🚀 go-pugleaf

In article <47279@bbn.COM> slackey@BBN.COM (Stan Lackey) writes:
>In article <20764@usc.edu> vorbrueg@bufo.usc.edu (Jan Vorbrueggen) writes:
>
>>Eugene, ever seen a transputer? Overhead for receiving or sending a 
>>message is 19 cycles (630 ns for a 30 MHz part). The actual transfer
>>is done by a dedicated DMA machine at a maximum rate of 1.7 Mbyte/s
>>unidirectional or 2.4 MByte/s bidirectional.
>
>1) Although it has high? peak bandwidth, the latency is still there.
>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. This
use of `excess' parallelism is important precisely because it can hide
latency. This is one reason why the transputer not only incorporates
high performance communication links but also a hardware scheduler. The
use of excess parallelism to hide latency is not new, it was used in the
HEP.

>Please recall that even the hypercubes (well inmos, ncube and thinking
>machines anyway) do not use commodity microprocessors, but proprietary
>ones that do the extras in hardware that they need 
>

I'm not sure what you mean by commodity microprocessors. The Inmos
transputers ARE commodity microprocessors, they are feely available to
anyone who wants to buy them. 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))!





Roger Shepherd, INMOS Ltd   JANET:    roger@uk.co.inmos 
1000 Aztec West             UUCP:     ukc!inmos!roger or uunet!inmos-c!roger
Almondsbury                 INTERNET: roger@inmos.com
+44 454 616616              ROW:      roger@inmos.com OR roger@inmos.co.uk