🚀 go-pugleaf

In article <1993Jan26.073432.16561@ntuix.ntu.ac.sg> eoahmad@ntuix.ntu.ac.sg (Othman Ahmad) writes:
>Why do say we are stupid when the ATBus version is surely faster than the
>SCSI version?

I'd be interested in hearing why you think this is so.  It would seem
to me that the SCSI CD-ROM drive would be faster than a SonyBUS drive,
if indeed there is a performance difference (which I doubt, see
below).  Why?  Well, the controller/host adapter provided with the
cheap SonyBUS drives is usually an 8-bit unit that is engineered with
price in mind, not performance -- since the SonyBUS units are usually
the cheaper, lower-end devices.  With a SCSI device, the manufacturer
doesn't have to include a host adapter in the total development and
manufacturing budget, and so can dedicate a higher percentage of the
total budget to the drive itself.  In addition, SCSI host adapters (at
least the ones I'd use in *my* system) are usually 16-bit or 32-bit
cards that do bus-mastering DMA.

However, I doubt there's going to be any noticable performance
difference at all.  Keep in mind that with a CD-ROM drive, the
bottleneck is getting the data off the CD, not getting the data from
the CD-ROM drive to the host.  The CD<->drive transfer rate is going
to be 150K/s (or 300K/s, if you have one of the fancy double-speed
drives), no matter what you do with the host adapter; 150K/s is much
lower than the maximum transfer rate of even an 8-bit card using
programmed I/O.

-- 
Marc Unangst, N8VRH         | "Of course, in order to understand this you
mju@mudos.ann-arbor.mi.us   |  have to remember that the nucleus of the atom
                            |  is squishy."
                            |    -W. Scheider, from a Physics lecture