[J3] Proposals for new intrinsic procedures
Van Snyder
van.snyder at sbcglobal.net
Thu Jan 5 21:03:41 UTC 2023
Some of these were proposed in 04-184r1 at meeting 167. In the scoring
in the 04-265r1 spreadsheet, the "hate dislike like love" score was 0-
0-11-2. The impact on the standard was considered to be small. THey
didn't meet the development priority, so detailed papers were not
written.
The following were in a draft of 04-227r1, but they were removed in
subgroup:
----------------------------------------------------------------------
13.7.58a EXPM1 (X)
Description. Exponential function of X, less 1.
Class. Elemental function.
Argument. X shall be of type real.
Result characteristics. Same as X.
Result value. The result has a value equal to a processor dependent
approximation to $e^{X}-1$.
Example. EXPM1(1.0E-11) has the value 1.0E-11 (approximately).
NOTE 13.xxx
It is recommended that the result be calculated using a method that
does not incur loss of precision by forming $e^X$ and then
subtracting 1.
Note to J3: EXP(1.0E-11) has the value 1.0 + 1.0E-11 to first order.
This is 1.0 in IEEE single precision, so EXP(1.0E-11)-1.0 has the
value 0.0 in IEEE single precision, not the "correct" value of 1.0E-11.
----------------------------------------------------------------------
13.7.98a LOGP1 (X)
Description. Natural logarithm of X+1.
Class. Elemental function.
Argument. X shall be of type real.
Result characteristics. Same as X.
Result value. The result has a value equal to a processor dependent
approximation to $\log_e^{X+1}$.
NOTE 13.xxx
It is recommended that the result be calculated using a method that
does not incur loss of precision by forming X+1.
Example. LOGP1(1.0E-11) has the value 1.0E-11 (approximately).
Note to J3: 1.0 + 1.0E-11 has the value 1.0 in IEEE single precision,
so LOG(1.0 + 1.0E-11) has the value 0.0 in IEEE single precision, not
the "correct" value of 1.0E-11.
An optional part of the standard containing interfaces for special
functions defined in ISO 31-11:1992 part 11 was proposed in WG5/N1688.
Van
On Thu, 2023-01-05 at 19:35 +0000, Brad Richardson via J3 wrote:
> Hi all,
>
> Would proposals for new intrinsic functions be acceptable to submit
> for
> the upcoming meeting? Specifically I've been asked to write up the
> papers for:
>
> * prefix sum (commonly known as scan, but of course that name's
> already
> taken), both inclusive and exclusive
> * All functions recommended by IEEE-754, most of which Fortran
> already
> has, but the whole list is reproduced below
>
> exp
> expm1
> exp2
> exp2m1
> exp10
> exp10m1
> log
> log2
> log10
> logp1
> log2p1
> log10p1
> hypot
> rSqrt
> compound
> rootn
> pown
> sin
> cos
> tan
> sinPi
> cosPi
> tanPi
> asin
> acos
> atan
> atan2
> sinh
> cosh
> tanh
> acosh
> atanh
>
> Some points of order questions also. Which subgroup is in charge of
> the
> intrinsic procedures so that I can coordinate with them? Would it be
> best to write a separate paper for each new procedure, put them all
> in
> a single paper, or some scheme for grouping certain subsets? And
> whilst
> we don't have final version of F2023 to draft the edits against,
> would
> it be worthwhile to go ahead and detail out what they will look like?
>
> Thanks in advance for any feedback.
>
> Regards,
> Brad
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20230105/7fb9eb13/attachment.htm>
More information about the J3
mailing list