Re: The Hunt for a Tiny Diagonal Spaceship
Posted: March 22nd, 2016, 7:06 am
Forums for Conway's Game of Life
https://conwaylife.com/forums/
I'm traveling this week, so thanks for taking over #133-143 for me.thunk wrote:Because I have plenty of impatience and spare cpu cycles, I went ahead and searched some of the spaces reserved by others. Apologies to bullet51, dvgrn, and muzik:
#6, #133-143 is empty; #172, #174, #175 only have trivial p5+p2 oscillator combinations.
Speaking of which, biggiemac, how's this coming along?dvgrn wrote:It still seems just slightly worrisome that CopperSearch would throw away interesting things like diagonal c/8 or c/15 spaceships without reporting them. I suppose if a more efficient search program is set up to run a diagonal search with a few more bits added, then it will naturally re-search the smaller space and notice anything that was missed.
Of course there most likely isn't anything... but something new and amazing is sure to show up at some number of bits.
biggiemac wrote:I'll give it some time this weekend. I think I can build from apgnano. Dare I expand the box of interest, perhaps to 37 cells?
When I saw v0.4 had achieved a bigger speed up than I think apgnano would have given it, I refocused my efforts to running it instead of writing it. But I will maybe at least start looking at what can be built from apgmera this week.BlinkerSpawn wrote:biggiemac, how's this coming along?biggiemac wrote:I'll give it some time this weekend. I think I can build from apgnano. Dare I expand the box of interest, perhaps to 37 cells?
The life-simulating algorithm in apgmera is in the file "vlife.h" (which #includes "params.h" together with a bunch of lifeasm-* and lifelogic-* files for different instruction sets). It contains an example program which runs the methuselah Lidka for 30k generations in 45ms -- by comparison, Golly's QuickLife is about three times slower (owing to the fact that vlife is explicitly compiled for the specific rule, whereas qlife resorts to using a lookup table) and HashLife even more so.biggiemac wrote:When I saw v0.4 had achieved a bigger speed up than I think apgnano would have given it, I refocused my efforts to running it instead of writing it. But I will maybe at least start looking at what can be built from apgmera this week.
Code: Select all
for i:=1 to 41 do
begin
Step;
if Compare then
begin
if (((Offset.X <> 0) or (Offset.Y <> 0)) and (i <> 4)) or
(not (i in [1,2,3,4,5,6,8,15])) then // Change for debug
Save(EternalN,i);
Break;
end;
end;
until EternalN = MaxInd;
end;