←2020-05-13 2020-05-14 2020-05-15→ ↑2020 ↑all
00:07:23 -!- rodgort has joined.
00:25:19 -!- zseri has quit (Quit: zseri).
01:16:38 -!- jonathanr has joined.
01:17:34 <jonathanr> Hello, does anyone used AlphaBeta before?
01:22:04 <int-e> Hmm, doesn't look too interesting
01:24:56 <jonathanr> What languages did you use before?
01:29:57 <shachaf> I used alpha-beta pruning before.
01:34:33 <int-e> Did it help?
01:38:13 <shachaf> Yes! It was great.
01:38:22 <shachaf> Do you understand PVS/Scout?
01:38:37 <shachaf> I would like to improve my understanding of the whole null window search deal.
01:39:05 <int-e> No I don't.
01:40:59 <int-e> I could barely expand PVS without looking it up. My first thought was what theorem provers have to do with all this.
01:47:40 <jonathanr> shachaf How did you use Alpha Beta without the use of comments?
01:48:23 <int-e> I did rewrite my cellular automaton update function for Ponder This in assembly though, just for laughs... it ended up 20% faster than the corresponding C, which I think is decent nowadays.
01:48:48 <shachaf> Did you do some sort of fancy SIMD thing?
01:48:50 <zzo38> In assembly langauge for what instruction set?
01:48:50 <int-e> > 1/1.2
01:48:52 <lambdabot> 0.8333333333333334
01:48:56 <int-e> no SIMD at all
01:49:23 <int-e> or is that 16% faster...
01:49:50 <jonathanr> 16 and 2/3
01:50:02 <int-e> jonathanr: it's not that precise
01:50:08 <int-e> or accurate
01:50:47 <int-e> > 21/25.5
01:50:49 <lambdabot> 0.8235294117647058
01:50:54 <int-e> > 25.5/21
01:50:56 <lambdabot> 1.2142857142857142
01:51:12 <jonathanr> Using fractions is more precise than decimal, unless I find a way to write ­­¯ over the numbers
01:51:18 <int-e> (those are the approximate cycle counts)
01:51:34 <int-e> shachaf: I did think about it but I don't think it'll pay off.
01:52:13 <int-e> shachaf: The proper way to do SIMD is to move it to a GPU, which will be much slower for an individual update but make up for it by massive (compared to CPU) parallelization.
01:52:33 <shachaf> Sure.
01:53:09 <shachaf> Do you like this? where this = alpha-beta pruning
01:53:20 <int-e> `? this
01:53:22 <HackEso> This is something people on the channel like to talk about. We're often unsure what this is, though. Nobody likes this.
01:55:49 <int-e> shachaf: The trouble is, 128 bit shifts are not a good fit for Intel's SIMD stuff. With x86_64 you have those cute shld/shrd instructions. To do that in SSE, you need to shift, then use the weird byte-wise shift on top, then shift in the other direction, and combine the results. And I think the CPU can issue twice as many integer operations as SSE instructions.
01:56:35 <int-e> But I didn't try.
01:57:34 <shachaf> It's great how you use these tricks to do an exhaustive search of a game tree to depth 13 or however much.
01:57:38 <shachaf> Such a big tree.
01:57:46 <int-e> (gcc knows how to use the shld/shrd instructions, too. The speedup I have is mostly from better instruction scheduling... I save a few mov instructions as well, but those are essentially free.)
01:58:40 <int-e> Those modern CPUs are such beasts.
02:00:46 <int-e> not %rax; or %rbx, %rax <-- as far as I understand it, this fuses into a single micro-op. My first instinct was to move the `not` instructions further up to get rid of a data dependency and stall... and that made the code slower.
02:02:19 <zzo38> I don't really like the way they do it so much, though. I like MMIX instruction set.
02:02:24 <int-e> And the CPU will reorder instructions on its own as well...
02:02:54 * int-e shrugs.
02:03:07 <int-e> Nobody uses MMIX though.
02:04:13 <int-e> "the CPU" is an i7-6850K; I expect Ryzen does similar things. The low power models are less crazy. ARM as well, I suppose.
02:04:27 <int-e> (That is, ARM is less crazy as well, I suppose.)
