Page 4 of 24
Re: Larger than Life
Posted: July 6th, 2017, 7:23 am
by calcyman
How is the convolution step implemented? If the neighbourhood is a n-by-n square, you can first perform a convolution by a 1-by-n rectangle, followed by a convolution by a n-by-1 rectangle. So you only have to do ~ 2n additions instead of ~ n^2 additions per cell.
Also, if n is composite, you can get a further speedup. For example, if n = 15, you can replace the convolution by:
with two successive convolutions by each of:
For large neighbourhoods which don't easily 'factorise' in this manner, such as a circular neighbourhood, then an FFT (over F_p for large p) is probably the way to go.
Re: Larger than Life
Posted: July 6th, 2017, 8:53 am
by Andrew
calcyman wrote:How is the convolution step implemented?
You answer
my questions and I'll answer yours.
(That's assuming I can figure out what you mean by "convolution step". Do you mean the calculation of a cell's neighborhood count?)
Re: Larger than Life
Posted: July 6th, 2017, 10:55 am
by blah
This rule generates tree rings, though they get kind of octahedral as the tree grows:
Code: Select all
x = 24, y = 24, rule = R9,C0,M1,S30..180,B24..45,NN
2b4ob2o3b2obo$3b2o5bobo3bob3ob2o$b7o3bob2ob3ob4o$5b4obob2obobo2bo2bo$
2bob2o2bo2b3o2bobobo2bo$3b2obo3b2o2bob4ob3o$3o3b4ob2ob2o4b4o$b3o4b4o2b
ob2obo2b2o$bo4bo2b2o2bobobob3o$bo4b2ob2o3bo5bobo$obobo2b2o3bo2b3ob5o$b
5o2bobob2o4b2obobo$5b2obo3b3o2bo2b2obo$ob2o2b3ob2o2bo4b2obo$3o2b2ob3ob
2o4b4obo$5bobobo2bo2b3ob2o$b4o5bo2b3o3bobo$bob2o3bo2bob5obo3bo$bobob2o
3b2o2b4obobo$3b5ob4ob4obo3bo$3b2ob6o2bobo3bobo$bobo2bob2o2bo2bob2obo2b
o$b2o3bob4o2bob3obo2bo$2o2b2ob3o2bobobobo3bo!
Re: Larger than Life
Posted: July 6th, 2017, 12:00 pm
by bprentice
Rule R3,C2,M1,S7..19,B12..23,NN supports a rich set of oscillators. An archive containing a collection of these oscillators is here:
http://bprentice.webenet.net/Larger%20T ... 20Life.zip
They are in file ../Version 1/Patterns/Rule 053.
Brian Prentice
Re: Larger than Life
Posted: July 6th, 2017, 12:54 pm
by bprentice
Rule R4,C4,M1,S34..35,B7..19,NN
3c/2 diagonal ships
Code: Select all
x = 105, y = 67, rule = R4,C4,M1,S34..35,B7..19,NN
30.C$30.CB$30.C2B$30.2CBA$30.2CBA$31.CBA$31.CB2A$31.CB2A$30.2C2BA$30.
2C2BA$30.2C2BA$31.CB2A$31.CB2A$31.CB2A$30.2C2BA$30.2C2BA$30.2C2BA$31.
CB2A$31.CB2A$31.CB2A$30.2C2BA$30.2C2BA$30.2C2BA$30.2CB2A$30.C2B2A$29.
2C2B2A$29.2C2B2A$29.2C2BA$30.CB2A$30.CB2A$30.CB2A$29.2C2BA$29.2C2BA
$29.2C2BA$29.2CB2A$29.C2B2A66.C$28.2C2B2A66.2C$28.2C2B2A65.2C2B$28.
2C2BA65.3C2BA$28.2CB2A65.2C3B2A$27.2C2B2A64.3C2B2A$27.2C2B2A63.3C2B
3A$26.2C3B2A62.3C3B2A$25.3C2B2A61.4C3B3A$25.2C3B2A60.4C3B3A$24.3C2B
2A59.4C4B3A$24.2C2B3A57.4C4B4A$23.2C3B2A57.4C4B4A$22.3C2B3A54.5C4B4A
$22.2C3B2A56.2C4B4A$21.3C2B2A58.4B4A$21.2C2B3A59.5A$20.2C3B2A61.2A$
19.3C2B3A$19.2C3B2A$18.3C2B2A$18.2C2B3A$17.2C3B2A$16.3C2B3A$14.4C3B
2A$12.4C4B2A$7.8C4B3A$11C6B4A$.2B3C9B4A$2.8B8A$3.11A$6.3A!
