|
Dieting Hippo posted:edit: Got sub-3000 cycles on TEC REDSHIFT  I love doubling the size of my program to save literally one single cycle. Edit: I feel kind of dirty about this, but...  Same basic solution, now with optimizations from looking at the test cases. (For anyone else trying: the lowest password that actually occurs is 002 and the highest is 990 Snake Maze fucked around with this message at 16:38 on Aug 14, 2018 |
#
?
Aug 14, 2018 15:30
|
|
|
|
| # ? Jun 7, 2024 00:38 |
|
|
Snake Maze posted:(For anyone else trying: the lowest password that actually occurs is 002 and the highest is 990 You can save even more cycles if you don't hold yourself to those but realise you can loop around after 1000. Starting a bit higher than the lowest can save a bunch of cycles still. 
|
|
#
?
Aug 15, 2018 03:18
|
|
|
New Exapunks update with an important bugfix:quote:Fixed a bug where you could use M in global mode to communicate across non-existent links, like in the "modem" puzzles. This will invalidate some players' solutions to these puzzles, so the completion flags, leaderboards, and histograms for them have been reset.
|
|
#
?
Aug 15, 2018 05:04
|
|
|
gently caress you too Zach
|
|
#
?
Aug 15, 2018 05:22
|
|
|
lol this definitely broke at least one of my solutions, because I was doing exactly that. On the other hand, I'm sure there's some cool puzzle design you could do where having multiple distinct global M networks could come in handy.
|
|
#
?
Aug 15, 2018 06:19
|
|
|
Welp, my tiny neat Pager solution is dead then.
|
|
#
?
Aug 15, 2018 11:27
|
|
|
ymgve posted:Welp, my tiny neat Pager solution is dead then.
|
|
#
?
Aug 15, 2018 11:52
|
|
|
I was honestly surprised this worked in the first place when I got to the radio station one. What a weird thing to overlook until solution stats started rolling in.
|
|
#
?
Aug 15, 2018 13:37
|
|
|
I've kinda hit a wall on the hack the highway sign puzzle - could anyone give me a soft hint in the right direction?!
|
|
#
?
Aug 15, 2018 20:14
|
|
|
You don't need to count both rows and columns. DIVI and MODI will be your friends in your time of need.
|
|
#
?
Aug 15, 2018 20:18
|
|
|
Chev posted:
Maybe I misremembered. It's not working like that now. I'm using TEST MRD instead.
|
|
#
?
Aug 15, 2018 20:37
|
|
|
Chev posted:You don't need to count both rows and columns. DIVI and MODI will be your friends in your time of need. what does MODI do?!
|
|
#
?
Aug 15, 2018 21:00
|
|
|
Shadski posted:what does MODI do?!
|
|
#
?
Aug 15, 2018 21:01
|
|
|
I wish the game’s plot revolved around your histogram results because I’ve had some really shameful solutions that pivot on general purpose unoptimized for test cases self-replication. “Hey the heart beats IAW the spec good enough eh just hope I’m never near something that reads an M or I die”. Why have anything more than one loop that spawns single-task EXAs 
|
|
#
?
Aug 16, 2018 03:08
|
|
|
Dieting Hippo posted:it should be possible to cobble together a Bitmap-to-EXA generator I might have taken Redshift a bit too far.  I threw together a program that'll take in a Monochrome BMP and spit out EXA code. At some point I'll throw it up on GitHub so other Redshift tinkerers can have an easier time. Here's EXAs loading data generated by the tool: https://www.youtube.com/watch?v=3AcAYCmNRNI
|
|
#
?
Aug 16, 2018 12:02
|
|
|
and as he stood atop the hack*match leaderboards he wept, for there were no more punks left to conquer
|
|
#
?
Aug 16, 2018 12:09
|
|
|
Hah. My solutions to the first two modem puzzles were still valid after the patch, and the pager puzzle wasn't too hard to set up a synchronized delay loop.
|
|
#
?
Aug 16, 2018 12:45
|
|
|
Zetsubou-san posted:and as he stood atop the hack*match leaderboards he wept, for there were no more punks left to conquer 
|
|
#
?
Aug 16, 2018 18:54
|
|
|
good thing i'm not friends with that punk then! 
|
|
#
?
Aug 16, 2018 20:29
|
|
|
SupSuper posted:I got bad news for you. oof
|
|
#
?
Aug 16, 2018 21:47
|
|
|
My zines arrived! And now the question becomes if I should open the envelopes, or leave them intact, collector's item-ish.
|
|
#
?
Aug 16, 2018 22:33
|
|
|
Neat, had a brainstorm on the wonderdisk, beat the TOP_PERCENTILE by over a 1000 cycles.
|
|
#
?
Aug 16, 2018 23:49
|
|
|
ymgve posted:My zines arrived!
|
|
#
?
