Re: hdparm and DMA

[Emil Pedersen <emilp@update.uu.se>]

[...]

> # hdparm /dev/hda
> /dev/hda:
>  multcount    =  0 (off)
>  I/O support  =  0 (default 16-bit)
>  unmaskirq    =  0 (off)
>  using_dma    =  0 (off)
>  keepsettings =  0 (off)
>  nowerr       =  0 (off)
>  readonly     =  0 (off)
>  readahead    =  8 (on)
>  geometry     = 77545/16/63, sectors = 78165360, start = 0
>  busstate     =  1 (on)
> 
> # hdparm -tT /dev/hda
> /dev/hda:
>  Timing buffer-cache reads:   128 MB in  0.46 seconds =278.26 MB/sec
>  Timing buffered disk reads:  64 MB in 16.12 seconds =  3.97 MB/sec

THIS is ok, without dma performance is lousy, but I really don't
understand why you can't turn it on (or get better perf when you do)!



> Thus all the same as before whether turned on or off.  I also tried
> setting MultiSect to max of 16 (which hdparm -i told me it could do),
> but this didn't change things either.
> 
> So are the read times as expected for an IDE ATA133, 7200RPM, 40GB drive?

Certainly not.  Bellow is what I got from my IBM 7200 rpm disk, (using
UDMA66):

/dev/hda:
 Timing buffer-cache reads:   128 MB in  0.81 seconds =158.02 MB/sec
 Timing buffered disk reads:  64 MB in  2.03 seconds = 31.53 MB/sec


At home I got Samsung 5400 rpm disks, they do about 20MB/s real single
stream write, tested with:
  time { dd if=/dev/zero of=zeroes bs=1M count=1000; sync; }


> This is one of the faster IDE drives now isn't it?

Don't know, but even the slowest you could by today should give a lot
better performance than 4 MB/s...


I'm out of ideas.  As far as my experience goes, hdparm should work. 
Try upgrade something, hdparm, kernel, anything.
  Btw, you do use properly working cables, right (and don't se any
errors in syslog or messages)?

Regards,
	Emil