6,5c/4 knight ships
Code: Select all
x = 76, y = 17, rule = R4,C4,M1,S34..35,B7..19,NN
11.C4.C50.C$12.B3.CBA47.2C$11.2CA.2C2BA46.2CB$11.2CB.2C2BA44.3C2BA$
10.2C2B.2C2BA42.4C3B2A$8.4C2BAC2B2A26.4C3.2C3.7C3B2A$2C3.5C3B2A.CB2A
26.2B13C6B3A2.2C$.7C4B3A2.BA28.18B4A3.2CB$2.9B3A2.2CA29.5A2B3A2B7A4.
2C2B$3.2A2B6A3.2CB32.13A5.4C2BA$4.7A4.2C2B48.4C3B2A$13.4C2BA38.2C3.
7C3B3A$5.2C3.5C3B2A39.7C6B3A$6.7C4B3A40.10B4A$7.9B3A42.2A2B7A$8.2A2B
6A44.7A$9.7A!
2,3c/2 knight ships
Code: Select all
x = 49, y = 14, rule = R4,C4,M1,S34..35,B7..19,NN
2.C$2.2C$3.CBA38.C$3.CB2A37.CB$3.C2BA38.CBA$3.C2BA38.CBA$3.2CBA37.2C
2BA$3.2CBA37.2C2BA$3.C2B2A36.2CB2A$2.2C2B2A35.2C2B2A$.3C2BA35.2C3B2A
$3C2B2A37.2B2A$.3B3A38.2A$2.B3A!
Gun that shoots 3,2c/2 knight ships
Code: Select all
x = 18, y = 20, rule = R4,C4,M1,S34..35,B7..19,NN
.2A$3A2BA$.3B2C3B2A$.2B3C2.2CA3.C$.2A2C2.ACA2B.2A$.CAB2.A.C.2BA2C2B
$2C.B2.A2.B2C2AC2B$3.2B2A3.C2A2.CB$3.A2CA.3C2B2.2C$2.2A2CB.2C3BA.2C
$3.A3B.2C2B2A.CB$3.A3BA3B3A2.B$5.ACA.B3A2.CBA$5.2CA2.A3.2CBA$5.A.A5.
3C2B$5.2B8C2BA$6.2B3C5B2A$6.2A7B3A$9.8A$11.3A!
Brian Prentice
Re: Larger than Life
Posted: July 7th, 2017, 3:45 am
by Dean Hickerson
bprentice wrote:Rule R3,C2,M1,S7..19,B12..23,NN supports a rich set of oscillators.
...
Oscillators with periods 2 thru 26 have now been found.
Here are some more; the first missing period is now 45. (Although it's not listed in your post, your collection includes a p44 oscillator.)
