Search found 54 matches

by rokicki
October 11th, 2019, 3:21 pm
Forum: Other Cellular Automata
Topic: Rule 30
Replies: 5
Views: 390

Re: Rule 30

Well, Rule30 is pretty fundamental, as it's the simplest CA exhibiting the type of behavior that it does. And while I don't disagree with you, the only reason I do any of this is because it's fun, not because it has any fundamental importance. For instance, in this particular case, I was able to use...
by rokicki
October 8th, 2019, 2:37 pm
Forum: Other Cellular Automata
Topic: Rule 30
Replies: 5
Views: 390

Rule 30

So because of the Rule 30 prizes that were announced https://blog.wolfram.com/2019/10/01/announcing-the-rule-30-prizes/ I've been spending too much time playing with rule 30. The left side of Rule 30 shows some patterns, so I was wondering what those looked like. The period of the leftward diagonal ...
by rokicki
October 3rd, 2019, 10:19 pm
Forum: Website Discussion
Topic: Updating website
Replies: 45
Views: 1251

Re: Updating website

Active topics doesn't seem to work. Or is it me?
by rokicki
July 18th, 2019, 12:33 pm
Forum: General Discussion
Topic: Challenge: Linear Growth
Replies: 8
Views: 1749

Re: Challenge: Linear Growth

Wow, amazing! Thanks! I can't help but giggle watching the numbers increase completely identically to astronomical values . . . this will be very useful. What's a good name for this? How did you manage to make glider constructions so quickly? I know there are databases for such things but even so . ...
by rokicki
July 17th, 2019, 8:40 pm
Forum: General Discussion
Topic: Challenge: Linear Growth
Replies: 8
Views: 1749

Challenge: Linear Growth

Here's one that might be easy, maybe not. Construct a pattern for Conway's Game of Life that, in all generations g > t (for some t), the population count p is equal to g. If that's possible, then try to minimize t. I've tried this and figured out some possible ways to do it, but they aren't particul...
by rokicki
July 16th, 2019, 10:44 pm
Forum: General Discussion
Topic: Golly 3.3b2
Replies: 36
Views: 7760

Re: Golly 3.3b2

Thanks for the excellent report (on the generation count advancing spuriously).
It was indeed exactly as Andrew has described. I've submitted a change that
will fix this.

No one was sacked over this bug.
by rokicki
October 17th, 2018, 12:09 pm
Forum: General Discussion
Topic: Golly 3.2
Replies: 53
Views: 29686

Re: Golly 3.2

On drawing outside +/- 2B: if we did decide to support it, another decision that needs to be made is whether we support drawing inside bounds of 2^63 (about +/- 9E18), or whether we truly support all editing in bignums throughout. The latter is probably a fair bit of work; the former would probably ...
by rokicki
October 16th, 2018, 4:58 pm
Forum: Scripts
Topic: LifeAPI - New GOL API (in C++)
Replies: 46
Views: 39280

Re: LifeAPI - New GOL API (in C++)

How does this compare with the algorithm test suite in

https://github.com/rokicki/lifealg/
by rokicki
May 25th, 2018, 3:07 am
Forum: General Discussion
Topic: Golly 3.2b1
Replies: 16
Views: 7027

Re: Golly 3.2b1

wildmyron wrote:Sorry, that's entirely my fault. In my defense, It did appear to be publicly available.
It is public, just didn't think anyone would notice it; no worries.
by rokicki
May 24th, 2018, 8:50 pm
Forum: General Discussion
Topic: Golly 3.2b1
Replies: 16
Views: 7027

Re: Golly 3.2b1

Would you care to expand on the improvements to HashLife and RuleLoader? Tom did all those changes so best to ask him (golly-test might be a better place -- not sure how often he checks these forums). The hashing algorithms should be somewhat faster (varies). I changed the hashtable code to use pow...
by rokicki
May 8th, 2018, 6:55 pm
Forum: General Discussion
Topic: Golly suggestions
Replies: 176
Views: 82480

Re: Golly suggestions

Allow us to set the maximum number of states available, rather than having it fixed at 256, or if not possible, set it to something high like 4098. If you can motivate us to do this by providing a compelling example of why this is interesting, it might happen. Golly has a lot of functionality, but ...
by rokicki
April 1st, 2018, 12:36 pm
Forum: Patterns
Topic: Elementary knightship
Replies: 136
Views: 224496

Re: Elementary knightship

Very impressive, Dave!
by rokicki
March 16th, 2018, 12:56 am
Forum: Scripts
Topic: Merging search programs
Replies: 20
Views: 6862

Re: Merging search programs

Oh, interesting! So it looks like JLS was pretty comparable. When I ran my tests on a 16-core box, I had 32 other threads running, so each run was about half a logical core. It would be nice to nail this down. I can run this exact search under more controlled conditions (idle box) and I can run JLS ...
by rokicki
March 15th, 2018, 8:48 pm
Forum: Scripts
Topic: Merging search programs
Replies: 20
Views: 6862

Re: Merging search programs

I performed the following experiment. I cut out an nxn square at the bottom right of Sir Robin, and I asked glucose to complete it. (I used my own Perl script to create the SAT problem, and I used Knuthian encoding). Here are timings for various n: 7 0:01.19 8 0:02.46 9 0:23.82 10 1:29.90 11 5:44.08...
by rokicki
March 14th, 2018, 7:43 pm
Forum: Scripts
Topic: zfind discussion
Replies: 237
Views: 129515

Re: zfind discussion

Can the code from zfind 1.6 be potentially addable, so larger width knightship searches could run? Sure, or I can "backport" the changes I made into zfind 1.6. The table changes are fully separable. Concurrency and dynamic table generation don't play well together so for the initial stab at multith...
by rokicki
March 12th, 2018, 4:52 pm
Forum: Scripts
Topic: Can hashlife draw a long line?
Replies: 9
Views: 3998

Re: Can hashlife draw a long line?

