<<Prev Next>> Scroll to Bottom
Stuff goes here
L1[00:32:48] ⇦ Quits: Vexatos (~Vexatos@port-92-192-131-75.dynamic.as20676.net) (Quit: Insert quantum chemistry joke here)
L2[00:54:17] ⇦ Quits: Arimil (~Renari@64.67.31.239.res-cmts.bgr.ptd.net) (Ping timeout: 189 seconds)
L3[01:01:43] ⇦ Quits: Hawk777 (~chead@2607:c000:8272:e500:bffa:325b:5bd6:d311) (Quit: Leaving.)
L4[01:12:05] ⇨ Joins: lordpipe (~ba7888b72@66.109.211.50)
L5[02:28:41] ⇨ Joins: Renari (~Renari@64.67.31.239.res-cmts.bgr.ptd.net)
L6[02:42:20] ⇨ Joins: Hawk777 (~chead@2607:c000:8272:e500:c10c:9dea:1c62:ee14)
L7[03:03:21] ⇦ Quits: Renari (~Renari@64.67.31.239.res-cmts.bgr.ptd.net) (Ping timeout: 195 seconds)
L8[03:30:14] <Ar​iri> Izaya: My network has been down and this keeps motivating me to work on it, thank you
L9[03:45:53] <Izaya> :3
L10[03:56:41] * Amanda curls up around elfi, beams a midnight snack of a mini powdered doughnut near her, passes out
L11[03:57:24] <Amanda> Look elfi, I stopped hallucinating before those foxes want and compressed time again!
L12[03:57:30] * Amanda zzz
L13[03:57:44] <Amanda> Night nerds
L14[04:02:16] <ThePiGuy24> the to share this here with no context: http://notavirus.ru/ovo.jpg
L15[04:02:45] <ThePiGuy24> there are atleast 3 in-jokes here
L16[06:01:16] <Forec​aster> %tonk
L17[06:01:17] <MichiBot> Kapow! Forec​aster! You beat Cyborg​Potato's previous record of 7 hours, 40 minutes and 28 seconds (By 47 minutes and 12 seconds)! I hope you're happy!
L18[06:01:18] <MichiBot> Forecaster's new record is 8 hours, 27 minutes and 40 seconds! Forecaster also gained 0.00632 (0.00079 x 8) tonk points for stealing the tonk. Position #2. Need 0.40724254 more points to pass Va​ur!
L19[06:58:21] ⇨ Joins: ben_mkiv (~ben_mkiv@200116b814f61b00fe3497fffea975f2.dip.versatel-1u1.de)
L20[08:29:36] ⇦ Quits: Hawk777 (~chead@2607:c000:8272:e500:c10c:9dea:1c62:ee14) (Quit: Leaving.)
L21[08:46:52] ⇦ Quits: DBotThePony (~Thunderbi@31.220.170.28) (Read error: Connection reset by peer)
L22[09:20:49] <Va​ur> %sip
L23[09:20:51] <MichiBot> You drink a powdery silver potion (New!). Vaur turns into an octiron dragon boy until they have A Liquid Iron potion.
L24[09:29:30] ⇨ Joins: DBotThePony (~Thunderbi@31.220.170.28)
L25[09:54:29] ⇨ Joins: Vexatos (~Vexatos@port-92-192-131-75.dynamic.as20676.net)
L26[09:54:29] zsh sets mode: +v on Vexatos
L27[10:03:08] <Ash​irg> >Izaya: %tell Ariri https://fedi.absturztau.be…
L28[10:03:09] <Ash​irg> Nah, prefectly fine
L29[10:03:10] <MichiBot> Ash​irg: Ariri will be notified of this message when next seen.
L30[10:03:17] <Ash​irg> Fuck
L31[10:03:19] <Ash​irg> lmao
L32[10:03:53] <Ash​irg> It's kinda cute though
L33[10:05:56] <Izaya> there's an argument to be made that one shouldn't be using IRC at work at all but still, best to note that it's a worth a little extra caution
L34[12:27:06] <Va​ur> %sip
L35[12:27:07] <MichiBot> You drink a sedimented titanium potion (New!). Vaur suddenly craves pie.
L36[13:18:02] ⇨ Joins: flappy (~flappy@91-154-0-54.elisa-laajakaista.fi)
L37[13:35:05] <Z0id​berg> IRC at work is fiiiiiiiiine
L38[13:35:19] <Z0id​berg> One could argue that they SHOULD be using IRC at work
L39[13:47:45] <dequbed> Izaya: If I own the company is it still NSFW? :p
L40[13:48:01] <Izaya> that's up to you
L41[13:49:07] <dequbed> But still, oh no two people kissing how absolutely terrible! :p
L42[14:43:53] <Z0id​berg> Is it horrible that when I come up with game ideas for software projects I write a D&D / similar dice rolling library to do skill and success checks?
L43[14:43:56] <Z0id​berg> 😄
L44[14:44:07] <Z0id​berg> I'm thinking of making a Genesys dice rolling library now
L45[14:44:51] <Z0id​berg> you pass it some specs and it uses recursion to create a dice pool that is then used to determine outcomes.
L46[14:45:38] <Z0id​berg> The cool thing about doing it this way isthat I can store dice specs in a configuration file and edit it any time without changing the source code.
L47[14:45:59] <Z0id​berg> which is then use to weigh the balances of pseudo random outcomes.