02:05:26 <zzo38> There is not suitable hardware for MMIX made yet (other than a few FPGA stuff which is incomplete, as far as I know), although there are software implementations of MMIX that can measure the number of cycles.
02:07:46 <int-e> TBH I have not even read the MMIX document, because I hardly see the point.
02:09:22 <zzo38> Well, I think that it is a better instruction set than ARM and some other stuff. However, there is a few things I would have done differently, such as I would have explicit caching only and no implicit caching, and I would also have a "Muxcomp" operator.
02:57:48 -!- adu has joined.
03:47:18 -!- tromp has joined.
03:51:13 <salpynx> irony; I just realised that I had been "incorrectly" thinking of RBF as a group but actually had the concept of monoid in mind. Turns out RBF _is_ a group and I was making two errors that happened to be the inverse of each other...
03:51:47 -!- tromp has quit (Ping timeout: 265 seconds).
03:52:45 <esowiki> [[Dig]] https://esolangs.org/w/index.php?diff=72251&oldid=72189 * Emerald * (+69) Not associated with DigFill TM
04:06:57 -!- jonathanr has quit (Remote host closed the connection).
04:33:50 -!- adu has quit (Quit: adu).
04:41:23 -!- tromp has joined.
04:46:23 -!- tromp has quit (Ping timeout: 272 seconds).
04:48:32 -!- salpynx has quit (Remote host closed the connection).
05:11:42 -!- Antebrationist has joined.
05:14:11 -!- Antebrationist has quit (Remote host closed the connection).
05:14:19 -!- ArthurStrong has quit (Quit: leaving).
05:23:47 <esowiki> [[User:Programmer5000]] https://esolangs.org/w/index.php?diff=72252&oldid=52076 * Programmer5000 * (-250) Blanked the page
05:25:39 <esowiki> [[Meta Memes]] https://esolangs.org/w/index.php?diff=72253&oldid=52068 * Programmer5000 * (-89)
05:35:43 -!- tromp has joined.
05:40:17 -!- tromp has quit (Ping timeout: 244 seconds).
06:18:03 -!- tromp has joined.
06:21:30 -!- rain1 has joined.
06:35:38 <zzo38> Now I think the rendering functions in TeXnicard are fully implemented (preview, render to RGB, and render to separations output format). However, typesetting is not yet implemented, and I may also want to add a few more effects for rendering.
06:36:47 <zzo38> If someone will provide graphics for templates then I can make up the templates involving those graphics.
07:05:26 -!- rain1 has quit (Ping timeout: 272 seconds).
07:06:28 -!- rain1 has joined.
07:47:46 -!- kspalaiologos has joined.
08:45:06 -!- LKoen has joined.
09:27:13 -!- zzo38 has quit (Read error: Connection timed out).
09:30:58 -!- Lord_of_Life has quit (Ping timeout: 260 seconds).
09:31:12 -!- Lord_of_Life_ has joined.
09:32:31 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
09:57:54 -!- gnu-nobody has joined.
10:08:14 <esowiki> [[Bigspace]] https://esolangs.org/w/index.php?diff=72254&oldid=68031 * DmilkaSTD * (-106)
10:29:58 -!- imode has quit (Ping timeout: 246 seconds).
10:42:10 -!- arseniiv has joined.
11:05:57 <esowiki> [[User talk:Javamannen]] https://esolangs.org/w/index.php?diff=72255&oldid=31617 * DmilkaSTD * (+130)
11:26:19 -!- kspalaiologos has quit (Read error: Connection reset by peer).
11:26:48 -!- kspalaiologos has joined.
11:46:58 -!- atslash has quit (Ping timeout: 256 seconds).
11:47:58 -!- atslash has joined.
12:32:28 <esowiki> [[Bigspace]] M https://esolangs.org/w/index.php?diff=72256&oldid=72254 * PythonshellDebugwindow * (-9)
13:35:54 -!- adu has joined.
14:09:58 <esowiki> [[VTFF]] M https://esolangs.org/w/index.php?diff=72257&oldid=52077 * PythonshellDebugwindow * (+3289) /* Notes */
14:16:37 <esowiki> [[6969 Assembler]] https://esolangs.org/w/index.php?diff=72258&oldid=72067 * CMinusMinus * (+220) /* JMP */
14:25:28 <kspalaiologos> > The goal is to create a language hard enough so no usable program can be created with it in reasonable time
14:25:30 <lambdabot> <hint>:1:90: error: parse error on input ‘in’
14:25:31 <kspalaiologos> am I a joke to you?