Just write a tiny Lua script for Golly.

There is draw-lines.lua already distributed with Golly that will let you draw lines of arbitrary
length in any rule/algorithm.
by rokicki
March 12th, 2018, 2:41 pm
Forum: Patterns
Topic: Spaceship Discussion Thread
Replies: 873
Views: 496793

Re: Spaceship Discussion Thread

Here's my best partial so far: x = 29, y = 23, rule = B3/S23 3bo$2b2o$b3o$2o2bobo$3b3o2bo$8b2o$5bo4b2obobobo$2b4ob2o6bobo$8bo3bo4bo $15b2o$9b7o$8b2o$8b2o2b5ob3o$8b4o2b4o$12bo8bo$12b3ob2ob2o$14b2o3b7o$ 15bobo3bo4bo$15bobo3bo2bo$26bo$23b3o$24b3obo$26bo! Another slightly longer partial: x = 28, y = 27,...
by rokicki
March 7th, 2018, 3:23 pm
Forum: Patterns
Topic: Elementary knightship
Replies: 136
Views: 224496

Re: Elementary knightship

Since Adam completed the knightship, he's the official discoverer, so I think he should pick the name. Getting something that almost works doesn't really count, though I appreciate his generosity in sharing credit. This ship was discovered through the use of a SAT solver, which I think is interestin...
by rokicki
March 7th, 2018, 12:20 am
Forum: Patterns
Topic: Elementary knightship
Replies: 136
Views: 224496

Re: Elementary knightship

Adam,

Now that you have your knightship, should we call you "Sir Adam"?

-tom
by rokicki
March 6th, 2018, 8:31 pm
Forum: Patterns
Topic: Spaceship Discussion Thread
Replies: 873
Views: 496793

Re: Spaceship Discussion Thread

w10 c1/8 asymmetric is negative, and surprisingly never found any partial longer than 32! And it took 20 CPU days. w11 c1/8 is finding partials (so far) of length 203 (this translates to 208/8 actual rows in a generation). Here's one: x = 10, y = 26, rule = B3/S23 7b2o$5b4o$2b5o$b4o4bo$bo4bo$4b3o$4b...
by rokicki
March 3rd, 2018, 9:39 pm
Forum: Patterns
Topic: Spaceship Discussion Thread
Replies: 873
Views: 496793

Re: Spaceship Discussion Thread

For width 11, 3c/8, even, odd, gutter, and asymmetric are all negative. For width 11, 4c/9 even, odd, gutter, and asymmetric are all negative. Longest partial among all of these is this 4c/9. x = 20, y = 42, rule = B3/S23 5bo8bo$4bobo6bobo$3bo3bo4bo3bo$2bo4bo4bo4bo$2bobobobo2bobobobo$6b3o2b 3o$b5obo...
by rokicki
March 1st, 2018, 1:27 pm
Forum: Patterns
Topic: Spaceship Discussion Thread
Replies: 873
Views: 496793

Re: Spaceship Discussion Thread

I just completed a search for asymmetric width-11 spaceships at 2c/7 and the search completed with no results. The longest partial reported is pretty long: x = 11, y = 45, rule = B3/S23 8bo$7bobo$7bob2o$5b3ob2o$4b3o$4bob2o$4bo2b2o$5bob2o$4bo2bo$4bo2bo$4bo 2b2o$2bob2ob2o$bo3bo$bo3b2o$7bo2$5b3o$2b3obo...
by rokicki
February 28th, 2018, 4:12 pm
Forum: Scripts
Topic: zfind discussion
Replies: 237
Views: 129515

Re: zfind discussion

According to http://www.conwaylife.com/wiki/LifeWiki:Spaceship_Search_Status_Page 3c/8 searches have been run at w11 for all symmetries with gfind by Paul Tooke ... Hmm, my reading of that page says they were run by knight2 (and unattributed). Am I misreading it? Asymmetric searches are now twice a...
by rokicki
February 28th, 2018, 3:05 am
Forum: Scripts
Topic: zfind discussion
Replies: 237
Views: 129515

Re: zfind discussion

Great, thanks for the change! I'll make the change conditional on Windows, or maybe I'll just embed the Mersenne twister code which should be good enough for what we need. I'd be careful letting your machine swap; you might use the R4000 option to limit max memory consumption to 4GB (the program wil...
by rokicki
February 27th, 2018, 9:17 pm
Forum: Scripts
Topic: zfind discussion
Replies: 237
Views: 129515

Re: zfind discussion

All I did to speed up the search was to put a cache in front of the lookAhead routine. Sometimes the lookAhead routine can spend a lot of time (relatively speaking) trying to decide if it's worth exploring a branch. I hash the relevant parameters and use a direct-mapped cache to see if I've seen tho...