L48[15:00:48] ⇨ Joins: TPG24 (~ThePiGuy2@host-92-31-110-4.as13285.net)
L49[15:02:14] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-31-110-4.as13285.net) (Ping timeout: 189 seconds)
L50[15:02:49] * Amanda yawns, looks around, notices Sophia is speaking Marian again, tucks her tail over elfi as a blanket and observes
L51[15:03:03] * Elfi squeak?
L52[15:03:14] <Amanda> Martian*
L53[15:03:52] <Elfi> Which Sophia are we talking about?
L54[15:04:05] <Amanda> @Z0idberg
L55[15:04:11] <Elfi> Ohhh
L56[15:04:11] <Elfi> Okay
L57[15:04:24] <Elfi> Because my fiancée is also named Sophia and I was wondering if you followed her on fedi for a moment :o
L58[15:04:49] <Amanda> ah, no
L59[15:05:00] <Z0id​berg> LOL
L60[15:05:25] <Amanda> Though I'm not sure if you and Izaya have been quiet lately, or if the instance Lizzy maintains for us is actin gup again
L61[15:05:39] <Elfi> I've been posting stuff
L62[15:05:58] <Z0id​berg> I tried to get plemora or whatever running but it was problematic
L63[15:10:50] ⇨ Joins: ThePiGuy24 (~ThePiGuy2@host-92-31-110-4.as13285.net)
L64[15:11:35] ⇦ Quits: TPG24 (~ThePiGuy2@host-92-31-110-4.as13285.net) (Ping timeout: 189 seconds)
L65[15:12:04] <Z0id​berg> What is your favorite fediverse instance engine for general things like what mastodon or plemora does?
L66[15:20:09] <Z0id​berg> Ooh, lemmy is a reddit clone, this looks kind of neat...
L67[15:28:10] <Amanda> Elfi: also, I call @Z0idberg sophia from mobile because it's easier to type, and they originally joined #cc with the nick of "Sophia3" after a WW 2(?) tank
L68[15:29:45] <Z0id​berg> something like that
L69[15:29:57] <Z0id​berg> It was a tank from blaster master the NES game actually
L70[15:30:12] <Z0id​berg> which they recently made a remake of for PC
L71[15:31:03] <Z0id​berg> this thing: https://tinyurl.com/yh83djmo
L72[15:33:48] <Z0id​berg> https://www.youtube.com/watch?v=p2bfeq_RO7s
L73[15:34:04] <Z0id​berg> This is a pretty decent summary
L74[15:34:25] <Z0id​berg> It was kind of like a vechicular version of metroid
L75[15:45:46] <Co​mpu> can anyone help me with minitel networking?
L76[15:48:23] <Ocawes​ome101> what about it specifically?
L77[16:05:33] <Co​mpu> ok so
L78[16:05:44] <Co​mpu> i have a server running fserv with a wireless card
L79[16:06:01] <Co​mpu> then i have another server acting as a router with a wireless card and 2 network cards
L80[16:06:25] <Co​mpu> then i have some wired computers connected to the router's network cards via relays
L81[16:06:44] <Co​mpu> when i run fget on a computer with a wireless card it works and pulls info from fserv
L82[16:07:01] <Co​mpu> when i try to use one of the wired ones it doesn't and the server running fserv seems to freeze
L83[16:07:25] <Co​mpu> the router has minitel running in debug mode and shows traffic getting through
L84[16:07:41] <Co​mpu> but fserv never responds to the wired server and just seems to freeze the computer
L85[16:07:51] <Co​mpu> but fserv never responds to the wired computers and just seems to freeze the computer [Edited]
L86[16:07:57] <Co​mpu> but fserv never responds to the wired computers and just seems to freeze the computer running fserv [Edited]
L87[16:08:06] <Co​mpu> what am i doing wrong?
L88[16:08:35] <Co​mpu> i also tried disconnecting the router from the computers and using a relay with a wireless card in it, got the same result
L89[16:10:36] <Elfi> It's Shigeru Miyamoto's 69th birthday today
L90[16:10:44] <Ocawes​ome101> Izaya: Compu has a minitel question
L91[16:11:05] <Elfi> also re: fedi, I am looking forward to GoToSocial
L92[16:34:04] <Amanda> @Compu do you have relay mode turned on?
L93[16:34:14] <Amanda> It's an option in the minitel.cfg
L94[16:34:21] <Co​mpu> on the routing server yes
L95[16:34:23] <Co​mpu> but not on anything else
L96[16:36:31] <Co​mpu> oh apparently if i leave the fserv server alone for long enough it unfreezes on it's own, still doesn't solve my problem tho
L97[16:44:14] <Amanda> Is the wireless range set high enough?
L98[16:44:57] <Amanda> That's about the only things I can think of, I never used the relay stuff much
L99[16:54:22] <Amanda> I let OC do most of my routing by having a single LAN backbone that everything connected to via relay
L100[16:54:30] <Amanda> ( Relay as in OC relay block )
L101[16:58:21] <Co​mpu> yes it's set high enough
L102[16:58:33] <Co​mpu> i know cuz if i use fserv on the routing server it actually works
L103[16:58:55] <Co​mpu> and if it wasn't set high enough how would the wired computers be able to make fserv freeze
L104[16:59:42] <Co​mpu> having everything connected to a relay block seems to make the tier 2 computers complain about too many components
L105[17:01:03] <Co​mpu> i currently have 2 relays with 3 tier 2 computers each, both connected to one of the routing server's wired network cards
L106[17:07:00] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-31-110-4.as13285.net) (Ping timeout: 189 seconds)
L107[17:07:16] ⇨ Joins: ThePiGuy24 (~ThePiGuy2@host-92-31-110-4.as13285.net)
L108[17:15:19] <Va​ur> %tonk
L109[17:15:21] <MichiBot> Kapow! Va​ur! You beat Forec​aster's previous record of 8 hours, 27 minutes and 40 seconds (By 2 hours, 46 minutes and 22 seconds)! I hope you're happy!