Code: Select all
x = 203, y = 92, rule = R3,C0,M1,S7..19,B12..23,NN
3o2b3o10b18o39b2o46b3o2b3o2b3o2bobo10b30o$2bo4bo10b19o38b4o46bo4bo4bo
2bobo10b14ob2ob2o2b10o$3o4bo10b7o2b6ob3o38b5o43b3o2b3o4bo2b3o11b32o$o
6bo10b20o19b3o2b3o10b6o42bo4bo6bo4bo11b5o2b26o$3o4bo10b20o19bo6bo10bob
5o41b3o2b3o4bo4bo12b15ob2o2b13o$57b3o2b3o10b8o71b31o$59bo2bo12b9o$57b
3o2b3o10b10o$75b11o$75b12o$3o2b3o11b19o37b12o$2bo2bobo10b22o83b3o2b3o
2bobo2b3o10b31o$3o2bobo10b12ob7ob3o83bo2bobo2bobo2bobo10b2o2b17ob11o$
2bo2bobo10b24o81b3o2b3o2b3o2b3o10bo2b23ob7o$3o2b3o10b24o81bo4bobo4bo4b
o10bo2b23ob8o$123b3o2b3o4bo2b3o10b2o2b17ob14o$57b3o2b3o11b18o57b35o$
57bo4bo12b21o$57b3o2b3o10b3ob19o$59bo2bobo10b23o$2b3o2bo11b20o18b3o2b
3o10b23o$4bo2bo10b4o2b16o$2b3o2bo10b4o2b16o83b3o2b3o2b3o2b3o10b30o$4bo
2bo10b4o2b16o83bo6bo2bobo2bobo10b32o$2b3o2bo10b22o83b3o2b3o2b3o2b3o10b
5o2b26o$125bo4bo4bo4bo10b5o2b27o$57b3o2b3o11b17o30b3o2b3o2b3o2b3o10b
31ob3o$57bo6bo10b20o56b34o$57b3o2b3o10b22o$57bobo2bo12b22o$19b5o33b3o
2b3o10b22o$3o2b3o11b7o$2bo2bobo10b10o$3o2b3o10b12o88b3o2b3o2b3o2b3o2b
3o12b45o$2bo2bobo10b14o88bo2bobo2bobo4bo2bo13b40o2b6o$3o2b3o10b13o87b
3o2b3o2bobo2b3o2b3o10b40o3b8o$18b13o26b3o2b3o11b19o25bo4bo2bobo2bo6bo
10b41ob9o$57bo4bobo10b5o2b15o21b3o2b3o2b3o2b3o2b3o10b51o$57b3o2b3o10b
24o$57bobo2bobo10b5o2b17o$57b3o2b3o10b24o2$3o2b3o11b19o$2bo2bobo10b21o
76b3o2b3o2b3o2b3o2b3o2bo12b45o$3o2b3o10b21o76bo4bo4bobo2bo6bo2bo11b22o
b8ob16o$2bo4bo10b20o77b3o2b3o2b3o2b3o2b3o2bo10b43o2b5o$3o2b3o10b20o19b
3o2b3o12b18o22bo2bobo2bobo4bo2bo4bo10b43ob7o$59bo4bo11b19o20b3o2b3o2b
3o2b3o2b3o2bo10b23ob8ob19o$59bo2b3o10b2o2b7ob8o56b51o$59bo4bo10b21o$
59bo2b3o11b19o2$2bobo2bo10b12o$2bobo2bo10b14o$2b3o2bo10b15o$4bo2bo10b
8ob7o$4bo2bo10b8o2b3ob3o43b2o$18b16o42b5o$59b3o2bo11b6o$59bobo2bo10b2o
b5o$59b3o2bo10bob7o$59bobo2bo10b10o$59b3o2bo10b11o$obo2b3o10b16o41b12o
$obo4bo10b2o7b9o39b13o$3o2b3o10bo9b10o37b12o$2bo4bo10bo9b10o$2bo2b3o
10b2o7b11o$18b20o3$57b3o2b3o12b19o$57bobo2bobo11b22o$57b3o2b3o10b20ob
4o$obo2b3o11b19o19bobo4bo10b21ob3o$obo4bo11b19o19b3o2b3o10b25o$3o4bo
10b20o$2bo4bo10b21o$2bo4bo11b19o3$75b12o$75b2ob9o$75bob4ob5o$57b3o2b3o
10b6o2b4o$57bobo2bobo10b6o2b4o$57b3o2b3o10b12o$59bo2bobo10b12o$57b3o2b
3o10b12o$75b5o2b5o$75b5o3b4o$75b12o!
As the last column shows, you can get many large periods by starting with a 5xN or 6xN rectangle and modifying the ends in various ways.
Re: Larger than Life
Posted: July 7th, 2017, 9:02 am
by bprentice
Dean Hickerson wrote:Although it's not listed in your post, your collection includes a p44 oscillator.
A few more oscillators were found and the archive was updated but alas the forum post was not. The archive does contain a file 'periods' which is current.
Thank you for contributing to the oscillator collection.
The oscillators were found using the program here:
http://bprentice.webenet.net/Larger%20T ... 20Life.zip
This archive contains two versions of the program. Each version includes the Java source code together with patterns organized into rules which can be run. The programs support five neighborhoods (SQUARE, CROSS, SALTIRE, STAR and DIAMOND) which were briefly discussed here:
viewtopic.php?f=7&t=2932#p45697
Rule 53, which you contributed to, is a modification of rule 44 which used the STAR neighborhood. Both rules generate a large assortment of oscillators.
