2001-02-07 15:17:22

by CaT

[permalink] [raw]
Subject: suspecious ide hdparm results with 2.4.1 (and a minor capacity question)

well I got my ATA100 IBM HD and my Promise ATA100 card to go with it and
shoved them into my ultrafast VX MB (cough). I then ran hdparm
to see how fast the sucker could get and got this:

[02:02:24] root@gozer:/root>> free; hdparm -tT /dev/hde; free
total used free shared buffers cached
Mem: 94520 91760 2760 0 63272 6780
-/+ buffers/cache: 21708 72812
Swap: 266032 4 266028

/dev/hde:
Timing buffer-cache reads: 128 MB in 3.54 seconds = 36.16 MB/sec
Timing buffered disk reads: 64 MB in 3.13 seconds = 20.45 MB/sec
Hmm.. suspicious results: probably not enough free memory for a proper test.
total used free shared buffers cached
Mem: 94520 91828 2692 0 63608 6524
-/+ buffers/cache: 21696 72824
Swap: 266032 12 266020

Now... that suspicious results note makes me ponder that the results
aren't for real but... if they are, gawddamn. That's more then I hoped
for. 8)

Also, as an aside question, I want to get as much capacity out of this
HD as possible, back and windows compatability be damned. It's currently
in LBA mode (I believe) and that, to my knowledge, wastes the most space.
Is there anything I can do to get more of my HD back for use?

[02:02:41] root@gozer:/root>> hdparm -i /dev/hde

/dev/hde:

Model=IBM-DTLA-307045, FwRev=TX6OA50C, SerialNo=YMDYMT8X423
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=40
BuffType=3(DualPortCache), BuffSize=1916kB, MaxMultSect=16, MultSect=16
DblWordIO=no, OldPIO=2, DMA=yes, OldDMA=2
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=90069840
tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4
UDMA modes: mode0 mode1 mode2 mode3 mode4 *mode5

--
CaT ([email protected]) *** Jenna has joined the channel.
<cat> speaking of mental giants..
<Jenna> me, a giant, bullshit
<Jenna> And i'm not mental
- An IRC session, 20/12/2000


2001-02-07 17:44:56

by Andries E. Brouwer

[permalink] [raw]
Subject: Re: suspecious ide hdparm results with 2.4.1 (and a minor capacity question)

> It's currently in LBA mode (I believe) and that, to my knowledge,
> wastes the most space.

There are two entirely different things both called LBA.
Neither of them wastes any space.

2001-02-07 22:38:37

by CaT

[permalink] [raw]
Subject: Re: suspecious ide hdparm results with 2.4.1 (and a minor capacity question)

On Wed, Feb 07, 2001 at 06:44:34PM +0100, [email protected] wrote:
> > It's currently in LBA mode (I believe) and that, to my knowledge,
> > wastes the most space.
>
> There are two entirely different things both called LBA.
> Neither of them wastes any space.

You sure? I admit to not having much knowledge of this but BIOS always
reports different sized for LBA, NORMAL and LARGE. This is what I'm
referring to.

--
CaT ([email protected]) *** Jenna has joined the channel.
<cat> speaking of mental giants..
<Jenna> me, a giant, bullshit
<Jenna> And i'm not mental
- An IRC session, 20/12/2000

2001-02-08 00:22:35

by Andries E. Brouwer

[permalink] [raw]
Subject: Re: suspecious ide hdparm results with 2.4.1 (and a minor capacity question)

>> There are two entirely different things both called LBA.
>> Neither of them wastes any space.

> You sure?

Yes. One is the hardware disk access - all disk access is LBA
these days, certainly by Linux, but if the disk is small
one can also use CHS access, when it is old one has to use CHS.

The other thing called LBA is a translation scheme.
Translation means that the BIOS will use one set of parameters
when talking to the disk, and another when talking to DOS or so.
This just means that the same sector on the disk has different
names.

But when you partition your disk using DOS fdisk, it will want
to use an integral number of "cylinders" for each partition,
and for example with 8MB cylinders you may have the bad luck
that the rounding costs you almost 8MB. But here each of the
possible translations can be lucky or unlucky.
Under Linux there is no restriction that a partition has to
start on a cylinder boundary, but an integral number of blocks
is used, so for example with 1KB blocks and an odd number of
sectors this may cost you 512 bytes.
These days one sees various (rather ugly) patches floating around
just to get at this last sector. (But they are unnecessary, I think.)

Andries