L110[17:15:22] <MichiBot> Vaur's new record is 11 hours, 14 minutes and 3 seconds! Vaur also gained 0.02493 (0.00277 x 9) tonk points for stealing the tonk. Position #1.
L111[17:24:08] <Kristo​pher38> Maybe it's a software issue
L112[17:28:45] <Ne​uro> hi all
L113[17:29:27] <Ne​uro> I have a problem with "setBundledOutput" having inconsistent behaviour
L114[17:31:37] <Amanda> @Compu then yeah, you might have to wait for Izaya to wake up, it's the middle of the night for him though
L115[17:32:01] <Co​mpu> ok, he's free to DM me
L116[17:32:40] <Ne​uro> I figured out the problem but its yet one more annoying facet of lua indexing
L117[17:32:59] <Ne​uro> why does the bundled cable api indexing start at 0 when lua's starts at 1?
L118[17:33:11] <Ne​uro> unbearably irritating
L119[17:36:06] <Amanda> Because bundled cables' api is meant to be used with the colours api, which is based off the MC colour enum
L120[17:36:37] <Amanda> It's meant to be a key-value kind of table, not a list kind of table
L121[17:36:47] <Ne​uro> hm, I can get that
L122[17:37:02] <Ocawes​ome101> @Compu izaya uses irc, i don't think he'd be able to dm you (unless Corded does DMs?)
L123[17:37:05] <Ne​uro> im working on a computer tho so for me, the cables are 16-bit encodings, and the colour is only relevant insofar which direction is MSB
L124[17:37:11] <Co​mpu> ah
L125[17:37:15] <Ne​uro> im working on a computer tho so for me, the cables are 16-bit encodings, and the colour is only relevant insofar as which direction is MSB [Edited]
L126[17:38:05] <Ne​uro> i managed to wrangle lua into writing binary decodes starting from 0 but it was a mess
L127[17:38:28] <Co​mpu> tbh i think the problem here isn't a minitel problem but rather an fserv problem
L128[17:38:41] <Co​mpu> fserv freezing and thus not sending a reply
L129[17:38:43] <Ocawes​ome101> izaya is still the right person to talk to :P
L130[17:38:52] <Ocawes​ome101> i believe he wrote fserv
L131[17:39:12] <Co​mpu> it probably does send the reply when it unfreezes but by then the client has given up
L132[17:40:42] <Amanda> It's probably hanging waiting for the ACK
L133[17:41:42] <Amanda> which implies either something wrong with the routing, or one wireless device is too far and not transmitting back
L134[17:41:50] <Mic​hiyo> >Ocawesome101: <@!208512715008770048> izaya uses irc, …
L135[17:41:50] <Mic​hiyo> Corded can do DMs, but it's... annoying. lol
L136[17:43:20] <Mic​hiyo> to PM someone on IRC from Corded you have to send Corded a DM with: IRCName: Your message here, and to reply they have to do DiscordNick: Reply back, and it MAY not work with special characters, spaces, and if the time ends with a digit between 0 and 9
L137[17:43:54] <c​1> is OC out of dev now?
L138[17:44:37] <Co​mpu> i don't think it's the routing or wireless device
L139[17:44:52] <Ocawes​ome101> >c1: is OC out of dev now?
L140[17:44:52] <Ocawes​ome101> the "dev" builds are pretty stable
L141[17:44:57] <Ocawes​ome101> though build 217 broke the redstone IO block
L142[17:45:02] <Co​mpu> because the message switches to wireless on the routing server, and if i fget from the routing server it works just fine
L143[17:45:14] <Amanda> It's not fserv if "fget" works from the router
L144[17:45:31] <c​1> >Ocawesome101: the "dev" builds are pretty stable
L145[17:45:31] <c​1> the ones pinned right side of the forums?
L146[17:45:37] <Amanda> There's no protocol difference betweeen routed packages and non-routed packages.
L147[17:45:46] <Amanda> Correction: There _shouldn't_ be a difference.
L148[17:45:51] <Amanda> s/packages/packets/g
L149[17:45:52] <MichiBot> <Amanda> There's no protocol difference betweeen routed packets and non-routed packets.
L150[17:46:00] <Co​mpu> then idk
L151[17:46:14] <Co​mpu> the routing should be fine, the router is set up to route packets
L152[17:46:21] <Mic​hiyo> >c1: the ones pinned right side of the forum…
L153[17:46:21] <Mic​hiyo> the OC-Staging post? No
L154[17:46:37] <Mic​hiyo> https://ci.cil.li/job/OpenComputers-MC1.12/
L155[17:46:55] <c​1> ah ok
L156[17:46:56] <c​1> thanks
L157[17:46:56] <Mic​hiyo> But like Ocawesome said, 217 broke the redstone io block, so... maybe try 216
L158[17:47:02] <Co​mpu> and i have debug mode turned on on the router and can see packets from both fserv and the wired computers on the debug output
L159[17:47:27] <Mic​hiyo> Unless you specifically need one of the fixes from 217 https://ci.cil.li/job/OpenComputers-MC1.12/changes
L160[17:52:19] ⇦ Quits: notherobrine (~1@47.199.226.237) (Read error: -0x7880: SSL - The peer notified us that the connection is going to be closed)
L161[17:52:34] ⇨ Joins: notherobrine (~1@47.199.226.237)
L162[18:09:43] <Kristo​pher38> the regression present in 217 was introduced earlier
L163[18:10:20] <Kristo​pher38> 213 should be a safe build if you want working redstone IO I believe
L164[18:10:35] <Amanda> "79517 store paths deleted, 171673.60 MiB freed"
L165[18:10:43] <Amanda> Oops.