The oscillator collection has been updated to include your contribution.
Brian Prentice
Re: Larger than Life
Posted: July 7th, 2017, 6:04 pm
by drc
Dean Hickerson wrote:I wrote:Can anyone find anything slower? Or maybe a slow knightship?
Here's the slowest yet, an orthogonal spaceship with range 43:
How are you finding these?
Re: Larger than Life
Posted: July 8th, 2017, 9:30 am
by Dean Hickerson
drc wrote:How are you finding these?
I don't remember how I found the first one, back in 2002. For the others, I've just made minor changes to that rule (including rescaling as I mentioned
here), and then tried random stuff. Most stuff either dies out or turns into an oscillator, but occasionally slow spaceships arise.
Re: Larger than Life
Posted: July 8th, 2017, 9:36 am
by Saka
Dean Hickerson wrote:drc wrote:How are you finding these?
I don't remember how I found the first one, back in 2002. For the others, I've just made minor changes to that rule (including rescaling as I mentioned
here), and then tried random stuff. Most stuff either dies out or turns into an oscillator, but occasionally slow spaceships arise.
Either way, you are a great guy
I got bored watching those spaceships move, especially the really slow ones. Wish golly would run faster at 10^2 and so on, but I know it is hard for Andrew.
Anyways, what script should I make next? My next project is to make a small script to... well, it's a secret!
Also, I have also been making a little collection of patterns I found in LTL hehe
Re: Larger than Life
Posted: July 8th, 2017, 1:59 pm
by muzik
Why does it run so slow anyway? Can't you make larger than life run on hashlife instead somehow?
Re: Larger than Life
Posted: July 8th, 2017, 2:22 pm
by AforAmpere
Awesome spaceships in this rule, a 3c/4, a c/2, a P3 knightship, and a 2c/4 diagonal:
Code: Select all
x = 44, y = 7, rule = R2,C5,M0,S2..2,B3..4,NN
3.2A10.2A8.5A7.A2B3A$3A2B3A6.A2BA6.BC4BA6.A3C2BA$B6CB6.C2.C6.C4DBA7.A
2.2BA$B6DA21.2B9.B.CA$CD4.B22.BA11.CB$41.ACB$29.D12.2A!
This variant has a 12c/56 (3c/14), and c/2, p3 knightship and the 2c/4 diagonal from the other rule, along with a p12 spinning oscillator.
Code: Select all
x = 78, y = 7, rule = R2,C4,M0,S2..2,B3..4,NN
15.2A17.4A15.5A13.A2B3A$.AB2A9.A2BA16.BACA14.BC4BA12.A3C2BA$BCBCA9.C
2.C17.2CA14.5CBA13.A2.2BA$C3.A31.BA19.2B15.B.CA$3.BA52.BA17.CB$75.ACB
$57.C18.2A!
EDIT, a rule with a common 4c/6 and a 5c/14 diagonal:
Code: Select all
x = 25, y = 11, rule = R2,C6,M0,S3..4,B3..4,NN
.3A14.C2D$A3BA13.D2EC$A3BA13.D3EDA$EB.BE14.2ECB2A$.A.A15.DA3BA$20.2AB
$14.B.C4.2A$13.AC2DC$13.B2AD$13.A.ACB$14.ABA!
EDIT 2, a better variant of the 1st rule, with an extra 4c/5:
Code: Select all
x = 67, y = 9, rule = R2,C5,M0,S2..2,B3..5,NN
9.3A13.2A11.2A8.5A7.A2B3A$5.4A4B3A6.3A2B3A7.A2BA6.BC4BA6.A3C2BA$2.3AB
A6C4BA4.B6CB7.C2DC6.C4DBA7.AD.2BA$A5BC8DBCB4.A6DB22.2B4.D4.B.CA$BCB5D
7.D2C5.B4.DC22.BA4.D5.DCB$2CD.D11.CB46.ACB$BC50.D12.2A2$62.2D!
