On Friday, 11-25-1994  Raetus Egli wrote to Lynn Nash about
"Colorados/2\Warp" as follows:

RE> I was very skeptical when I read your post.  Not that I don't
RE> respect your opinions and knowledge but because I ran extensive
RE> tests on IDE and SCSI busmasters about 2 years ago.  The difference
RE> then between HPFS16 and HPFS386 was significant.
RE> 
RE> However, after I read your post I did run a few tests using WARP
RE> on an  IDE machine and have to agree with you.  There is no
RE> discernible difference between the two.
RE> 
RE> Can you think of some obvious reasons why this would be the case?

Hi Raetus,

Putting HPFS386 on a PIO adapter is like putting lipstick on a pig -
no matter how much you try to pretty it up it is still intrinsically
ugly.

The busmaster approach allows multiple SCSI transfers to be chained
into a single physical transfer while the CPU goes off to do other
things. Using PIO the CPU sits there polling the adapter on a
byte-by-byte or word-by-word basis; no other significant amount of CPU
work gets done. And HPFS386 exploits the busmastering architecture more
fully than does "vanilla" HPFS.

Regards

Dave

 * KWQ/2 1.2g * My strength is as the strength of ten because my code is pure.

--- Maximus/2 2.02