14:29:12 -!- gnu-nobody has quit (Ping timeout: 260 seconds).
14:31:47 <esowiki> [[6969 Assembler]] https://esolangs.org/w/index.php?diff=72259&oldid=72258 * CMinusMinus * (+53)
14:37:08 -!- kspalaiologos has quit (Quit: Leaving).
14:37:55 -!- kspalaiologos has joined.
14:45:15 <esowiki> [[6969 Assembler]] https://esolangs.org/w/index.php?diff=72260&oldid=72259 * CMinusMinus * (+94) /* CMP */
14:45:24 <esowiki> [[6969 Assembler]] https://esolangs.org/w/index.php?diff=72261&oldid=72260 * CMinusMinus * (-1) /* CMP */
14:46:56 -!- esomaker has joined.
14:48:37 <esowiki> [[6969 Assembler]] https://esolangs.org/w/index.php?diff=72262&oldid=72261 * CMinusMinus * (+165)
14:51:19 <esowiki> [[6969 Assembler]] https://esolangs.org/w/index.php?diff=72263&oldid=72262 * CMinusMinus * (+1)
14:51:33 <arseniiv> topology is pointless, probability is free and logic is categorical
14:54:22 <arseniiv> or maybe “topology is pointless, logic is categorical, only probability is free”
14:55:53 <esowiki> [[6969 Assembler]] https://esolangs.org/w/index.php?diff=72264&oldid=72263 * CMinusMinus * (+462) /* Memory variables */
14:57:18 <esowiki> [[6969 Assembler]] M https://esolangs.org/w/index.php?diff=72265&oldid=72264 * PythonshellDebugwindow * (-2) /* Beginners explanation */
15:14:01 -!- esomaker has quit (Remote host closed the connection).
15:27:04 -!- gnu-nobody has joined.
15:32:04 -!- opticnerve has joined.
15:36:06 <esowiki> [[Ports]] N https://esolangs.org/w/index.php?oldid=72266 * AnimaLibera * (+10) create the Ports page
15:56:42 -!- adu has quit (Quit: adu).
15:57:05 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72267&oldid=72266 * AnimaLibera * (+93) add categories
16:06:07 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72268&oldid=72267 * AnimaLibera * (+320) Adding stuff
16:07:44 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72269&oldid=72268 * AnimaLibera * (+0) e
16:18:24 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72270&oldid=72269 * AnimaLibera * (+3120) add the Definitions and Execution paragraphs
16:19:00 -!- catern has quit (Quit: catern).
16:22:29 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72271&oldid=72270 * AnimaLibera * (+72) fancy definitions
16:25:08 -!- imode has joined.
16:25:40 <esowiki> [[User:AnimaLibera]] M https://esolangs.org/w/index.php?diff=72272&oldid=71487 * AnimaLibera * (+49) add my new esolang: Ports
16:29:37 -!- catern has joined.
16:34:47 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72273&oldid=72271 * AnimaLibera * (+308) add nop
16:40:04 -!- adu has joined.
16:43:01 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72274&oldid=72273 * AnimaLibera * (+8589) add instructions and syntax
16:49:13 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72275&oldid=72274 * AnimaLibera * (+1470) add Hello world
16:52:39 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72276&oldid=72275 * AnimaLibera * (-13) better looking hello world
16:55:23 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72277&oldid=72276 * AnimaLibera * (+485) add Cat
16:56:06 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72278&oldid=72277 * AnimaLibera * (-10) not a Stub anymore I think
16:57:30 <esowiki> [[Language list]] M https://esolangs.org/w/index.php?diff=72279&oldid=72248 * AnimaLibera * (+12) add Ports
17:01:49 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72280&oldid=72278 * AnimaLibera * (+107) improve intro
17:03:10 -!- gnu-nobody has quit (Ping timeout: 256 seconds).
17:11:54 -!- FreeFull has quit (Quit: rebooting).
17:12:57 -!- FreeFull has joined.
17:14:43 -!- kspalaiologos has quit (Ping timeout: 256 seconds).
17:15:20 -!- kspalaiologos has joined.