Re: Larger than Life
Posted: July 8th, 2017, 7:45 pm
by bprentice
Oscillators with periods 2 thru 66 have now been found using rule R3,C2,M1,S7..19,B12..23,NN. The archive referenced here:
viewtopic.php?f=11&t=2933&start=75#p46105
has been updated.
Brian Prentice
Re: Larger than Life
Posted: July 8th, 2017, 10:57 pm
by Ethanagor
This may or may not be known (I'm not sure where to look) but here is a one-time 90 degree reflector of a diagonal bug in the same rule as Bosco:
Code: Select all
x = -10, y = -20, rule = R5,C0,M1,S34..58,B34..45,NM
4b3o$3b7o$2bo2b6o$bo4b5o$2o5b4o$b2o4b4o$b4ob5o$b9o$2b7o$3b5o$4b3o22$
15b5o$14b7o$13b8o$13b4o2b2o$13b3o3b2o$13b4ob3o$14b6o$15b4o!
Re: Larger than Life
Posted: July 8th, 2017, 11:25 pm
by Saka
This is literally just hilarious
Bug: WINGS; EXTEND!
Computer: Wing extension failed; still life ahead
Bug: Noooo
boom
Code: Select all
x = 90, y = 43, rule = R5,C0,M1,S34..58,B34..45,NM
83b3o$82b5o$30b2o13b2o13b2o19b7o$28b6o9b6o9b6o16b4ob4o$27b8o7b8o7b8o
14b4o3b4o$26b5o2b3o5b5o2b3o5b5o2b3o13b3o5b3o$26b2o2bo3b2o5b2o2bo3b2o5b
2o2bo3b2o13b3o5b3o$26b2o3b2ob2o5b2o3b2ob2o5b2o3b2ob2o13b4o3b4o$26b3ob
6o5b3ob6o5b3ob6o14b4ob4o$27b8o7b8o7b8o16b7o$29b4o11b4o11b4o19b5o$30b2o
13b2o13b2o21b3o6$4b3o$3b6o$2b9o$b2o3b6o$bo5b5o$2o6b4o$bo5b5o$b2o3b6o$
2b9o$3b6o$4b3o6$30b2o13b2o13b2o$29b4o11b4o11b4o$27b8o7b8o7b8o$26b3ob6o
5b3ob6o5b3ob6o$26b2o3b2ob2o5b2o3b2ob2o5b2o3b2ob2o$26b2o2bo3b2o5b2o2bo
3b2o5b2o2bo3b2o$26b5o2b3o5b5o2b3o5b5o2b3o$27b8o7b8o7b8o$28b6o9b6o9b6o$
30b2o13b2o13b2o!
Also, who can find the first bug in Bugs Range 50? (Convert using my script/Dean Hickerson's algorithm)
Re: Larger than Life
Posted: July 10th, 2017, 4:08 am
by Dean Hickerson
Saka wrote:Also, who can find the first bug in Bugs Range 50? (Convert using my script/Dean Hickerson's algorithm)
Here's one:
Code: Select all
#C Symmetric spaceship that moves southeast with speed 3000/929 ~ 3.229
x = 92, y = 92, rule = R50,C0,M1,S2866..4890,B2866..3794,NM
36b14o$33b20o$31b24o$29b28o$27b32o$25b36o$24b38o$22b42o$21b45o$19b48o$
18b51o$17b53o$16b56o$15b59o$14b17o2b42o$13b17o4b43o$12b16o7b43o$11b16o
8b45o$10b16o10b45o$9b15o14b44o$9b14o16b44o$8b14o19b43o$7b14o21b43o$7b
13o23b43o$6b13o26b42o$5b14o27b41o$5b13o29b41o$4b13o31b41o$4b12o33b40o$
3b13o34b40o$3b12o36b39o$2b12o38b38o$2b12o38b39o$b14o38b38o$b15o38b37o$
b17o37b37o$19o36b37o$19o37b36o$20o36b36o$21o35b36o$21o35b36o$22o34b36o
$23o33b36o$24o32b36o$24o32b36o$25o31b36o$26o30b36o$27o29b36o$28o28b36o
$29o26b37o$b29o24b38o$b30o23b37o$b32o20b38o$2b32o18b39o$2b33o15b40o$3b
34o12b41o$3b86o$4b85o$4b84o$5b83o$5b82o$6b80o$7b78o$7b77o$8b75o$8b74o$
9b71o$10b69o$10b68o$11b66o$12b64o$12b63o$13b61o$13b60o$14b58o$15b56o$
15b55o$16b53o$17b51o$17b50o$18b48o$19b47o$20b45o$21b43o$22b41o$23b39o$
24b37o$26b34o$27b31o$29b27o$32b22o$35b16o!
