single-user mode with many runs averaged together.
>
> But what I REALLY want to know is where you found such an excellent memory
> I/O subsystem. Even with cas/2 PC100 SDRAM I can't seem to get better
than
> 54 MB/sec.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/