LLL (HLL too?): biased towards today's technology

Dr. J. Van Sckalkwyk (external) SCHALKW@odie.ee.wits.ac.za
Sun, 15 Jan 1995 22:22:41 SAT

Dear Chris

>       It seems that much of our recent discussion about the LLL seems to 
> have assumed that it will be a linear language, fit to run on the 
> [snip] ...
> working at the same time.  I can see computers that are not based on CPUs 
> and centralized clocks, but are a collection of small, extremely 
> specialized units, each syncronizing [sic] and communicating with 
> only a small subset of the others.  I can see linear memory being 
replaced by the [snip]...
>       Any thoughts on this?

Agreed, in principle. Parallel processes that can communicate should 
be part of the architecture (regardless of whether the machine we are 
working on has one or several processors to do the work). Some sort 
of fork/join or whatever should be implemented at a fairly low level, 
and be simulated on a single processor machine, or used to distribute 
the load on a multiprocessor machine! Actual inter-process 
communication can be fairly tricky, when you look at it! It may 
become a real bummer when you're working on a connection machine or 
something, I would imagine!!

Bye, JVS. <