[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: reproducible format



I'd call such hardware buggy, but again this is a general system architecture problem that occurs even in integer, and so it is not floating point's problem to solve.

Let the market deal with it - don't buy defective hardware, even if it can produce wrong results quickly. And, to get portable results, write a portable software emulator to define the floating point semantics.

Nick Maclaren wrote:
Ivan Godard <igodard@xxxxxxxxxxx> wrote:
  
You're right in the presence of communicating threads, but they are not 
reproducible anyway, even in integer, and so this is not floating 
point's problem to solve. Absent threading, the hardware of my 
acquaintance goes to considerable effort to provide the illusion of 
in-order execution even with an out-of-order reality, and some vendors 
(such as my own company) are in-order to begin with.
    

I have used quite a lot of hardware that makes no such attempt.
But, even when the attempt is made, you will often find that it
applies only if no exception arises - and that includes ones that
are handled internally, and the application (or even operating
system) never sees.

This leads to the effect that a program fails for no apparent
reason, and rerunning it works.  Or it gives answer A and, when
rerun, it always gives answer B.  Yes, that happens, and it is
far more common than most people realise.


Regards,
Nick Maclaren,
University of Cambridge Computing Service,
New Museums Site, Pembroke Street, Cambridge CB2 3QH, England.
Email:  nmm1@xxxxxxxxx
Tel.:  +44 1223 334761    Fax:  +44 1223 334679

  

754 | revision | FAQ | references | list archive