Re: Ram-Fast vs. Turbo IDE was (Re: RAM-Fast happy and sad)

[behrenss@Informatik.TU-Muenchen.DE (Soenke Behrens)]

In article <CBpywG.9FD@wave.scar.utoronto.ca>, 90taobri@wave.scar.utoronto.ca (TAO BRIAN T) writes:
|> In article <1993Aug11.120815.13753@Informatik.TU-Muenchen.DE>, Soenke Behrens writes...
|> > This is because the Turbo IDE uses the GS/OS caching mechanism, the RF
|> > does not.
|>     Not even the cache set by the RAM CDev?

Well, there is no other. Right. Not even that. Drew claimed that since they
had their own cache, using the GS/OS cache would slow things down. However,
their own cache (naturally) treats all data equal. But some data is more equal
than other (i.e.: directory blocks and the like), and the RF doesn't take
that into account.

Tell you what: The reason the Turbo IDE is faster than the RF is that it does
no caching of its own and relies on GS/OS for directory block caching. Sounds
weird? Not at all. More often than not, you don't really need the same data
twice in a row from the HD, except directory blocks. Having to write to a
cache and then to GS memory slows things down (only a little bit, granted).
Not keeping directory blocks in memory slows things down a lot.

Yours
Soenke

-- 
===========================================================================
More than silver, more than gold, I need   !    behrenss@informatik.
electricity. I need it for my dreams.      !    tu-muenchen.de
(RACTER)                                   !  FIDO: 2:246/13 Soenke Behrens