On 27 Mar 2000, at 3:06, der Mouse wrote:


> > that statement needs to be cleared!  It depends upon what one wants
> > to do with the computer.  As ide drives still need the processing
> > power of the CPU to do transferrs
> 
> Is this actually true?  I thought modern IDE interfaces were just as
> capable of "start it and forget it" operation as any SCSI interface.
> (And don't forget that SCSI interfaces demanding PIO exist, though
> quite possibly not for i386 machines.)
> 

Depending upon what you mean with "modern" that could be true (I 
contacted Promise and HPT this weekend to get such information). 
To my experience with up to Intels BX chipset ide controllers, cpu 
usage and PCI bus activity influences ide transfer *much* stronger 
than scsi transfers, not to speak about the ability to copy data from 
one device to another without influencing the host cpu and bus too 
much. I have not too much experience on non Intel machines, but 
even the lowcost sparcs that use ide disks show weaker 
performance than when equipped with SCSI hardware (at least 
according to the tests I read). 


> > there are many environments where SCSI is still more performant.
> 
> Oh, certainly.  For example, I have many machines for which IDE drives
> are totally useless. :-)
> 
> In fact, quite aside from all the reasons to prefer one or the other
> in isolation, I would always go with SCSI because then I can use the
> drives anywhere; with IDE I'm much more restricted (Intel, some
> Alphas, a few of the more recent Suns, anything else?).  Of course,
> whether this is a valid argument for *you* only you can tell.
> 

Not to talk avout the expandability, looking on my desktop, I see five 
external devices. But when it comes to a disk only subsystem, and 
the newer HPT or Promise controllers had a host controller CPU, not 
only for their cost were ide drives shurely worth to be considered. 
For Intel platforms of course.

mike