L166[18:10:54] <Amanda> That explains how "/nix" filled up
L167[18:11:08] <Amanda> guess I need to remember to trim the store some more
L168[18:11:34] <Co​mpu> also all the wireless computers have tier 2 wireless cards so they all have the same range
L169[18:11:44] <Amanda> The range is configurable
L170[18:11:49] <Amanda> with a component call
L171[18:12:01] <Co​mpu> i didn't configure the range, i didn't see any option to do so
L172[18:12:08] <Co​mpu> so they're all at the default
L173[18:12:20] <Kristo​pher38> well it's 400 by default if you don't explicitly call setRange or whatever the method name was
L174[18:12:31] <Amanda> Clarification, are you using OC Relay blocks to do the routing, or a server with multiple cards in it?
L175[18:13:02] <Co​mpu> server with multiple cards, connected to 2 relay blocks, connected to 3 computers per relay block
L176[18:13:22] <Co​mpu> i also tried connecting a computer to one of the cards directly, same problem
L177[18:14:06] <Co​mpu> the fserv computer is definitely closer than 400 blocks away from the server
L178[18:14:17] <Co​mpu> and it works when using fget on the routing server
L179[18:14:22] <Co​mpu> so it's not a range issue
L180[18:14:25] <Amanda> is this in single-player or multiplayer?
L181[18:15:26] <Co​mpu> multiplayer
L182[18:15:27] <Co​mpu> server
L183[18:15:50] <Amanda> Are all the relevent chunks chunk-loaded or close enough?
L184[18:16:11] <Co​mpu> for the record i'm using enderio's OC cables to connect everything, it simplifies things cuz i can use color channels
L185[18:17:30] <Ne​uro> mmm thats good https://tinyurl.com/yk6pdne5
L186[18:18:34] <Ocawes​ome101> what is this
L187[18:18:38] <Co​mpu> lol not sure there's a point to an OC memtest
L188[18:18:44] <Ne​uro> give me a sec
L189[18:18:44] <Ocawes​ome101> looks neat
L190[18:18:46] <Ne​uro> bit more context
L191[18:18:47] <Co​mpu> OC doesn't have a function for failing memory
L192[18:19:05] <Ne​uro> yes yes 1sec
L193[18:19:15] ⇦ Quits: gartral (~gartral@162.243.117.98) (Ping timeout: 192 seconds)
L194[18:19:22] <Ne​uro> here's your context 🙂 https://tinyurl.com/yey7uxf7
L195[18:19:50] <Co​mpu> oboi
L196[18:19:54] <Co​mpu> physical memory
L197[18:20:00] <Ocawes​ome101> nice
L198[18:20:06] <Ne​uro> just making something prodecural to test it
L199[18:20:17] <Co​mpu> i bet it's slow
L200[18:20:23] <Ne​uro> 33x3x33 high, 512b
L201[18:20:26] <Ne​uro> no shit it's slow
L202[18:20:32] <Co​mpu> lol
L203[18:20:56] <Ne​uro> tbh it wouldnt be too bad if redstone ticks werent double real ticks
L204[18:21:03] <Ne​uro> that's really what makes it irritating
L205[18:21:19] <Ne​uro> it can do a writeop in 3 redstone ticks and a read in 2
L206[18:21:24] <Ne​uro> so 6/4 game ticks
L207[18:21:49] <Ocawes​ome101> ~~just use OC for all the memory xd~~
L208[18:21:55] ⇨ Joins: gartral (~gartral@162.243.117.98)
L209[18:22:06] <Ne​uro> the point is to build a computer using projectred chips and wires
L210[18:22:11] <Ne​uro> using OC defeats the purpose
L211[18:22:16] <Ne​uro> im just using OC for procedural testing
L212[18:22:38] <Ocawes​ome101> 😛
L213[18:22:41] <Ne​uro> this was the old design from the prev project btw, much less dense, 5 blocks high https://tinyurl.com/ydnskdlk
L214[18:22:51] <Ne​uro> im quite proud for getting it 2 blocks shorter
L215[18:22:51] <Ocawes​ome101> i prefer using ICs generally
L216[18:22:55] <Ocawes​ome101> that is neat
L217[18:23:17] <Ne​uro> fabrication ics are neat yeah
L218[18:23:36] <Ne​uro> it takes 4 ticks to read because its 2 ticks for address line decode and then 2 ticks for the memory chip to send its contents
L219[18:23:37] <Ocawes​ome101> i just wish the 1.16 implementation of fabrication was ready x.x
L220[18:23:40] <Ne​uro> >Ocawesome101: i just wish the 1.16 implementation o…
L221[18:23:41] <Ne​uro> same bruh
L222[18:23:46] <Ne​uro> ive pestered tjp a bit about it
L223[18:23:47] <Ne​uro> no juice
L224[18:24:11] <Ne​uro> technically you could do a write in 4 ticks as well but it would get finnicky
L225[18:24:25] <Ne​uro> youd do a 1 redstone tick pulse on the second tick
L226[18:24:30] <Ne​uro> and it should work in theory
L227[18:24:33] <Ne​uro> but it would be messy
L228[18:26:01] <Ne​uro> but yes. OC wise, at least, the plan is to scatter around a bunch of wireless controlled microcontrollers that I can control from some main computer and have them stream outputs to me
L229[18:26:33] <Ne​uro> with each functioning as a sort of self contained unit test
L230[18:26:57] <Ne​uro> also for testing the actual machine when built obviously will use OC because how else to set the memory
L231[18:27:46] ⇨ Joins: TPG24 (~ThePiGuy2@host-92-31-110-4.as13285.net)
L232[18:29:35] ⇦ Quits: ThePiGuy24 (~ThePiGuy2@host-92-31-110-4.as13285.net) (Ping timeout: 189 seconds)
L233[18:29:43] <Ne​uro> now the real plan is to stack one/more of these layers on top of each other
L234[18:29:49] <Ne​uro> which is specifically why I wanted it thin
L235[18:30:01] <Ne​uro> altho I fear if I try to worldedit copy another one on top my computer will instantaneously implode
L236[18:34:43] <Ne​uro> well, here goes nothing
L237[18:40:59] <Amanda> %lua "pong!"
