Cross posting to the beta forum where we should continue the discussion.
I can't reproduce the error shown in J9.03 (linux jconsole) or J9.02
linux/windows jconsole/jqtide
x9=: 1e6 ?@$ 2e9
y9=: 1e5 ?@$ 2e9
f9=: x9&i.;
type 'f9'
++
|verb|
++
JVERSION
Engine: j903/j64avx2/lin
I have made a change to the page noting that the Kahan summation feature
was added in J9.03.
Imre - this thread relates to the current beta version of J. My expectation
is that there will be bugs and that this is our opportunity to help the
team iron them out before release.
If you have come acros
Kahan summation is a 9.03 feature,
Henry Rich
On 6/14/2021 7:46 PM, Imre Patyi wrote:
There seem to be many J examples that no longer work.
The second page on NuVoc about Kahan type of summation
does not work either.
https://code.jsoftware.com/wiki/Vocabulary/AccurateAccumulation
It has a tim
There seem to be many J examples that no longer work.
The second page on NuVoc about Kahan type of summation
does not work either.
https://code.jsoftware.com/wiki/Vocabulary/AccurateAccumulation
It has a timestamp from 2021.
0j20 ": (+/!.0) 1e7 # 100.1
|domain error
| 0j20":(+/ !.0)100
To get the fast copy you must use
old , new
in a context where old is inplaceable.
(":x) will never be inplaceable, as mentioned earlier.
Inplaceable contexts are
old =: old , new
and
old =. old , new
and
(string-creating expression) , new
Read in SpecialCombinations for details. It ans
OK, thank you. I had not realized that J did the linear scan backwards.
I guess I saw that in some examples such as alternating sum -/ and
continued fraction (+%)/ but plain forgot. Here is an experiment again.
slash=: 1 : 0 NB. For associative binary operation u.
N=.# y
if. 1=N do. y return. en
Did you really want the ; in f9?
Henry Rich
On 6/14/2021 12:52 PM, Lorenz Köhl wrote:
x9=: 1e6 ?@$ 2e9
y9=: 1e5 ?@$ 2e9
f9=: x9&i.;
--
This email has been checked for viruses by AVG.
https://www.avg.com
--
For information
I was running the some examples from
https://web.archive.org/web/20150103031253/http://www.jsoftware.com/help/release/midot.htm
x9=: 1e6 ?@$ 2e9
y9=: 1e5 ?@$ 2e9
f9=: x9&i.;
|out of memory: f9
| f9 y9
The process does not show a spike in memory allocated in task manager.
Maybe this is r
