« first day (586 days earlier)      last day (863 days later) » 

00:37
Not sure if you guys are already aware, but your work was mentioned in the latest Stack Overflow Podcast episode!
4
They start talking about it at 46:50.
@Oliver I didn't know that! That might well be the first podcast I ever listen to (and even then only for a short time :P).
@El'endiaStarman :D that particular episode wasn't that great until they started talking about the challenge.
Pretty great hearing them talk about it. :D
They did get one detail wrong though. They took Mego saying that he had to write out the full GoL computer bit by bit as meaning that we had to simulate the GoL computer bit by bit. We never actually simulated even a single generation of the full computer, I think.
@El'endiaStarman It may be worth leaving a comment or Tweeting them @StackPodcast.
01:03
@El'endiaStarman Mego ran it in Hashlife, no?
Yeah, I'm fuzzy on that.
I'm looking at the traffic stats for the GCC backend and almost no one is even looking at the code or cloning it -- they're just seeing "qftasm" and "GCC" in the repo title and going "zomg half life 2 episode 3 in GOL confirmed" and staring it XD
I'm seriously tempted to fork linux to "qftasm-linux" but not change it and see if anyone notices
I don't understand why something that doesn't compile and hasn't been posted anywhere is getting 13 stars I guess. GH stars are more mysterious than chat stars
01:24
I am tempted to star that comment
 
3 hours later…
04:00
Bad news: There's still a lot to go... and stuff is already starting to look like spaghetti.
Impressive nonetheless.
@quartata If the computer hits and unused instruction, I think (I haven't checked) that it would write the value 0 to the destination (argument 3). The computer would run normally aside from that.
@KZhang OK. I can probably work with that for version detection
Anonymous
04:43
@quartata I did run it in Hashlife. It took a few minutes to simulate a single instruction, even with a large step size.
Anonymous
And by "a few" I mean like 20
Well I mean we know it's not fast.. :P
@Mego Ah, I need to correct the comment I made on the podcast...
But we did simulate it outside of Varlife
Anonymous
Mainly I was checking for parity between the GoL and the Varlife versions. After waiting ~20 minutes for one instruction, I called it good enough :P
10:51
@KZhang What's the total number of different 8m by 8n logic tiles that would still need to be rewired to replace the 11x11 based ones?
I think the Hashlife-friendly metacell will turn out to be period 4096, though I shouldn't say that until I've done more of the neighbor wiring. Definitely 512x512, anyway.
Between that and the 8x8 rework, Golly should run the computer orders of magnitude faster.
 
4 hours later…
14:34
P4096? Deep down I knew I was being way too optimistic. I still have some hopes that it will work at p8192, but of course we can just keep doubling till there's enough time for everything (and Golly will still run it just as fast). conwaylife.com/forums/viewtopic.php?f=&p=51330#p51330
14:58
Hey guys, someone wants to publish our work in a "journal". github.com/QuestForTetris/WriteUp/issues/6
I've taken a quick look at one of their issues, and I'm iffy about it.
@DaveGreene Nice work. Is the red box a 512x512 area?
15:20
@El'endiaStarman Yup -- or a 512x512 grid anyway, so a 513x513 box.
The circuitry can still be packed quite a bit tighter, and otherwise the only missing required piece is the wiring connecting the metacell to each neighbor. So there should be something like a 256x256 area left where a metapixel piece can be added.
 
6 hours later…
Anonymous
21:00
@El'endiaStarman It seems to be a semi-serious journal written in an archaic, comedic fashion. I don't know how much extra attention it will give us, but I'm also not sure that it will hurt.

« first day (586 days earlier)      last day (863 days later) »