0E0P constructions

[calcyman]

-- Meant to ask: do you still happen to have the once-per-hour saved snapshots of your one-metatick run for a single cell, especially near the billion ticks that you mentioned? I suppose the whole collection might be useful for generating a routing blueprint.

I do, yes. I'm now running one of the backups so that I can get a snapshot just 500 ticks prior to the glider deleting the eater2; then, you'll be able to run that snapshot in Golly at a low step size to pinpoint the minimum population.

[dvgrn]

I'm now running one of the backups so that I can get a snapshot just 500 ticks prior to the glider deleting the eater2; then, you'll be able to run that snapshot in Golly at a low step size to pinpoint the minimum population.

Looks like I'm allowed to attach files this big:

rrocell_pop.mc.gz

0E0P metacell fast-forwarded to (relatively) near a likely minimum

(6.66 MiB) Downloaded 356 times

According to an email from calcyman, the attached snapshot is just before the switch to construction mode happens, as promised above. Once gliders start getting duplicated, the population will never go back to anywhere near the minimum.

At T=495 the switch is technically opened, via a glider removing an eater2 in the S corner:

Code: Select all

x = 615, y = 375, rule = B3/S23
177b2o$177b2o35$498bo$497bobo$498b2o49$354bo191b2o$353bobo190b2o7b2o$
354b2o199bo$553bobo$553b2o6bo$559b3o$558bo$348b2o208b2o$348b2o2$557b2o
$557b2o$540b2o$539bobo22b2o$539bo24bobo$538b2o25bo7$536bo7b2o$354b2o
179bobo7bo$354b2o179b2o5b3o$542bo$579b2o$578bo2bo$347b2o230bobo$347b2o
211b2o18bo$560bobo$562bo$562b2o$535bo$534bobo7b2o$534bo2bo6b2o$535b2o
3$552b2o34b2o$543b2o7b2o33bo2bo$544bo43bobo$544bobo22bo19bo$545b2o21bo
bo$568bobo$403b2o164bo$403b2o154b2o20bo9bo$559bobo18bobo7bobo$561bo18b
o2bo5bo2bo$561b2o18b2o7b2o4$583bo$583b3o$586bo$585b2o$600b2o$600bo$
597b2obo$596bo2bo$409b2o186b2o$409b2o156b2o13b2o$567b2o13b2o3$312b2o
88b2o$312bo89b2o$314bo228b2o$294b2o14b5o228b2o$295bo13bo259b2o6bob2o$
295bobo12b3o256bobo3b3ob2o$296b2o15bo257bo2bo$310b4o229b2o26b2o2b3ob2o
$305b2o3bo3b2o227b2o32bobo$305b2o4b3o2bo240b2o18bobo10b2o$313bob2o240b
obo18bo11b2o$313bo245bo$312b2o235b2o8b2o$550bo$547b3o$304b2o241bo$304b
o307b2o$305b3o304bobo$307bo305bo20$397bo$396bobo$397b2o6$347b2o$347b2o
19b2o$368b2o3$385b2o$385b2o$352b2o$352b2o8$452b2o4bo$452b2o3bobo4b2o$
456bo2bo4b2o$457b2o$428b2o$391b2o34bo2bo$391b2o34bobo$428bo26bo$455b3o
$458bo$384b2o71b2o$384b2o86b2o$472bo$469b2obo$468bo2bo$469b2o$439b2o$
370b2o44b2o21b2o$370b2o44b2o5$374b2o$374b2o73bob2o$447b3ob2o19b2o$446b
o24bobo$447b3ob2o19bo$449bobo$422b2o8b2obo13bobo10b2o$423bo8b2ob3o12bo
11b2o$420b3o15bo$420bo11b2ob3o$433bobo$433bobo$434bo2$76bo$75bobo$76b
2o8$477b2o$477bobo$478bo4$109bo$108bobo$109b2o5$103b2o$103b2o15$109b2o
$109b2o4$102b2o$102b2o12$21b2o$2o19b2o$2o5$16b2o$16b2o4$80b2o$80bo$82b
o$62b2o14b5o$63bo13bo$63bobo12b3o$64b2o15bo$78b4o$73b2o3bo3b2o79bo$73b
2o4b3o2bo77bobo$81bob2o77b2o$81bo$80b2o3$72b2o$72bo$73b3o$75bo18$381b
2o$380bobo$382bo13$112b2o$112bobo$113bo!

See (68499, 65162) in the full pattern in Golly for the location of this one-time switch.

Oddly enough, the minimum doesn't appear immediately after the eater2 is removed. The loss of the glider and eater2 aren't very noticeable in the short term; the signal gets swamped by the huge bank of Snarks reflecting gliders.

Up to T = 50,000 there are several local minima, with the lowest being slightly below 18,650,000 cells somewhere between 13,000 and 15,000. (This evening I'll edit in the exact numbers that showed up in getminpopandminbb.py.)

EDIT: The local minimum I saw was 18649387 at T=13434 -- but new local minima keep showing up, most recently 18649378 at T=56093 followed by 18649352 at T=71004 (with maybe a few intermediate local minima that I didn't see). Seems like the real minimum might be somewhere near there -- all but the last two digits, maybe?

EDIT: Nope, now there's population=18649178 at T=105800, and nothing has beaten that for another 50,000+ ticks. There will be a few more possible minimum points somewhere near T=105800+262144N, I think, for single-digit N, and another minimum might still show up before that. Not sure if I'll get bored with running this population check before T = 2 million, but if I do I'll post the last checkpoint that I reach.

A lower global minimum could happen any time before T=2,000,000 or so, when duplicate gliders start escaping past that missing eater2. At the current phase the single-channel recipe gliders are still working their way through the last ten U-turns in the northeast edge of the memory loop, so they have a long way to go before they're out of the minpop danger zone.

-- I believe the attached is the metacell programmed for B2-an3-eiky4aiqw5ijnr6ak/S012ik3-acir4kqw5acq6ack7c8. Some clever gene editing on the southwest side might produce an S0 metacell or a MAPAC[3A]I[15A]I[64A]* metacell -- but I haven't yet tracked down what still lifes are currently present or absent to encode the metacell's state (7), so that might have to be changed also to find a more minimal minimum.


* Why didn't we think to RLE-encode MAP encodings, anyway? -- either pre- or post-conversion to base64. We seem to end up with a lot of MAP rules that are mostly A's.

[rowett]

[Why didn't we think to RLE-encode MAP encodings, anyway? -- either pre- or post-conversion to base64. We seem to end up with a lot of MAP rules that are mostly A's.

Good question. The answer is because I was expecting that MAPs would mostly be machine (script) generated. If I'd wanted them human read/writeable I'd have encoded them in hex rather than base64 and probably picked a different bit order.

[dvgrn]

Now there's population=18649178 at T=105800, and nothing has beaten that for another 50,000+ ticks. There will be a few more possible minimum points somewhere near T=105800+262144N, I think, for single-digit N, and another minimum might still show up before that. Not sure if I'll get bored with running this population check before T = 2 million, but if I do I'll post the last checkpoint that I reach.

Between 250,000 and 400,000 ticks the minimum is 18649268 at T=375768, so there's nothing that beats the minimum at T=105800 yet. If anyone wants to check the remaining 1,600,000 ticks, here's the T=400,000 checkpoint:

rrocell_pop_400000.mc.gz

checkpoint at T=400,000 (where T=0 is 495 ticks before the key eater2 is destroyed)

(6.65 MiB) Downloaded 329 times