L238[18:41:01] <MichiBot> pong!
L239[18:42:54] <Kristo​pher38> @Neuro how many bytes do you have per IC
L240[18:42:58] <Ne​uro> 16#
L241[18:43:00] <Ne​uro> 16 [Edited]
L242[18:43:06] <Ne​uro> its not possible to have more
L243[18:43:14] <Ne​uro> well no
L244[18:43:18] <Ne​uro> you can have more but you can't have 32
L245[18:43:36] <Ne​uro> you could do some weird split cable encoding nonsense to get more than 16 but not 32
L246[18:43:44] <Kristo​pher38> yeah that's what i found out as well
L247[18:43:56] <Ne​uro> a friend did actually build a 32bit computer using the same ics and cables and stuff
L248[18:44:05] <Ne​uro> but she just uses 2 ics per memory word
L249[18:44:11] <Ne​uro> as well as a doubled buses everywhere
L250[18:44:17] <Ne​uro> its a decently impressive project
L251[18:44:32] <Kristo​pher38> i managed to cram 64-byte ROM with address decoding into one IC though
L252[18:44:49] <Ne​uro> yeah not surprised
L253[18:44:51] <Ne​uro> but well, ROM
L254[18:45:20] <Kristo​pher38> but that's about as high as you can go with memory density
L255[18:45:22] <Ne​uro> this is their design https://tinyurl.com/yhkxcj9t
L256[18:45:41] <Ne​uro> simple, works
L257[18:45:56] <Kristo​pher38> said ROM https://tinyurl.com/ydqwp7jg
L258[18:46:10] <Ne​uro> mm thats beautifully compact
L259[18:46:17] <Ne​uro> im guessing you have to set it in the blueprint UI?
L260[18:46:30] <Kristo​pher38> set what exactly?
L261[18:46:36] <Ne​uro> well, the ROM contents
L262[18:46:41] <Kristo​pher38> ah, right
L263[18:47:12] <Kristo​pher38> i guess it could serve really well as a microcode decode ROM
L264[18:47:21] <Ne​uro> yeah I could see the value
L265[18:47:29] <Ne​uro> densest memory I made is a field programmable lookup table
L266[18:47:40] <Ne​uro> let me load into the world that has that
L267[18:47:54] <Ne​uro> I think it holds 8x16 programmable memory entries, with addresses
L268[18:48:00] <Ne​uro> I think it holds 8x16b programmable memory entries, with addresses [Edited]
L269[18:49:26] <Ne​uro> sorry its like, 8 10 bit words? not sure https://tinyurl.com/ygedr2ek
L270[18:49:37] <Ne​uro> I made it specifically for integer decoding to a bus output panel
L271[18:49:42] <Ne​uro> not as a more generalized thing
L272[18:50:02] <Ne​uro> as for memory density, my design is pretty damn good if I say so myself but you could go denser with microblock abuse
L273[18:50:16] <Ne​uro> specifically, you could probably make it 2 blocks high insteadof 3
L274[18:50:19] <Ne​uro> specifically, you could probably make it 2 blocks high instead of 3 [Edited]
L275[18:51:11] <Ne​uro> further still would be .. hard but still not impossible
L276[18:51:33] <Ne​uro> especially if you were willing to make it a lot slower
L277[18:51:35] <Kristo​pher38> this is the design i've got, i had to have this weird bundled wire crossing thingy cause i couldn't avoid crossing them at some point https://tinyurl.com/yhlhtcqg
L278[18:51:41] <Ne​uro> if you were not, not sure if you could make it better
L279[18:52:05] <Ne​uro> >Kristopher38: this is the design i've got, i had to…
L280[18:52:05] <Ne​uro> yeah, cell fields are inevitable in fabrication
L281[18:52:11] <Ne​uro> you just gotta sit back and love the bomb
L282[18:52:20] <Kristo​pher38> 😔
L283[18:52:25] <Ne​uro> what exactly does that hold?
