Bill Stevens wrote:
> I will appreciate any advice. The chemists seem to want the computer
> facility to be PowerMacs and not Unix workstations. I am not sure of myself
> here, but it seems we have options of either ftp-ing the data to the Macs
> and processing it there with some non_varian product or using the Macs to
> establish X-windows sessions on the spectrometer host computers and
> processing there.
I will offer the following information. I have a PowerMac 7100/80 running
eXodus 5.2 (PowerPC native). With this configuration running off a SPARC 10
host computer running VnmrX 5.1, most operations are actually FASTER than
they are when sitting at the SPARC 10 itself! When I first saw this I rubbed
my eyes and thought I must have done something wrong, but on reflection it
makes perfect sense. After all if the Sun alone is being used, it is functioning
as both the X client and the X server. With the PowerMac in the picture, it
is now the X server, so that the system as a whole functions as a
multi-processor
computer. True the two processors are connected via Ethernet,
not via some
internal bus (e.g., M-Bus), but given typical NMR operations (display a 1D
spectrum, do an expansion, etc.) the Ethernet is not the bottleneck (obviously,
YMMV depending on Ethernet traffic).
This observation is NOT true on earlier Macs, even on my Mac IIfx, but definitely
so on the PowerMacs. The difference I allude to above is small, but the
important thing is that it is not in the negative direction, that is, the Mac
users don't suffer any penalty.
Plus they get to have a Mac. :-) Which once they get out of your university
and into "industry" will at least give them some "good old days" to think back
on. ;-)
Steve Patt
Varian NMR Instruments