[bionet.molbio.news] CSLG|COMM: From Alex Reisner

daemon@ig.UUCP (12/01/87)

From: Sunil Maulik <MAULIK@BIONET-20.ARPA>

		Computers and the Sequencing of Large Genomes

	
      Searching the database is not, however, a problem really suited to von 
Neumann machines.  At present the use of so-called 'fine-grained' general 
purpose parallel processor machines such as the Transputer (UK) or Connection 
Machine (US) appear to be best suited to such problems not only because of 
their very high speed in dealing with problems that can be posed 
appropriately but because of their flexibility and cheapness.  Adapting 
programmes such as FASTN or FASTP if optimum efficiency isn't an initial 
necessity, would be an almost trivial problem for the Connection Machine, a 
somewhat more troublesome exercise for the Transputer.  In addition the 
modularity of such machines allows them to be built in virtually any size. 
Moreover, the fact that these fine grained instruments are near first 
generation commercial efforts and are deliberately designed using non-exotic 
components clearly indicates that the aspects of simple searches and data 
storage can be dealt with quite easily.

-------