L284[18:52:44] <Kristo​pher38> 16x8-bit words
L285[18:52:52] <Ne​uro> solid
L286[18:53:47] <Ne​uro> and yeah dense rom could be nice as an exu
L287[18:53:54] <Ne​uro> my previous exu design was more of a custom bit logic thingy
L288[18:54:22] <Ne​uro> that implemented a bunch of instruction classes as defined here https://tinyurl.com/yz5g4osb
L289[18:55:56] <Kristo​pher38> that's pretty valid too tbh, there's additional difficulty in designing ISAs which don't rely on a microcode yet are still relatively easy to implement in hardware
L290[18:56:23] <Ne​uro> yeah neither my previous project or current one are gonna have microcode
L291[18:56:44] <Ne​uro> just instruction -> do the instruction
L292[18:57:09] <Ne​uro> i feel like microcode is a luxury for systems that have a few more than 16 bits
L293[18:57:17] <Ne​uro> and a lot more transistors heh
L294[18:57:47] <Kristo​pher38> more like uh
L295[18:57:50] <Kristo​pher38> lower difficulty :P
L296[18:58:06] <Ne​uro> difficulty?
L297[18:58:14] <Ne​uro> do I look like i'm designing computers in minecraft because it's easy?
L298[18:58:36] <Ne​uro> easy would be to place down an opencomputers computer and call it a day
L299[18:59:21] <Kristo​pher38> I've actually designed a physical computer yet to be built using 74 series chips but i wouldn't dare to try and implement the decoding logic with discrete chips
L300[18:59:39] <Ne​uro> as a famous person once said:
L301[18:59:39] <Ne​uro> "We do these things not because they are easy, but because I am bored"
L302[19:00:18] <Kristo​pher38> the decoding logic would probably take the same amount of space as rest of the computer itself :P
L303[19:00:24] <Ne​uro> not sure why
L304[19:00:45] <Ne​uro> if you mean decoding from an existing, complicated instruction set
L305[19:00:50] <Ne​uro> if you make your own, its really not that hard
L306[19:01:20] <Kristo​pher38> okay, how would you design an 8-bit ISA then
L307[19:01:25] <Ne​uro> well
L308[19:01:29] <Ne​uro> I can show you my old 16 bit one 🙂
L309[19:01:37] <Ne​uro> 8 bit tougher yeah, much less bits
L310[19:01:46] <Ne​uro> https://docs.google.com/spreadsheets/d/1CV8DKCSygddtTjly2xFnYni2VbiYqZwAIC_b8ikaqBA/edit#gid=0
L311[19:01:51] <Ne​uro> this is the old project spreadsheet
L312[19:04:29] <Ne​uro> this was admittedly not fully implemented
L313[19:04:40] <Ne​uro> that project was ended at about 75% complete
L314[19:04:53] <Ne​uro> the memory, cache, prefetch, jump and ret, and control logic was implemented
L315[19:05:17] <Ne​uro> most of the instruction chips were not
L316[19:05:52] <Ne​uro> but I think you think microcode is more important because you are thinking like real chips
L317[19:06:09] <Ne​uro> in a real chip you might have one adder, and then everything that remotely has to use an adder has to be routed through it
L318[19:06:28] <Ne​uro> in minecraft/projectred, if you want another adder, you just, well, place down another adder chip
L319[19:06:46] <Sagh​etti> ooooh
L320[19:06:47] <Sagh​etti> this looks cool
L321[19:06:50] <Kristo​pher38> lol
L322[19:06:55] <Ne​uro> in a real chip you might have one adder, and then everything that remotely has to use an adder has to be routed through it (well, a real chip that's quite old, not a newer one) [Edited]
L323[19:06:56] <Kristo​pher38> well yeah
L324[19:06:57] <Sagh​etti> speaking of cpus, made one in logisim
L325[19:07:03] <Ne​uro> logism is neat
L326[19:07:16] <Sagh​etti> >Saghetti:
L327[19:07:17] <Sagh​etti> here's a photo
L328[19:07:17] <Kristo​pher38> i'm thinking of resource usage as well
L329[19:07:27] <Ne​uro> this is literally the pack im using lmao https://tinyurl.com/yf7od6ds
L330[19:07:31] <Ne​uro> or rather, what I called it
L331[19:07:38] <Sagh​etti> cool
L332[19:07:40] <Sagh​etti> i'd play tbh
L333[19:07:54] <Kristo​pher38> i guess in minecraft you could do an out-of-order machine with as many functional units as you want lol
L334[19:07:55] <Ne​uro> lol its ancient and bad
L335[19:08:04] <Ne​uro> >Kristopher38: i guess in minecraft you could do an …
L336[19:08:04] <Ne​uro> yeah thats kinda whats going on here
L337[19:08:08] <Ne​uro> its not even that bad perf wise
L338[19:08:15] <Ne​uro> the memory is the real thing that the game struggles to load
L339[19:08:15] <Sagh​etti> me when i copy/paste my cpu (it's dual core now)
L340[19:08:22] <Ne​uro> >Saghetti: me when i copy/paste my cpu (it's dua…
L341[19:08:23] <Ne​uro> ez life
L342[19:08:28] <Ne​uro> my old design was 4 core
L343[19:08:37] <Ne​uro> the new one will, well, uhh, technically it will be 64 core
L344[19:08:40] <Ne​uro> but thats . . complicated
L345[19:09:34] <Ne​uro> >Saghetti: i'd play tbh
L346[19:09:34] <Ne​uro> i can link you the pack if you want
L347[19:10:00] <Ne​uro> its 1.12 because fabrication latest version is 1.12
L348[19:10:02] <Ne​uro> no way around that
L349[19:10:53] <Kristo​pher38> are you using the fabrication fork with copy-paste?
