Re: Motion P1788/0024.01: Rounding Mode as Operation
I'm glad I waited before I replied, because Hossam Fahmy did all the
work for me! He covered all the points I was going to make, and then
some.
I think there is a generic way of rephrasing the essence of the motion:
Every (essential) IEEE-754 arithmetic operation shall be provided
in a form with an explicit rounding direction, in addition to the
usual forms that respect a static global rounding direction. All
rounding modes provided for a given radix shall also be supported
in this manner.
Whether this is done via an additional parameter, or via a different
operation name, is (just like the actual names of operators or functions)
beyond the scope of the standard, as it falls within the purview of
language and library standards.
I might also point to two existing implementations of this motion,
both for DFP only at this point:
The decNumber library, where each operation has (besides the numeric
operands) a context argument. This context contains (among other
things) the rounding direction for this operation. (A global rounding
direction would simply be in a common context object.)
The DFP arithmetic instructions of IBM's z196 processor, where each
opcode includes a rounding mode field. Among those rounding modes
is the inherited rounding mode from the global FPC register.
Michel.
---Sent: 2011-04-18 18:43:04 UTC