🚀 go-pugleaf

In article <3vtkg5$d0n@inews.sc.intel.com>, mvilas@zws298.sc.intel.com (Mark Vilas) writes:
:In article <3vrf2o$ffq@meaddata.meaddata.com>, christw@meaddata.com (Christopher C. Wood) writes:
:Actually, there is a convention to describe bitness, while maybe not "generally agreed-upon",
:most processors adhere to it, including the 8088 and the 68000.  This is data path width.
:This is not to be confused with register size or external data bus width.  Both the
:8088 and the 68000 have 16 bit data paths and are thus classified as 16 bit processors.

Hmm... I always thought 8088 was only 8-bit, but I guess you know better (looking at
your emailaddress...). However, I was tought bitness is usually an ambitious description,
but with false 32-bit do we mean a processor with 32 bit internal architecture but using
only 16 (or less) bits for datatransfer. A true 32-bit processor uses 32 bits both internally
and for datatransfer. By this definition is 68000 a false 32-bit, while 68020 is a true 32-bit.
I'm a bit uncertain about intel's circuits, but you get the point. 

IMHO do I believe we use such terms to have a handy term for differentiate from newer and older
chips, not very often to compare 80x86 and 680x0 with each other. We could susally use the
words "newer" and "older", but it's kind of defacto standard to use "32-bit" and "16-bit"
instead, and usually is this enough. Do we want to compare performance is it much
better to use those test-diagrams that are ment for these purposes.

/Henrik