On  1 Dec 98 at 9:32, Luan Nguyen wrote:

> I have just installed a new UDMA Quantum FB 4.0Gb on my computer,
> CMOS detected that it was a 4.0Gb hdd and installed it in LBA mode.
>...
> yesterday, i have found that all the information abt my hdd in CMOS
> are completely different from the number provided on the disk
> (C=14### - i forgot :),H=9,S=63). i change the number in CMOS to
> the number provided be Quantum anh the computer still work, but not
> boot from hdd, it's like there is nothing on the hdd. I reenter the
> old number and it works ok again. The two information of the factor
> and the current informations both are detected by CMOS(the current
> is i LBA mode and the provided is in NORMAL mode or AUTO mode).

  The reason for using LBA is that the ordinary BIOS routines are
limited in the number of cylinders and sectors they support.  LBA
gets around this limitation, usually by successive steps of halving
the number of cylinders reported while doubling the number of heads
reported.  Internally, the process is simple to reverse to get the
drive "real" locations.
  So the numbers you see when LBA is selected are perfectly normal.

> I think i have to run fdisk again to use the hdd with new number.
> Since the hdd is now full of important informations, so it is wiser
> to make sure if the new number will improve the speed of the drive
> or not first. and that's what i need help on.

  You will need to re-FDISK the drive to change whether you use LBA
or not.  You're correct that far.

> Are there any problems using the hdd with informations differ from
> the informations provided by the factory (slower...)?

  I know of no significant performance penalty for using LBA.  Since
without LBA, the drive will only be accessible as about 500 MB, there
would have to be a really horrendous penalty to justify doing this to
a 4GB drive.

> every thing seemed ok, however it takes me abt 1 min to benchmark
> the drive with NU98, and the point was only 49 (my old Quantum
> 2.0Gb no UDMA, got more than 50 with a very short benchmark time) I
> don't know what was wrong since the drive work pretty good (except
> the problem in benchmark).

  Quantum's Fireball line offer reasonably good access times, but at
5400 rpm they are behind the "high performance" curve.  If the number
of *sectors* is the same, two 5400 rpm drives are likely to sustain
pretty similar throughput even if a larger cache helps one deal
better with short bursts of data.
  I don't know why it's taking longer to run the benchmark on the new
drive.  Does your (motherboard?) EIDE controller/chipset support
UDMA?

David G

                         PCBUILD's List Owner's:
                      Bob Wright<[log in to unmask]>
                        Drew Dunn<[log in to unmask]>