On Thu, May 25, 2006 at 08:49:47AM -0400, Brian Cuttler wrote:
> Jon,
>
> On Wed, May 24, 2006 at 07:25:46PM -0400, Jon LaBadie wrote:
> > On Wed, May 24, 2006 at 02:54:34PM -0400, Brian Cuttler wrote:
> > > The -e made a huge difference in runtime.
> > > # date ; amtapetype -e 400g -f /dev/rmt/3n ; date
> > > Wed May 24 09:23:58 EDT 2006
> > > Writing 2048 Mbyte compresseable data: 33 sec
> > > Writing 2048 Mbyte uncompresseable data: 33 sec
> > > Estimated time to write 2 * 409600 Mbyte: 912 sec = 0 h 15 min
> > > wrote 12386304 32Kb blocks in 189 files in 6539 seconds (short write)
> > > define tapetype unknown-tapetype {
> > > comment "just produced by tapetype prog (hardware compression off)"
> > > length 386048 mbytes
> > > filemark 0 kbytes
> > > speed 61853 kps
> > > }
> > >
> > > I really should run it again with HW compresson ON.
> >
> > If like my ultrium 1, the block count will be exactly the same,
> > but the speed will be slower.
>
> You are right, largely the same, run time of 3h 38 min and the speed
> was a little lower, but only by about 1%, but I don't know enough
> about the amtapetype and maybe the pattern of the data written has
> a lot to do with that.
>
> > date ; amtapetype -e 400g -f /dev/rmt/3un ; date
> Wed May 24 14:55:53 EDT 2006
> Writing 2048 Mbyte compresseable data: 34 sec
> Writing 2048 Mbyte uncompresseable data: 33 sec
> Estimated time to write 2 * 409600 Mbyte: 912 sec = 0 h 15 min
> wrote 12320768 32Kb blocks in 94 files in 6320 seconds (short write)
> wrote 12386304 32Kb blocks in 189 files in 6487 seconds (short write)
> define tapetype unknown-tapetype {
> comment "just produced by tapetype prog (hardware compression off)"
> length 386048 mbytes
> filemark 0 kbytes
> speed 61742 kps
> }
Well, actually a little different. I got my Ultrium 1 from ebay and
like a kid with a new toy ran several tapetypes on it. Three different
physical tapes (two brands), one two times, and with 3 different block
sizes, and a couple with and without compression on.
The thing that flabbergasted me was the absolute consistancy of the
total blocks written (adjusted for block size).
Two of your report lines were like mine:
> > > wrote 12386304 32Kb blocks in 189 files in 6539 seconds (short write)
> wrote 12386304 32Kb blocks in 189 files in 6487 seconds (short write)
Oh, I did some with a different scsi controller too as the first one I
used was too low performance. Same block count though.
--
Jon H. LaBadie jon AT jgcomp DOT com
JG Computing
4455 Province Line Road (609) 252-0159
Princeton, NJ 08540-4322 (609) 683-7220 (fax)
|
