Fascinating, Watson. The system is behaving properly. I will explain...
SPOILER ALERT
(-&2 +&$: -&1) ^: (1&<) M. 98
1.35302e20
(-&2 +&$: -&1) ^: (1&<) M. 98x
135301852344706746049
r =. (-&2 +&$: -&1) ^: (1&<) M.
r 98
1.35302e20
r 98x
1.35302e20
A memoized verb has a table of results for each input it has been
presented with. When you reuse the name f0b, you are getting the result
for the first time that value was presented - possibly in a different
precision.
You know my methods.
Henry Rich
On 9/25/2021 5:37 PM, Skip Cave wrote:
JVERSION
Engine: j903/j64avx2/windows
Beta-q: commercial/2021-09-14T15:38:42
Library: 9.03.01
Qt IDE: 1.9.2/5.12.10(5.12.10)
Platform: Win 64
Installer: J903 install
InstallPath: c:/users/skip/j903
Contact: www.jsoftware.com
I tried Elijah's extended precision fix. It didn't help on my machine:
pp=.9!:11
pp 20
f0b=: (-&2 +&$: -&1) ^: (1&<) M.
f0b 98
1.353018523447067607e20
x: f0b 98
135301852344706760704
x: f0b 98x
135301852344706760704
f2a=: 3 : '{. +/\@|.^:y 0 1x'
f2a 98
135301852344706746049
f2a 98x
135301852344706746049
x: f2a 98x
135301852344706746049
Skip Cave
Cave Consulting LLC
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
--
This email has been checked for viruses by AVG.
https://www.avg.com
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
