🚀 go-pugleaf

In article <45b7rk$qt4@aadt.sdt.com>, hawkins@PROBLEM_WITH_INEWS_GATEWAY_FILE (Tony Hawkins) writes:

|> >   I guess that what all this proves is that the phrase "64-bit
|> > machine" is nonsensical. 64-Bit by what definition? External or
|> > internal bus, data or address, register size? Without this
|> > information, "64-bit machine" tells you _nothing_.

Common usage has long been that if all you say is "I have an N-bit CPU",
you meant N = programmer-visible size of integer registers ...
	NOT FP register size (often different, like 64- or 80 for 32-bit CPU)
	NOT bus-size, of which there are often simultanously different sizes,
		even based off the same CPU core.
	NOT virtual-address size, which is usually <= N, although sometimes
		more with segmentation, sort of.
	NOT physical-address size, which has often been != N.

BYTE, Sept 1991, discussion + Table 1, p 136 + goes through all that.
(I'm sorry to keep pointing back to an article I wrote; smacks a bit too much
of self-advertising ... but *REALLY*, most of the topics discussed in this
discussion thread were gone through in more detail, in that article, 4 years
ago, and with charts and graphs difficult to put on the net:
	What does it mean to call a chip 64-bit?
	Year, size-called, integer-register-size,virtual-address-size,
		physical address size, data bus size for ~20 CPUs.
	Applications that get performance from bigger integers
	Bigger-addressing and why ... and why not:
		NOT to address 2**64 BYTEs of DRAM, mile-square 300-feet high.
	History & errors of addressing: mainframe, minicomputer, microprocessor
	Needs for big computers
	32-bit pressure in 1993/1994
		Graph of maximum memories in Mips & Sun systems, for example
	Why so much memory, and why 64, rather than (for example) 48-bitters
		What applications care; disk-drive issues.
		Most applications will stay 32-bit.
	On the desktop?

Four years later, I wouldn't change much, except to say a few more words to
emphasize even more that file-mapping and sparse addressing are useful even on
smaller-memory machines.  I gave a couple examples, but more would have
been useful.


|> I think the original assumption was:
|> 
|> 	data bus size = register size = address size
|> 
|> Of course this isn't always true --- but its not a bad assumption, for 
|> arguments sake!
Well, actually, of the 20 machines listed in BYTE, this was clearly untrue for
a majority of the CPUs...


-john mashey    DISCLAIMER: <generic disclaimer, I speak for me only, etc>
UUCP:    mash@sgi.com 
DDD:    415-390-3090	FAX: 415-967-8496
USPS:   Silicon Graphics 6L-005, 2011 N. Shoreline Blvd, Mountain View, CA 94039-7311