L350[19:11:05] <Ne​uro> no because I didnt know it existed
L351[19:11:14] <Kristo​pher38> well now you know
L352[19:11:22] <Ne​uro> tbh copy paste isnt the biggest thing
L353[19:11:25] <Ne​uro> there's two things I would want
L354[19:11:32] <Ne​uro> change I/O of entire row at one time
L355[19:11:39] <Ne​uro> and click/drag for many components (cells)
L356[19:11:54] <Ne​uro> change I/O of entire row at once [Edited]
L357[19:12:04] <Kristo​pher38> you have to compile it yourself because projectred shit license doesn't allow you to distribute forks in compiled form
L358[19:12:32] <Ne​uro> rly? what sorta license is that?
L359[19:12:50] <Ne​uro> also its worth noting you can do a sort of copy paste between blueprints with blueprint duping
L360[19:13:01] <Ne​uro> altho yeah copy paste within a blueprint might be nice
L361[19:19:56] <dequbed> @Neuro have you considered playing with FPGA so you have to work around less of the cursed shit MC throws at you? :P
L362[19:20:17] <Ne​uro> you mean a real physical one?
L363[19:20:23] <dequbed> yes
L364[19:20:27] <Ne​uro> or a logical one
L365[19:20:32] <Ne​uro> hm, not really
L366[19:20:50] <Ne​uro> there's something about the arcane limitations of minecraft that makes it interesting
L367[19:20:51] <Ne​uro> right?
L368[19:21:10] <Ne​uro> if I tried to make a CPU from scratch with real world dynamics all I would end up getting are the real world cpus that already exist
L369[19:21:18] <Ne​uro> or at least, minor variations thereof
L370[19:21:40] <Ne​uro> minecraft (or minecraft + projectred) changes a lot of those paradigms in interesting ways
L371[19:21:51] <dequbed> Nah FPGA give you bunch of limitations real CPU don't have either :P
L372[19:22:00] <Ne​uro> besides, it's not so bad, main problems are world loading takes a bit and chip design has messy components
L373[19:22:06] <Ne​uro> besides, it's not *so* bad, main problems are world loading takes a bit and chip design has messy aspects [Edited]
L374[19:24:01] <Ne​uro> but yeah I certainly could
L375[19:24:23] <Ne​uro> just think MC is cool
L376[19:25:04] <Ne​uro> its silly and grossly efficient in many ways, but, I mean, so is doing this at all
L377[19:25:07] <Ne​uro> not like anyone's paying me
L378[19:25:18] <Ne​uro> would be the same if I used logism or an fpga or whatever the hell
L379[19:29:02] <Ne​uro> certainly theres a lot of interesting things ive had to think up minecraft, like the ring address encoding, that i probably wouldnt have had to make in other sstems
L380[19:29:21] <Ne​uro> those other systems would have had challenges and problems of their own, im not saying they're worse, but I'm also saying they're not really better either
L381[19:29:26] <Ne​uro> certainly theres a lot of interesting things ive had to think up minecraft, like the ring address encoding, that i probably wouldnt have had to make in other systems [Edited]
L382[19:30:32] <Ne​uro> but hey if you've done work along those lines
L383[19:30:34] <Ne​uro> show me
L384[19:30:37] <Ne​uro> I might find it interesting
L385[19:34:12] <Kristo​pher38> >work along those lines
L386[19:34:19] <Kristo​pher38> you're underestimating dequbed
L387[19:35:21] <Ne​uro> I have been here for a total of slightly more than an hour
L388[19:35:29] <Ne​uro> i know nothing about anything
L389[19:35:32] <Ne​uro> i know nothing about anyone [Edited]
L390[19:36:32] <Kristo​pher38> yeah i know, it just occurred to me as funny given that i rougly know what dequbed does for a living :P
L391[19:37:00] <Ne​uro> customer FPGA design?
L392[19:37:01] <Ne​uro> neat job
L393[19:37:13] <Ne​uro> this is very much not my living heh
L394[19:37:18] <Kristo​pher38> i didn't say that
L395[19:38:00] <Ne​uro> yes, I assumed it
L396[19:38:06] <Ne​uro> because what else would involve FPGA work?
L397[19:38:19] <Ne​uro> (which in turn was implied by your comment)
L398[19:38:28] <Kristo​pher38> i'll let him handle talking about himself lol
L399[19:38:42] <Ne​uro> if he wants
L400[19:39:01] <Ne​uro> im busy doing the wiring on second memory layer so I'm happy to listen 😛
L401[19:42:43] <Kristo​pher38> personally i enrolled in this fpga course at uni so i now have one, but i didn't get to writing a cpu yet
L402[19:43:42] <Ne​uro> cool
L403[19:43:46] <Ne​uro> I dropped out of uni a while back heh
L404[19:44:26] <Kristo​pher38> feels really enjoyable to write something and instantly have it work in hardware, apart from the fact that you have to use a baroque language
L405[19:45:04] <dequbed> @Neuro nah my job more often than not boils down to "Nice CPUs you have there. Want me to burn them down?"