I found this by starting with a large random rectangle of density 40%. I saw something that moved for a while before vanishing, and noticed that some phases were almost symmetric across a diagonal. I made it symmetric by ORing it with its reflection. The result moved for quite a while before vanishing, so I started making small, symmetric modifications. Eventually I found one that lasted forever, becoming the spaceship shown above.
Re: Larger than Life
Posted: July 10th, 2017, 6:39 am
by Saka
Dean Hickerson wrote:Saka wrote:Also, who can find the first bug in Bugs Range 50? (Convert using my script/Dean Hickerson's algorithm)
Here's one:
It's shockingly similar to this bug I found in Bugs Range 15, another rule which is similarly empty.
Code: Select all
x = 29, y = 29, rule = R15,C0,M1,S269..461,B270..357,NM
11b9o$10b11o$9b13o$8b15o$7b17o$6b19o$5b20o$4b22o$3b23o$2b25o$b26o$13o
5b10o$12o7b9o$11o9b9o$11o9b9o$11o10b8o$11o12b6o$11o13b5o$12o13b3o$13o
11b4o$b14o8b4o$2b14o7b4o$3b13o6b4o$4b13o3b5o$5b13ob5o$7b16o$9b13o$11b
9o$13b5o!
Also, what's ORing? Oring, or Or-ing?
Re: Larger than Life
Posted: July 10th, 2017, 9:42 am
by Gamedziner
Why does that code show a hollowed out box in Viewer?
(EDIT: Snipped out most of quote at Saka's request.)
Re: Larger than Life
Posted: July 10th, 2017, 9:48 am
by Saka
Gamedziner wrote:
Why does that code show a hollowed out box in Viewer?
The grid size is too big. And also, please snip out the really big posts.
Re: Larger than Life
Posted: July 10th, 2017, 5:26 pm
by AforAmpere
RRO:
Code: Select all
x = 161, y = 67, rule = R3,C9,M0,S6..8,B6..7,NM
127.A$124.A5.2A$124.AB2.2B2.A$123.2A4.2B.A$122.3A4B.B.A$122.2AC2.4B.A
5.H$122.C2A.2DB2E.A4.H$124.A2D2.B2G5.H$124.2DA.FGF3H4.G$123.3D.4F6.G$
123.4DFDF.4H.EH$125.ED2FGF5HEF$122.C3.E2C.G5FG$124.C2.CE.G3F2.G$129.E
3.FG$133.F2.E36$2.E2.F140.E2.F$4.GF3.E138.GF3.E$2.G2.3FG.EC2.C131.G2.
3FG.EC2.C$2.G5FG.2CE3.C129.G5FG.2CE3.C$.FE5HFG2FDE131.FE5HFG2FDE$.HE.
4H.FDF4D129.HE.4H.FDF4D$.G6.4F.3D129.G6.4F.3D$G4.3HFGF.A2D129.G4.3HFG
F.A2D$.H5.2GB2.2DA130.H5.2GB2.2DA$.H4.A.2EB2D.2AC128.H4.A.2EB2D.2AC$H
5.A.4B2.C2A127.H5.A.4B2.C2A$6.A.B.4B3A133.A.B.4B3A$6.A.2B4.2A134.A.2B
4.2A$6.A2.2B2.BA135.A2.2B2.BA$7.2A5.A136.2A5.A$11.A143.A!
Rule with ships at speeds (reduced) 2c/1, (2,1)c/1, (3,2)c/2, c diagonal, and 3c/2 diagonal, and a (3,1)c/2 pufffer engine that I don't know if it can be stabilized:
Code: Select all
x = 239, y = 43, rule = R3,C8,M0,S1..1,B6..7,NM
4.4A92.4A21.A2.3A16.2A23.3A28.2A26.3A$3.A2B2.A89.2A4.2A19.A2.3A15.A2.
