mersenneforum.org  

Go Back   mersenneforum.org > Fun Stuff > Lounge

Reply
 
Thread Tools
Old 2020-10-23, 19:38   #34
sweety439
 
sweety439's Avatar
 
Nov 2016

2×13×97 Posts
Default

Quote:
Originally Posted by R2357 View Post
I noticed that if we classify the bases according to how efficient they are to identify at once (by the last digit) composites, base 10 is "quite" good, but not the best!
Not even up to 10, as base 6 is better.

I saw above that some wish a shifting to base 12, which is better than 10, but it's on the same rank as 6.

If we want to change our number system, as we already have existing bases, why not change for an even more efficient, how about 30, (I don't know if we have enough symbols for base 210).
Dozenal is the best base!!! See page https://dozenal.fandom.com/wiki/10 and the links at the bottom of this page.
sweety439 is offline   Reply With Quote
Old 2020-10-24, 04:21   #35
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

22·2,239 Posts
Default

Quote:
Originally Posted by sweety439 View Post
Dozenal is the best base!!!
Actually, I believe that 60060 is the best base. It is divisible by a lot of small numbers, and it fits in 16 bits so that you can use an uint32 to multiply two digits in a single clock cycle without loss of precision on many of the 32-bit processors nowadays (like ARM and Cortex M).

Hihi.
[/trolling]

Last fiddled with by LaurV on 2020-10-24 at 04:22
LaurV is offline   Reply With Quote
Old 2020-10-24, 04:30   #36
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

5×11×107 Posts
Default

Quote:
Originally Posted by R2357 View Post
(I don't know if we have enough symbols for base 210).
There are plenty of symbols available. One look at the Unicode tables will convince you that a lack of symbols is not the problem here.

ETA: Even LaurV's base 60060 can be accommodated with Unicode symbols. If one has the necessary desire.

Last fiddled with by retina on 2020-10-24 at 04:32
retina is online now   Reply With Quote
Old 2020-10-24, 07:30   #37
S485122
 
S485122's Avatar
 
Sep 2006
Brussels, Belgium

32·181 Posts
Default

Quote:
Originally Posted by LaurV View Post
Actually, I believe that 60060 is the best base. It is divisible by a lot of small numbers, and it fits in 16 bits so that you can use an uint32 to multiply two digits in a single clock cycle without loss of precision on many of the 32-bit processors nowadays (like ARM and Cortex M).

Hihi.
[/trolling]
Euh ... to fit in 16 bits it would be 30030 (13#).

Jacob

It it ain't you trolling, but them dozenalers. ;-)
They forget that most humans have 5 digits and two hands, that the supposed benefits of base 12 don't weight up to having (almost) the whole world change their way of counting, and having to replace (almost) all things displaying numbers...

Last fiddled with by S485122 on 2020-10-24 at 07:31
S485122 is offline   Reply With Quote
Old 2020-10-24, 07:58   #38
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

5·11·107 Posts
Default

Quote:
Originally Posted by S485122 View Post
Euh ... to fit in 16 bits it would be 30030 (13#).


2^16 = 65536, right?
retina is online now   Reply With Quote
Old 2020-10-24, 08:40   #39
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

100010111111002 Posts
Default

Quote:
Originally Posted by S485122 View Post
Euh ... to fit in 16 bits it would be 30030 (13#).
Nope. 16 bits can store a number between 0 and 65535. Why should you bother with negatives? In fact, I explicitly mentioned uint32. Please read again. Here "u" in front of "int" stands for "unsigned".
And for the 13# part, you would need to multiply it by 2, because mersenne classes for k need to be multiple of 4 to have q=2kp+1 be 1 or 7 (mod 8), see discussions in this thread. That's why mfaktc uses 4620 classes (2*11#) and not 2310. This way, you can very fast discard half of the filtered candidates (those which are 3 or 5 (mod 8)) without any test (because they will all be in the same class, which you just ignore), getting things done much faster.


