🚀 go-pugleaf

In article <2658@ganymede.inmos.co.uk> roger@inmos.co.uk (Roger Shepherd) writes:
>In article <47279@bbn.COM> slackey@BBN.COM (Stan Lackey) writes:
>>[in a hypercube]
>>the interconnect system waits for the processor to do an access, then
>>the processor waits for the interconnect to get the data.  Many
>>microseconds go by between the time the CPU needs dependent data and
>>it is usable.
>This is true BUT the time between a request being made and the data
>returning is NOT WASTED. It can be used to execute other processes.

The context of my posting was of getting high speedups on a single
application on a massively parallel computer in an easy-to-use,
general purpose way.  My statements concerned the use of commodity
micros without special architectural mechanisms intended to support
massively parallel processing.  By 'commodity' I meant true general
purpose micros produced with massive volumes and multiple sources.

>The communication system which is
>provided with every transputer is there because it is generically
>useful in multiprocessor system (and you might be surprised just how
>many electronic systems are multiprocessor - my PC has at least 3
>microprocessors in it (before I plug in my transputer card))!

If the three micros in your PC don't provide a speedup for Lotus by
running the Lotus code in parallel, it is outside the scope of this
discussion.  Which should probably be in comp.parallel anyway.

I'm not saying that the communication mechanism in a transputer is
'bad' or unnecessary; I mentioned it as a supportive example of
architectural extensions at the processor level for making massively
parallel systems more effective.
-Stan