🚀 go-pugleaf

Alan Hewat (hewat@ill.fr) wrote:
: gusm@dmp.csiro.au (Gus Mumme) writes:

: > There seems to be a general agreement on the following :-
: > Indigo 2, 48-64 Meg RAM, "extreme" graphics,
: >

: The machine Bob quoted at $15K was apparently an Indy 4400SC, which is 
: indeed a good machine for inorganic crystallography.  But it is much 
: less expensive than an Indigo**2 extreme, for which you now conclude 
: $15-28K is a 'reasonable price'.  

: You have apparently been confused by the various advice you have 
: received.  Go and have a look at the machines running the programs
: that you are interested in, and decide yourself how much you want to
: pay.  You may conclude that you can start with the cheapest Indy 
: 4600PC model at $5-10K, which will compare advantageously with a 
: 'near top of the range PC'.

Extreme graphics may be overkill.  Your best bet is to sit
down at such a beastie and try the software you want to use.

We have a range of machines here from R3K Indigoes to an Indigo2 Extreme.

The comment I've heard is that there is very little difference 
between XZ and Extreme graphics for most applications.  The 
four-fold increase in Geometry Engines really only makes a
significant difference if you are using textures in your images.
Most crystallographic software does not; they rely on vectors and
perhaps triangles.  (Grasp and Setor are the only exceptions to
the rule that I'm aware of and I'm not sure how much they exploit
the advanced GE features.)  We do fitting on all our machines.
For small macromolecular work, an Indy with the large  monitor
would suffice (but don't try a better machine or you won't
go back).  By small, I mean around 100 residues.  For a larger
structure (~1700 residues) you will want at least XZ graphics
and either an R4K or R4400 processor.  The TFP will likely
be overkill unless you intend to also use the machine for
refinement runs.

Also keep in mind that you will need dials and buttons (well...
we need them for our software/users) and may want to consider
the CrystalEyes stereo eyeware.

For the record, we use: Alberta/Caltech TOM and MMS (UCSD) primarily.
(We are starting to use O more, but inertia is a fact of nature
for end-users as well as "objects in motion" :-).)

Ron Meleshko
MRC Group in Protein Structure and Function
Department of Biochemistry
University of Alberta
Edmonton, Alberta
Canada