Last fiddled with by LaurV on 2020-10-24 at 08:49
LaurV is offline   Reply With Quote
Old 2020-10-24, 09:44   #40
S485122
 
S485122's Avatar
 
Sep 2006
Brussels, Belgium

32·181 Posts
Default

Quote:
Originally Posted by retina View Post


2^16 = 65536, right?
Quote:
Originally Posted by LaurV View Post
Nope. 16 bits can store a number between 0 and 65535. Why should you bother with negatives? In fact, I explicitly mentioned uint32. Please read again. Here "u" in front of "int" stands for "unsigned".
And for the 13# part, you would need to multiply it by 2, because mersenne classes for k need to be multiple of 4 to have q=2kp+1 be 1 or 7 (mod 8), see discussions in this thread. That's why mfaktc uses 4620 classes (2*11#) and not 2310. This way, you can very fast discard half of the filtered candidates (those which are 3 or 5 (mod 8)) without any test (because they will all be in the same class, which you just ignore), getting things done much faster.

Oups : it is so obvious ... now.

Jacob

Instead of thinking I mishandled a program : I tried first with another number 2*2*2*3*3*5*7*11*13 But must have typed in something else, then deduced from the 18 bits that your number would use more than 16 while 655536 is a known quantity so closely associated with 16 bits ;-)

I'd wish somebody deleted that stupid post of mine.
S485122 is offline   Reply With Quote
Old 2020-10-24, 10:14   #41
Nick
 
Nick's Avatar
 
Dec 2012
The Netherlands

101111100012 Posts
Default

We all have moments like that!
Nick is online now   Reply With Quote
Old 2020-10-24, 10:53   #42
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

22·2,239 Posts
Default

Quote:
Originally Posted by S485122 View Post
I'd wish somebody deleted that stupid post of mine.
Buddy, you are not alone. All of us have some brain farts regularly, I for sure lead the list and these guys here are all eager and waiting to byte me when that happens. And I, them, haha. But that's the good part of science, when you make mistakes (general you), others correct you, and you realize the mistake, and you learn something from it. Then, next time you know, and can come with something better, therefore helping others in return.

Last fiddled with by LaurV on 2020-10-24 at 10:54
LaurV is offline   Reply With Quote
Old 2020-10-24, 11:16   #43
S485122
 
S485122's Avatar
 
Sep 2006
Brussels, Belgium

32×181 Posts
Default

Quote:
Originally Posted by LaurV View Post
Buddy, you are not alone. All of us have some brain farts regularly, I for sure lead the list and these guys here are all eager and waiting to byte me when that happens. And I, them, haha. But that's the good part of science, when you make mistakes (general you), others correct you, and you realize the mistake, and you learn something from it. Then, next time you know, and can come with something better, therefore helping others in return.
The problem in this case is that I didn't learn anything new about powers of two or numerical representations. It is just a case of not using my brain correctly.
You could say I learned that my brain can "fart", but that is something I know and despair ;-) about since a very long time. Also this brain fart is now saved for posterity :-D

But thanks for your and Nick's nice answers !

Jacob
S485122 is offline   Reply With Quote
Old 2020-10-24, 11:32   #44
xilman
Bamboozled!
 
xilman's Avatar
 
"π’‰Ίπ’ŒŒπ’‡·π’†·π’€­"
May 2003
Down not across

19×547 Posts
Default

Quote:
Originally Posted by S485122 View Post
They forget that most humans have 5 digits and two hands, that the supposed benefits of base 12 don't weight up to having (almost) the whole world change their way of counting, and having to replace (almost) all things displaying numbers...
See my avatar.
xilman is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
'All Your Data ❝Я❞ Belong To Us' Thread ewmayer Tales From the Crypt(o) 269 2020-11-16 16:02
Numbers Sierpinski to multiple bases Gelly Miscellaneous Math 18 2020-07-23 02:57
Other Bases? wblipp GPU Computing 50 2012-10-11 13:23
Cause this don't belong in the milestone thread bcp19 Data 30 2012-09-08 15:09
Riesel and Sierp numbers bases <= 1024 R. Gerbicz Conjectures 'R Us 22 2009-12-29 20:21

All times are UTC. The time now is 13:36.

Wed Dec 2 13:36:31 UTC 2020 up 83 days, 10:47, 2 users, load averages: 3.85, 4.34, 4.41

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.