2A20.3B2A26.A3.2A22.2B2.3A$4.3B2.A87.2A.4B.2A17.BA.3B.A13.4B.2A18.3C
2B.A24.3B3.2A20.BA2.2B3A$3.4CB2.A86.A2B4.2BA17.BA.3B.A13.2C2.2B2A16.
3D2C.B.A22.2C2.3B2A18.A3C3.2B.A$2.C2D2BCB2.A85.2B.4C.2B16.C.A3C.B.A
11.CBA2C.2BA15.3E2D.C.B2A20.C.A.2C3BA18.2DBA3C2B$3.3D2BCB2.A84.B2C4.
2CB16.C.A3C.B.A10.4D.2C.BA14.3F2E.D.C2BA19.3DA2.2C.B.A16.DC.B2D3CB2.A
$2.4ED2BCB2.A83.2C.4D.2C12.B2.DC.3D.C.BA10.2EB.2D2C.BA12.3G2F.E.D2CBA
18.2E2A3D2C3.A14.B3E.B.2D.C.B.A$.E2F2DED.BCB2.A82.C2D4.2DC8.2A2B3.DC.
3D.C.B10.EDC2E.2DC2A15.2G.F.E2DCB18.E.C.2E3DC2.BA14.2FD.3E2D3ABA$2.3F
2DED.BCB2.A81.2D.4E.2D8.AB2A2CE2C3E.D.CB9.4F.2E.DC19.G.F2EDC18.3FC2.
2E.DAC.B14.F2.D2F3EDA.CB$.4GF2DED.BCB2.A80.D2E4.2ED8.AB2A2CE2C3E.D.C
10.2GB.2F2EBDCB18.G2FED18.2G2C3F2E2.ACB14.3G2D.2F.EBD.C$G2.2FGF.DED.B
CB2.A79.2E.4F.2E7.2B2C2DFE.3F.E.DC9.GFE2G.2FE2C20.2GFE18.G.E.2G3FE2BD
C17.E3G2F3.DC$4.2FGF.DED.BCB2.A78.E2F4.2FE9.2C2DFE.3F.E.D14.2G.FE23.G
F21.E2.2G.FBE.D15.GF3.3GFACED$5.2FGF.DED.BCB2.A77.2F.4G.2F8.C2.DGCE3G
.F.ED15.2GDFED21.G20.2E3.2G2.BED13.G3.C5.GDFDE$7.FGF.DED.BCB2.A76.F2G
4.2GF8.C2.DGCE3G.F.E11.G5.G2E43.G6.G2DFE17.G5.3EFE$8.FGF.DED.BCB2.A
75.2G6.2G7.AD2C.G5.G.FE17.G45.G6.DG.F26.GF$9.FGF.DED.BCB2.A74.G8.G8.D
.A.G5.G.F11.F5.F.GF41.2G7.DGF26.FGD$10.FGF.DED.BCB2.A91.2G.2G6.GF17.
2G50.2F.G23.3G.G$11.FGF.DED.BCB2.A90.2G.2G6.G69.F27.G$12.FGF.DED.BCB
2.A100.G69.F30.E$13.FGF.DED.BCB2.A$14.FGF.DED.BCB2.A193.F$15.FGF.DED.
BCB2.A$16.FGF.DED.BCB2.A$17.FGF.DED.BCB2.A$18.FGF.DED.BCB2.A$19.FGF.D
ED.BCB2.A$20.FGF.DED.BCB2.A$21.FGF.DED.BCB2.A$22.FGF.DED.BCB2.A$23.FG
F.DED.BCB2.A$24.FGF.DED.BCB2.A$25.FGF.DED.BCB2.A$26.FGF.DED.BCB2.A$
27.FGF.DED.BCB2.A$28.FGF.DED.BCB2.A$29.FGF.DED.BCB2.A$30.FGF.DED.BCB$
31.FGF.DED.BCB$32.FGF.DED.BCB$33.FGF.DED.BC$34.FGF.DED.B$35.FGF.DED$
36.FGF.DED!
Re: Larger than Life
Posted: July 10th, 2017, 5:51 pm
by bprentice
A fascinating combination of isolated blocks interspersed with oscillators.
