🚀 go-pugleaf

>>>>> On 31 May 91 02:06:26 GMT, prender@ohstpy.mps.ohio-state.edu said:

|> Hi.  Our research group needs to do real-time (or as close as possible) 
|> display
|> (grayscale-ish images) and 2D FFTs of 100+ x 100+ arrays of numbers, along
|> with
|> standard 2- or 3D rotations of the generated images (basically all just very
|> fast (compared to DOS on 386s) numbercrunching and graphics), and I've drawn
|> the short straw on 'finding a suitable system' for the job.  So far, OS/2
|> 2.0
|> (32 bit), SUN (and similar) workstations, and the NeXT Machine have all come
|> up for consideration.  I have myriad DOS (and related area) experience, but
|> nothing of consequence on the others, and so I'm looking for some practical
|> opinions of people actually USING SUNs (or actually, ANY of the above 
|> systems)
|> in a laboratory enviroment.

|> is VERY tempting to us, but I have NO experience AT ALL with even SEEING a
|> SUN run (I am a mainframe UnixBaby, though), and so hope to get some real
|> nitty-gritty on the subject.  Anything provided would be greatly 
|> appreciated.
|> As usual, EMAIL is sure to be read but postings to this group should be 
|> noticed
|> as well :-) ...

|> Thanks.

|> S Prendergast

	I administer a lab dealing with signal processing and speach
	recognition.  We work with HUGE databases of data and run
	large jobs dealing with these databases.  Thus number crunching
	is a big part of the usage of our computers.  Now, this is our
	set-up:

	4 SparcStation 2's  one with 48 MB and the others with 32MB.  
	Sun Slc (small - IMHO opinion, useless.. (not quite useless, but...))
	They all run off of a Sun 4/380 server.  
	We have about 35 users.

	Now, you might want to take a look at SGI workstations as well.  
	They are *the* number crunching personal workstation.  I have not
	worked with them all that much...Now, they do NOT yet run X which
	is a BIG strike against them. 

	My suggestion to you:  get a Sun server and some Sparc2's with
	color monitors.  If the data needs to be shared between everyone
	this is probably a good choice.

	Now, I also administer a smaller lab of about 6 people.  This lab
	has just started.  There are two Sparc2's.  The server is one of
	the Sparcs.  This is ok, but is simply not enough.  

	Hope this helps!!

jc

--
					-- James Cameron  (jc@raven.bu.edu)

Signal Processing and Interpretation Lab.  Boston, Mass  (617) 353-2879