17:19:27 -!- zzo38 has joined.
17:38:58 <Lymia> !ztest alive -.+--)*-1
17:39:08 <Lymia> guess that is well and truly dead :p
17:57:55 <esowiki> [[Esofun]] https://esolangs.org/w/index.php?diff=72281&oldid=72230 * Palaiologos * (+26)
18:26:31 -!- LKoen_ has joined.
18:27:43 -!- LKoen_ has quit (Read error: Connection reset by peer).
18:28:32 -!- gnu-nobody has joined.
18:29:39 <pikhq> Yup
18:57:56 -!- kspalaiologos has quit (Ping timeout: 272 seconds).
18:58:39 -!- kspalaiologos has joined.
19:14:37 <b_jonas> kspalaiologos: so about komalpsz. part of the problem is just that it may have been too ambitious a project for my weak skills back then, so I never finished it. I never even implemented function/subroutine calls, even though that would have made more programs runnable. I'm not sure I understood back then how to implement function calls with parameters and local variables in an interpreter.
19:15:25 <b_jonas> But the bigger problem is that there are parts where I kept modifying the source codes of psz programs to match the interpreter, rather than working harder to implement the syntax that the existing programs used.
19:15:36 <b_jonas> At least I had the sense to mark with comments where I modified the programs, but still.
19:16:27 <b_jonas> For example, many psz programs are missing keywords to mark the end of if/while statements. However, they do always have newlines and indentation determine where each if/while ends, so I should have modified the parser to ignore those end keywords and look at the indentation.
19:17:20 <b_jonas> I also didn't implement print statements that print literal strings and variables in the same statement, even though multiple psz programs use that, instead I just modified those statements in the programs. That's rather silly.
19:17:56 <b_jonas> In the end there are only like five programs that the interpreter can actually run.
19:18:23 <b_jonas> The coding style of the interpreter sucks, but then, so did the coding style of every program that I wrote back then.
19:18:37 <myname> what is psz
19:19:18 <b_jonas> myname: It's an interpreter. (1) You don't want to know. (2) If you really want to know, "http://www.math.bme.hu/~ambrus/pu/psz-snapshot.tgz" .
19:22:59 <b_jonas> also because of this, I never really did anything with the results of the runs
19:23:06 <b_jonas> so the goal failed too
19:23:26 <Lymia> What even is this?
19:24:18 <Lymia> This looks foreign language but not *extordinarily* esoteric
19:26:47 <b_jonas> Lymia: it's about KöMaL's mail-in programming competition between 2001-09 and 2003-01, to which the organizers released example solutions (after the deadline) in this nonsense language with no consistent syntax rules,
19:27:32 <b_jonas> presumably because if they released solutions in a production language that people can just get a compiler for (like pascal or C; they did occasionally did that for a few tasks), anyone could run their example solution and see that it's broken.
19:28:28 <b_jonas> So I decided to try to run those example solutions again to see that they are indeed completely broken, darn the obfuscated language they're written it, and indeed found that at least the example solution for I.10. is broken.
19:28:35 <b_jonas> But I never really got far with it.
19:29:09 <b_jonas> the tasks and solutions are linked from https://www.komal.hu/verseny/korabbi.h.shtml with the relevant dates,
19:29:23 <Lymia> language barriers yay
19:29:45 <b_jonas> plus the tasks are also linked from https://www.komal.hu/verseny/korabbi.e.shtml in English translation
19:30:05 <b_jonas> the solutions aren't, but at least the source codes for the solutions shouldn't require translation, even if the explanations do
19:30:46 <b_jonas> at least one of the early solutions that they released in Pascal is broken too by the way, which may have discouraged them from releasing real runnable programs
19:31:42 <b_jonas> and I don't mean broken as in it has a few portability problems or it fails in some weird corner case, but that the program source code is entirely not powerful enough to do what the problem asked for
19:41:43 <Lymia> ... eek.
19:42:47 <b_jonas> Note that this was back between 2001-09 and 2003-01. Me and everyone else were very young back then. And it's not like they had much money to pay to the college students that they roped in to write most of those problems and solutions.
19:42:53 <b_jonas> So there are excuses.
19:43:14 <b_jonas> Believe me, the programs I write these days are better. They're not perfect of course, but they're buggy in more advanced way,
19:43:25 <b_jonas> and I managed to write larger programs that are still usable.