Code: Select all
x = 12, y = 11, rule = R10,C3,M1,S47..177,B7..7,NN
4.A2.A$4.2A.A$3.3A.2A$2.3A.A.2A$.3A.3A.2A$9A.2A$2A.7A.A$A.A.4A.2A$5.
5A$6.3A$7.A!
Code: Select all
x = 8, y = 8, rule = R10,C5,M1,S29..87,B10..10,NN
5.A$4.3A$3.A.3A$2.3A.2A$.6A$6A$.4A$3.A!
Brian Prentice
Re: Larger than Life
Posted: July 10th, 2017, 8:34 pm
by AforAmpere
A way to make n*c ships for any number n, under 25c for 3 state LTL rules:
1. Draw a n by 2n box of state 1 cells and a box of equal size under it of state two, like this for 3c:
Code: Select all
x = 6, y = 6, rule = R5,C3,M0,S0..0,B18..18,NM
6A$6A$6A$6B$6B$6B!
2. Set the rule to R(2n-1),C3,M0,S0..0,(2n^2)..(2n^2),NM
3. That's it
An example would be a 3c ship, so R5,C3,M0,S0..0,18..18,NM
Code: Select all
x = 6, y = 6, rule = R5,C3,M0,S0..0,B18..18,NM
6A$6A$6A$6B$6B$6B!
4c works the same, with rule R7,C3,M0,S0..0,B32..32,NM:
Code: Select all
x = 8, y = 8, rule = R7,C3,M0,S0..0,B32..32,NM
8A$8A$8A$8A$8B$8B$8B$8B!
25c:
Code: Select all
x = 50, y = 50, rule = R49,C3,M0,S0..0,B1250..1250,NM
50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$50A$
50A$50A$50A$50A$50A$50A$50A$50A$50B$50B$50B$50B$50B$50B$50B$50B$50B$
50B$50B$50B$50B$50B$50B$50B$50B$50B$50B$50B$50B$50B$50B$50B$50B!
Re: Larger than Life
Posted: July 11th, 2017, 1:38 pm
by dvgrn
muzik wrote:Why does it run so slow anyway? Can't you make larger than life run on hashlife instead somehow?
One of HashLife's basic assumptions is that information can't travel faster than c. 2^(N+1)-by-2^(N+1) tiles are run 2^(N-1) ticks into the future, and the 2^N-by-2^N centers are stored and re-used without recalculating them, whenever the same large hashtiles show up again. This works only because it's impossible for anything outside of the large tile to affect the center area in that amount of time.
However, any neighborhood bigger than the eight-cell Moore neighborhood immediately invalidates that key assumption. For Hashlife to work at Larger Than Life range 50, you'd have to keep hashes of every 201-by-201-cell subpattern, just to avoid having to recalculate the state of the center cell two ticks later. That quickly turns into a whole lot of memory. And even on a machine with terabytes of RAM available, caching that information probably wouldn't make the simulation any faster. HashLife's advantage shows up when it can cache results for big tiles and long time steps, not just one cell and a couple of ticks.
... Now, 201 isn’t a power of two so it wouldn’t really fit into a HashLife-ish partitioning system anyway. But this should give some idea of how impossibly big a hashtable would have to be for the high-range Larger Than Life rules, to do the kind of speedup magic that HashLife can do for Moore neighborhoods.
Long story short, Golly just plain has to do a lot more work to simulate (most) Larger Than Life rules, simply because so many more neighbor cells might possibly affect the state of each cell. There are still bit-twiddling tricks available to speed up the calculation, but they’re more QuickLife-style tricks than HashLife-style memoization.
Re: Larger than Life
Posted: July 11th, 2017, 2:15 pm
by 83bismuth38
For me it don't work for me ):
Re: Larger than Life
Posted: July 11th, 2017, 4:01 pm
by Apple Bottom
dvgrn wrote:One of HashLife's basic assumptions is that information can't travel faster than c. 2^(N+1)-by-2^(N+1) tiles are run 2^(N-1) ticks into the future, and the 2^N-by-2^N centers are stored and re-used without recalculating them, whenever the same large hashtiles show up again. This works only because it's impossible for anything outside of the large tile to affect the center area in that amount of time.
Unrelated to LtL, but-- this is a very good description of how HashLife fundamentally works, so I pilfered it for the
HashLife article on the wiki. Hope that's OK!