P-1788 members:
Since Motion 24 has been made by Ulrich Kulisch and
seconded by Chenyi Hu, the discussion period now begins, and
will end after Sunday, May 8. I attach the motion.
Discussion on this motion will proceed according to the rules for
position papers.
Juergen: Please place the motion and associated information
in the appropriate place on the web page, as
you have aptly done in the past.
Acting secretary: Please record the transaction in the minutes.
Best regards,
Baker (acting as chair, P-1788)
P.S. (personal opinion):
This is an interesting motion. In particular, we have suffered in the
past from requiring many clock cycles to change the rounding mode
in current hardware, whereas this would not be so if the rounding
mode were in the opcode. Specifying separate operations for
each rounding mode may start hardware engineers thinking that way.
I note that, although such a specification does not seem to fit
centrally into an interval arithmetic standard, at least not at
the top level, it is clearly relevant and closely tied to any
such standard. Furthermore, it could be relevant to people (of
which there are significant numbers, although not ubiquitous)
who use directed rounding directly, without interval arithmetic.
P.P.S. I am beginning to get nervous about our plan for the rest of
our term, and hope good progress is made on the actual wording.
I'll work with the officers soon (and in any case after our semester
ends here in May) to both work out a schedule for a vote on
portions of the standard and to identify issues that still need
position votes. In the mean time, anyone who wishes to take the
initiative on this is welcome. Also, anyone who knows of an
issue which will need to be resolved with a position vote
(and which hasn't yet been so resolved) is welcome to present it.
On 4/15/2011 11:43, Ulrich Kulisch wrote:
Dear colleagues,
As an attempt to eliminate the usual difficulties that occur if interval arithmetic is implemented using IEEE 754 floating-point arithmetic I prepared a motion on Rounded Floating-Point Arithmetic. The text is attached.
Best withes
Ulrich
--
---------------------------------------------------------------
Ralph Baker Kearfott, rbk@xxxxxxxxxxxxx (337) 482-5346 (fax)
(337) 482-5270 (work) (337) 993-1827 (home)
URL: http://interval.louisiana.edu/kearfott.html
Department of Mathematics, University of Louisiana at Lafayette
(Room 217 Maxim D. Doucet Hall, 1403 Johnston Street)
Box 4-1010, Lafayette, LA 70504-1010, USA
---------------------------------------------------------------
<RoundOper.pdf>