19:44:15 <b_jonas> Plus the internet was young, there weren't many opportunities to learn good practices from others yet.
19:45:56 <b_jonas> Also KöMaL had a homepage where they listed a copy of those programs and solutions (besides the printed journal), but it only added a forum to the homepage in 2003-11 apparently, so I couldn't really talk to many people about those example solutions either.
19:51:27 <gnu-nobody> b_jonas: isn't I.10 just a simple convex hull?
19:52:12 <b_jonas> gnu-nobody: it's supposed to be, yes. but the example solution is broken.
19:53:13 <gnu-nobody> Also as far as I can see the solution is using an O(n^2) sort
19:59:56 <b_jonas> gnu-nobody: the larger problem is that the solution usually runs into a division by zero
20:05:38 <gnu-nobody> if you're talking about sorting the stuff by atan2 (I didn't read the code actually) there is a near-trivial way to bypass that
20:05:48 <gnu-nobody> except I don't remember it
20:08:23 <b_jonas> gnu-nobody: yes, there are better ways to write that program. that's not the question.
20:09:14 <Lymia> Algorithms code is nasty in general.
20:09:17 <Lymia> So easy to fuck everything up.
20:17:44 -!- kspalaiologos has quit (Quit: Leaving).
20:18:06 -!- gnu-nobody has quit (Ping timeout: 265 seconds).
20:57:49 <esowiki> [[User:PythonshellDebugwindow]] M https://esolangs.org/w/index.php?diff=72282&oldid=72249 * PythonshellDebugwindow * (+51)
21:08:01 -!- rain1 has quit (Quit: leaving).
21:08:20 -!- TheLie has joined.
21:14:36 <esowiki> [[User:PythonshellDebugwindow/(Unnamed language)]] M https://esolangs.org/w/index.php?diff=72283&oldid=70948 * PythonshellDebugwindow * (+2390)
21:31:05 -!- atslash has quit (Ping timeout: 265 seconds).
21:31:19 <Lymia> So.
21:31:29 <Lymia> I'm tempted to give BFJoust one last spurt of life before I clean my hands of this. :p
21:31:34 <Lymia> A last attempt at an evolver.
21:31:45 -!- atslash has joined.
21:32:14 -!- Lord_of_Life has quit (Ping timeout: 240 seconds).
21:32:24 -!- Lord_of_Life_ has joined.
21:35:16 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
21:44:02 <zzo38> I am still work on TeXnicard and on the card set that I was making with it. So far there is only one conspiracy, but I can add more. So far only one card has flavor text, but I can hope to add more; do you have some ideas? I also wanted to add one card is the otyugh, too. And, some idea of cards I had but do not know the name to put. And other thing might also be had, to improve the draft, including to adjust rarity perhaps.
21:44:48 <b_jonas> masterkratos: nobody else wanted to run the categories like small only or all levels, so DeeR just got a free world record there.
21:45:04 <b_jonas> glitched any% and glitched jackpot has other runs at least
21:45:29 <b_jonas> also there were like anniversaries as an excuse
21:45:45 <b_jonas> sorry, wrong chat again
22:23:54 -!- opticnerve has quit (Ping timeout: 256 seconds).
23:11:47 <esowiki> [[Ports]] M https://esolangs.org/w/index.php?diff=72284&oldid=72280 * AnimaLibera * (+10) rename github link to include that it contains the TC proof
23:27:48 -!- TheLie has quit (Remote host closed the connection).
23:38:10 <esowiki> [[Esofun]] M https://esolangs.org/w/index.php?diff=72285&oldid=72281 * PythonshellDebugwindow * (+13) Fix title
23:40:37 -!- FreeFull has quit (Ping timeout: 265 seconds).
23:41:14 -!- adu has quit (Ping timeout: 260 seconds).
23:45:21 <esowiki> [[User:PythonshellDebugwindow/(Unnamed language)]] M https://esolangs.org/w/index.php?diff=72286&oldid=72283 * PythonshellDebugwindow * (+149)
23:54:06 -!- adu has joined.
23:54:15 -!- adu has quit (Client Quit).
←2020-05-13 2020-05-14 2020-05-15→ ↑2020 ↑all