Autor: Mariusz Kowalczyk (mkov_at_cc.uni.torun.pl)
Data: Thu 06 Mar 1997 - 00:00:17 MET
In pl.comp.pecet Zbigniew Olearczyk <ZBOLE_at_star.kul.stalwol.pl> wrote:
: On 5 Mar 97 at 0:48, Robert Slaski wrote:
: > skutek byla taki, ze transfer poprawil mu sie z 1.2MB/s (wtedy to bylo
: > fiuuu...) na jakies niecale 300kB/s (zgroza, nawet wtedy).
:
: Nalezalo po preformacie ustawic przeplot i mialbys ponownie 1.2MB/s.
: > Reasumujac: dyskow twardych IDE sie NIE formatuje, maja one fabrycznie
: > naniesiona strukture sciezek i kazda taka operacja, jesli nie go
: > uszkodzi, to przynajmniej mocno go rozkalibruje.
: Nie jest to prawda
Podobnie jak i te cuda o ustawianiu przeplotu dla dyskow IDE :-)
Jest taki fajny dokument, nazywa sie E-IDE/FAST ATA/ATA-2 FAQ.
Do znalezienia pod http://www.wi.leidenuniv.nl/ata/atafq.html
14 rozdzial, pozycja 14.1 prowadzi nas do Facts and Fiction by Hale
Landis. Krotki cytat z tego zrodla:
"The original ATA specification (now at rev 4.0c, also known as
ATA-1) documents the "full" function Format Track command but
leaves it to the drive vendor to decide what a drive will really
do. It recommends a minimum action of writing binary zero into
the data field of each sector formatted. The ATA-2 specification
says that the function of the command is "vendor specific" -- it
doesn't even recommend the minimum action of writing binary zero
data -- a major step towards (finally) making the command
obsolete."
IMHO low-level format dyskow IDE moze byc albo bardzo niebezpieczny albo
tez moze byc wielka strata czasu. Zalezy to od daty produkcji iproducenta
dysku. Generalnie im nowszy dysk, tym strata czasu wieksza.
Wracajac jednak do przeplotu:
"MFM/RLL/ESDI and some older ATA drives use the same controller
command code, known as Format Track, to do the low-level format
of a track. This command is issued once for each track on the
hard disk. This command writes the inter sector gaps, the sector
ID fields and the sector data fields. Each sector on a track has
an ID -- an 8-bit binary number usually starting at 1. The order
is which the sector IDs are written determines the interleave.
If the sector IDs are written as 1, 2, 3, ..., n, this is 1-to-1
interleave. When written as 1, n, 2, n+1, 3, ..., n-1, you have
2-to-1 interleave. MFM controllers usually used 2-to-1 or 3-to-1
interleave with 17 sectors per track."
[...]
"Likewise, because most
drives are now zone recorded (they have different number of
sectors per track at different locations on the media), the inter
sector gaps, sector ID fields and sector data fields are also
written at the factory and can not be recreated later."
Amen, Enter.
Pozdrawiam,
kovi
-- <mkov_at_cc.uni.torun.pl> "Qui bibit, dormit, qui dormit, non peccat, sanctus est."
To archiwum zostało wygenerowane przez hypermail 2.1.7 : Tue 18 May 2004 - 15:57:45 MET DST