Aug 17, 2018 00:47
|
|
|
I know I write mostly bloated solutions (they're highly maintainable ) but there are some god drat wizards in this forum. How the gently caress do you get 179 cycles on the EXA-BLASTER radio station playlist hack? Dieting Hippo posted:I might have taken Redshift a bit too far. 
|
|
#
?
Aug 17, 2018 01:04
|
|
|
 Make sure your volume is turned down, redshift can be loud with 3 channels going
|
|
#
?
Aug 17, 2018 02:15
|
|
|
I'm looking forward to playing this https://steamcommunity.com/id/IShallRiseAgain
|
|
#
?
Aug 17, 2018 02:57
|
|
|
 Man, it gets pretty rough once you get above 50k.
|
|
#
?
Aug 17, 2018 03:05
|
|
|
I'm liking the way you can learn from seeing other sokuyions in the hacker battles, also the way that sometimes you have to write a different solution for a particular opponent, but that can so reveal a better general solution. This game, it's good.
|
|
#
?
Aug 17, 2018 15:26
|
|
|
Someone clue me in on the trick to lower cycles on Tutorial 4? Unrolling my SUBI X 1 X, COPY X F loop, with a final loop to fix it if we went negative, still only gets me in the high 200s.
|
|
#
?
Aug 17, 2018 22:29
|
|
|
Here's a hint: SUBI X 0 F SUBI X 1 F SUBI X 2 F ...
|
|
#
?
Aug 17, 2018 22:53
|
|
|
ymgve posted:Here's a hint: Perfect. Thanks!
|
|
#
?
Aug 18, 2018 00:42
|
|
|
Just did Equity First Bank, the one with the ATMs. I thought my solution that hit all the ATMs with one EXA was pretty good, but my cycles are at 4000+ and everyone else on the leaderboard is at ~1100. I feel like I'm probably missing something basic that's making me unoptimized...
|
|
#
?
Aug 18, 2018 02:48
|
|
|
CharlieFoxtrot posted:Just did Equity First Bank, the one with the ATMs. I thought my solution that hit all the ATMs with one EXA was pretty good, but my cycles are at 4000+ and everyone else on the leaderboard is at ~1100. I feel like I'm probably missing something basic that's making me unoptimized... Are you checking every cycle to see if the ATM is empty? It sounds like you're checking every cycle to see if the ATM is empty. Don't do that. I haven't tried too hard at the TV station hacking game, but I was pretty shocked that I could beat everyone except the few wizards on my friends list just by flooding the central room with KILL commands for like 40 cycles, and then half assedly try to move my two files.
|
|
#
?
Aug 18, 2018 03:07
|
|
|
GuavaMoment posted:Are you checking every cycle to see if the ATM is empty? It sounds like you're checking every cycle to see if the ATM is empty. Don't do that. Is there a limit to how many opponents you can get on the competitive levels? I see your score on the ATM level, but not in the opponent list for the TV station one.
|
|
#
?
Aug 18, 2018 15:29
|
|
|
GuavaMoment posted:Are you checking every cycle to see if the ATM is empty? It sounds like you're checking every cycle to see if the ATM is empty. Don't do that. Well, that makes sense. The code looks worse to me but it runs faster! Lesson learned
|
|
#
?
Aug 18, 2018 15:48
|
|
|
CharlieFoxtrot posted:Just did Equity First Bank, the one with the ATMs. I thought my solution that hit all the ATMs with one EXA was pretty good, but my cycles are at 4000+ and everyone else on the leaderboard is at ~1100. I feel like I'm probably missing something basic that's making me unoptimized... Play to the histogram. The histograms you see rate each solution on every metric. Solutions optimized for one metric generally are not optimized for the others. It’s the fast/cheap/good triangle. Then play to the test scenario. I’m guessing you’re replicating an EXA then sending it to nodes 0-1000. What range of nodes are in the tests? Bobcats fucked around with this message at 16:20 on Aug 18, 2018 |
|
#
?
Aug 18, 2018 16:17
|
|
|
 Gah! Edit: I''m never gonna finish the campaign at this rate. EditEdit:  WOO, I am the 0.1%
Broken Cog fucked around with this message at 22:14 on Aug 18, 2018 |
|
#
?
Aug 18, 2018 18:30
|
|
|
Guy Axlerod posted:Is there a limit to how many opponents you can get on the competitive levels? I see your score on the ATM level, but not in the opponent list for the TV station one. I had Guava on KGOG (uploaded a solution that beat the killbots) but don't have them there today. Everyone else is still there from what I can tell.
|
|
#
?
Aug 18, 2018 22:00
|
|
|
Would be nice if you could automatically run your solution against every opponent.
|
|
#
?
Aug 18, 2018 22:20
|
|
|
|
| # ? Jun 7, 2024 00:38 |
|
|
Just opened the envelope containing my second zine, feels pretty good. Also rad 3d glasses
|
|
#
?
Aug 18, 2018 23:10
|
|