Mainly for Ian, I expect.
I’ve been playing with Raul Miller’s “countdown” solution as listed in
https://rosettacode.org/wiki/Countdown#J
Actually, I’ve modified “task” slightly, in order to force the target number
and the selection of the 6 numbers.
Here’s a copy of a so far successful session. As usual, I hope it displays ok
if using a fixed width font even though it looks awful on this iPad:
NB. timer just shows the crude elapsed time, and result, if any.
NB. task produces an empty result as the echo statements report what’s required.
timer'154 task 1 5 6 9 25 100' NB. runs reasonably fast
terms: 1 5 6 9 25 100
target: 154
#solutions: 392
for example: (5+9)*((1+6)+(100%25)) NB. 100+6*9 is my preference!
┌───────┬┐
│10.6086││
└───────┴┘
NB. Try program monitor, wondering if {.a shows as costly, cf Raul’s recent
query.
start_jpm_ 100000000
1785713
timer'154 task 1 5 6 9 25 100'
terms: 1 5 6 9 25 100
target: 154
#solutions: 392
for example: (5+9)*((1+6)+(100%25)) NB. not 100+6*9 !
┌───────┬┐
│12.5705││
└───────┴┘
showdetail_jpm_ 'task'
recorded all lines
used and max record count: 1,785,713 1,785,713
the PM data area has overflowed and records have been lost
Time (seconds)
┌────────┬────────┬───┬─────────────────────────────────────┐
│all │here │rep│task │
├────────┼────────┼───┼─────────────────────────────────────┤
│0.000000│0.000000│0 │monad │
│0.000000│0.000000│0 │[0] ''task y │
│0.000000│0.000000│0 │total monad │
│ │ │ │ │
│0.000000│0.000000│0 │dyad │
│0.000000│0.000000│0 │[1] if. #y do. c=.y │
│0.000000│0.000000│0 │[2] else. │
│0.000000│0.000000│0 │[3] c=./:~deal'' │
│0.000000│0.000000│0 │[4] end. │
│0.000000│0.000000│0 │[5] if. #x do. t=.x │
│0.000000│0.000000│0 │[6] else. │
│0.000000│0.000000│0 │[7] t=.targ'' │
│0.000000│0.000000│0 │[8] end. │
│0.000000│0.000000│0 │[9] echo'terms: ',":c │
│0.000000│0.000000│0 │[11] echo'target: ',":t │
│0.000000│0.000000│0 │[12] echo'#solutions: ',":#a=.t all c│
│0.000000│0.000000│0 │[13] echo'for example: ',;{.a │
│0.000000│0.000000│0 │total dyad │
│ │ │ │ │
└────────┴────────┴───┴─────────────────────────────────────┘
NB. Not a lot of use. So I tried
start_jpm_ 200000000
3571427
NB. This might be relevant. What is the memory limit?
9!:20''
17179869184
terms: 1 5 6 9 25 100
target: 154
#solutions: 392
for example: (5+9)*((1+6)+(100%25))
┌───────┬─┐
│12.7607│0│
└───────┴─┘
NB. So far so good, BUT
showdetail_jpm_ 'task' NB. leads to wipeout.
JVERSION
Engine: j903/j64/iOS
beta-k: GPL3/2021-10-15T01:07:05
Platform: iOS/iPadOS
Version: 901.1 49
Installer: App Store
Contact: www.jsoftware.com
I confess I copied jpm over from my Windows installation, and its use in iOS
might be deprecated, but I suspect there’s a more general problem when
exceeding memory. In another Rosetta task, seeking sum-partitions of a given
integer into a strictly increasing set of primes with exactly m members, some
larger problems lead to a crash. The code is a lot longer, though not as
sophisticated as Raul’s approach to countdown.
Cheers,
Mike
Sent from my iPad
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
