Post Reply 
(Free42) not a bug - COMB
07-14-2020, 06:57 AM
Post: #9
RE: (Free42) not a bug - COMB
(07-14-2020 03:38 AM)Thomas Okken Wrote:  I'm coding the fix so that it starts the product with 1/s instead of 1, and then multiplies by s at the end, where s is 10^(1+floor(log10(p))), or 2^(1+floor(log2(p))) in the binary version. That fixes all the cases mentioned in this thread.

https://github.com/thomasokken/free42/co...487ac559e6

For Free42 Decimal, the largest value of C(n,p)*p that overflows while C(n,p) does not is for n=20421,p=10167, and that is also the only case that requires the constant s to be 1E5 i.o. 1E4. 1E5 will thus work for all cases in Free42 Decimal.
Likewise, s=2^10 will work for all cases in Free42 Binary. Anything larger will also work, of course.

Cheers, Werner

41CV†,42S,48GX,49G,DM42,DM41X,17BII,15CE,DM15L,12C,16CE
Find all posts by this user
Quote this message in a reply
Post Reply 


Messages In This Thread
(Free42) not a bug - COMB - Werner - 07-13-2020, 11:36 AM
RE: (Free42) not a bug - COMB - Paul Dale - 07-13-2020, 12:03 PM
RE: (Free42) not a bug - COMB - Werner - 07-13-2020, 12:22 PM
RE: (Free42) not a bug - COMB - Werner - 07-13-2020, 01:16 PM
RE: (Free42) not a bug - COMB - Werner - 07-13-2020, 12:28 PM
RE: (Free42) not a bug - COMB - Werner - 07-14-2020 06:57 AM



User(s) browsing this thread: 1 Guest(s)