Re: What is a word ?


toby wrote:
> robertwessel2@xxxxxxxxx wrote:
> > toby wrote:
> > > All the above "definitions" apply to one sense of "word" - as in what
> > > particular vendors happen to mean in the context of data types.
> >
> >
> > Which is how I interepreted th OPs question...
>
> Well no, the 1st and 2nd posters did not clearly distinguish the two
> senses, hence the confusion.


Actually I *was* the second poster, and my interpretation of the OPs
question was that it was clearly from a programming perspective, hence
my response in the same context. Whichever, the context has now been
clarified.


> > I agree, the size of the GP registers is probably the best short
> > defintion of word size for an ISA, but there are certainly a large
> > number of architectures to which that defintion is difficult,
> > impossible or illogical to apply.
>
> There are also a large number where it is a straightforward and useful
> definition. Stokers of controversy usually cite the 68000 here, but
> this rule clearly gives it a word length of 32 bits. Definitions based
> on other architectural properties of course may give different
> answers... I seem to recall a good thread on various competing
> definitions but cannot locate it at the moment.


I've always considered 68K to be 32 bit. AFAIK, the only justification
for calling the original 68K a 16 bit machine is that was: one, what
Motorola called it, and two, an accurate reflection of the
microarchitecture of the actual implementation. Niether of which
changes the fact that the ISA was essentially straight 32 bit.

Some of the more difficult cases include things like stack machines
(which may not have any programmable registers at all - and may have
different width data and return stacks), machines like the Cray-1
(where you have a choice between considering the 24 and 64 bit GP
registers - FWIW, I'd call it a 64 bit ISA), and some of the
storage-to-storage machines (also mainly without GP registers). Some
of the microcontroller architectures also have quite weird combinations
of features.

.

Relevant Pages

  • Re: GetThreadContext and debug register problem
    ... If you want to set a breakpoint in a debug ... CONTEXT structure, and then pass it to SetThreadContext. ... memory breakpoints have to be numerically below 0x80000000 - otherwise, ... debug registers via two ways. ...
    (microsoft.public.win32.programmer.kernel)
  • [PATCH 09/21] 2.6.19 perfmon2 : register read-write operations
    ... This patch contains the core read-write operations on the PMU registers ... There is no read operation on PMC because they are always set by the users. ... * The context is locked and interrupts are disabled. ...
    (Linux-Kernel)
  • [PATCH 09/22] 2.6.22-rc3 perfmon2 : register read-write operations
    ... This patch contains the core read-write operations on the PMU registers ... There is no read operation on PMC because they are always set by the users. ... * The context is locked and interrupts are disabled. ...
    (Linux-Kernel)
  • Re: WMMX registers mangled after a context switch
    ... MMX registers are aliases for the x86 FPU registers, ... There's really no way for a user app to know when its going to be context switched. ... to save WMMX registers over context switches. ... during context switch and the BSP flag: ...
    (microsoft.public.windowsce.platbuilder)
  • Re: Start Menu "Open", "Open All Users", etc. failing
    ... Right-click is slow or weird behavior caused by context menu handlers: ... Windows XP Shell/User ... > On one of my XP Pro SP2 machines, right-clicking a folder in the Start ...
    (microsoft.public.windowsxp.configuration_manage)