🚀 go-pugleaf

In article <1992Sep6.130707.2409@newstand.syr.edu>, amichiel@rodan.acs.syr.edu (Allen J Michielsen) wrote:
}In article <Bu34nA.3FB.2@cs.cmu.edu> ralf+@cs.cmu.edu (Ralf Brown) writes:
}> IDE won't max out for a while...if the drives themselves keep getting faster
}
}   I don't see how the drive can get much faster.  Good access time are already
}around 10 mS, and we are talking physics.  Mechanical parts, moving...

Simple: create a drive with 16 data surfaces and place each bit of a
16-bit word on a different surface.  Voila, 32+ MB/s throughput....  Or
interleave the bytes across as many surfaces as you want on a
byte-by-byte basis.  Seek times can be reduced by using multiple heads
per surface.  These techniques were in use 30 years ago on mainframes.

}>The more immediate limit to IDE (and ESDI) is likely to be the size of the
}>drive.  If (as I suspect), the WD-1003 only permits 16 heads, then the
}>limit on IDE is 504 MB (1024 cylinders * 16 heads * 63 sectors/track); even
}
}The WD-1003 spec does max out at 16 heads, but that isn't something cast in
}stone.  A new backwards compatable standard shoud be in the near future...
}The 1024 cylinder limit is a dos limit, ide and esdi controllers don't
}have that same problem because the actual drive geometry is hidden
}(same as scsi).  This makes the standard max size up around
}1024 cyl x 1024 heads x 1024 sectors 

Yeah, but the geometry translation is based on the limits of the BIOS, which
are 1024 cyl x 64 heads x 63 sectors/track--a little less than 4G sectors
per BIOS-recognized drive.  To get more on a physical drive, you'll have to
split it into multiple drives before the BIOS sees it.

-- 
Internet: RALF+@CS.CMU.EDU   |The University would disclaim this if it knew...
FIDO: Ralf Brown 1:129/26.1  |
BIT: RALF%CS.CMU.EDU@CARNEGIE|"Success has a simple formula: do your best,
AT&Tnet: (412)268-3053 school| and people may like it."   -- Sam Ewing