L406[19:45:28] <Ne​uro> not sure what that quite translates to sorry
L407[19:45:42] <Kristo​pher38> dequbed: i heard you saying that you bang computers into submission
L408[19:45:53] <Ne​uro> 😳
L409[19:46:11] <dequbed> I wouldn't use the term ... "bang" here because that'll give people the wrong association but sure :P
L410[19:46:23] <Kristo​pher38> ...just realized
L411[19:48:14] <dequbed> @Neuro High...ish performance computing and stuff I'd like to call "Infrastructure software", because you only going to get funds for it once, it's invisible to everybody until it breaks and there's never money around for maintenance.
L412[19:48:32] <Ne​uro> still pretty vague but ok
L413[19:48:48] <Ne​uro> if you wanna keep mum i understand
L414[19:49:17] <Amanda> you realise that she works in the kind of industry full of NDAs, she can't exactly point at something and say " I made this for them"
L415[19:49:32] <dequbed> Eh, current project is machine automatization but it's a project I can't talk in detail about for at least another 6 months.
L416[19:49:46] <dequbed> Because of what Amanda said and other legalese shenanigans
L417[19:50:06] <Ne​uro> >Amanda: you realise that she works in the kin…
L418[19:50:06] <Ne​uro> well yes thats why I understand
L419[19:50:15] <Ne​uro> also kristopher said he so which is it?
L420[19:50:27] <dequbed> yes :P
L421[19:50:47] <Kristo​pher38> i was never made aware of dequbed's preffered pronouns so i assumed
L422[19:50:58] <Ne​uro> 🤷‍♂️ yes it is
L423[19:51:36] <Ne​uro> well yes thats why I said I understand [Edited]
L424[19:51:49] <Ne​uro> anyway no as said before I have no work relation to this
L425[19:51:54] <Ne​uro> im currently unemployed infact
L426[19:51:59] <Ne​uro> just a fun project to whittle away time on
L427[19:52:35] <dequbed> My gender is something between "451 unavailable for legal reasons", "due to supply chain issues delivery could take another 6-8 weeks" and "Sorry but your clearance level is too low for that information. You require ULTRAVIOLET clearance, [REDACTED] or [REDACTED] for access"
L428[19:53:24] <Ne​uro> well
L429[19:53:29] <Ne​uro> does that gender have pronouns?
L430[19:53:45] <dequbed> But because most people don't want to deal with that I mostly go by she/her ¯\_(ツ)_/¯
L431[19:53:54] <Ne​uro> ok
L432[19:54:04] <Ne​uro> I mean I am happy to call people whatever but sometimes people make a stink about it
L433[19:54:12] <Ne​uro> so you might as well ask
L434[19:54:58] <Ne​uro> and yeah, your job sounds like it might be cool
L435[19:55:03] <Ne​uro> I can understand not being able to say much
L436[19:55:16] <Ne​uro> I have literally only ever done work experience and I have still signed not one but two NDAs
L437[19:55:25] <Ne​uro> so yeah, sometimes thats just how it is
L438[19:57:04] <dequbed> Honestly, I'm not going to get fussed either way unless people go out of their way to annoy me. I have bigger problems, such as my "L3 cache is too slow" and "This CPU will throttle hard when using SIMD and either shut down or melt if it doesn't, depending on the microcode iteration" :P
L439[19:57:58] <Ne​uro> mm yeah
L440[19:58:04] <Ne​uro> low level cpu work is fun
L441[19:58:22] <Ne​uro> I suppose at least it probably pays quite well
L442[19:58:29] <dequbed> HAH.
L443[19:58:56] <dequbed> Only if you sell your soul to the highest bidder and shoot your conscience.
L444[19:58:56] <Ne​uro> it . . doesn't?
L445[19:59:07] <Ne​uro> ah well most industries work like that
L446[19:59:09] <Ne​uro> unfortunately
L447[20:00:06] <Ne​uro> back when I was studying chemeng the devil was oil companies
L448[20:00:15] <Ne​uro> in tech it's the big tech companies
L449[20:00:23] <dequbed> Like 99% of people do not need fast and reliable at the pricepoint of my hourly wage. Hardware is cheap as fuck and most every algorithm can be scaled horizontally easy enough
L450[20:00:25] <Ne​uro> for you it's, well, idk but yeah
L451[20:01:54] <Ne​uro> actually for you its probably aerospace/defence. but anyway
L452[20:02:05] <Ne​uro> i've never cared too much for money
L453[20:02:10] <Ne​uro> if you make enough to live, good enough
L454[20:02:31] <Ne​uro> and if your hourly wage is notable as you say, you probably do
L455[20:03:55] <dequbed> Its not that its notable it's more that you can't wageslave engineering so salary in that bracket actually does move with general wealthiness of a society.
L456[20:04:17] <Amanda> 🎶 Now We're out of beta, we're releasing on time! 🎶
L457[20:04:47] <luna​r_sam> wew
L458[20:05:08] <Ne​uro> >dequbed: Its not that its notable it's more th…
L459[20:05:08] <Ne​uro> hopefully yeah
L460[20:05:10] <luna​r_sam> i'm gonna make a clone of a touhou game on the PS2
L461[20:05:27] <luna​r_sam> and i'm going
L462[20:05:34] <luna​r_sam> semi hardmode
L463[20:06:09] <Ne​uro> haha cool
L464[20:06:17] <luna​r_sam> https://tinyurl.com/ydknt5e2
L465[20:06:37] <luna​r_sam> 1280x720, too, because i feel like a masochist
L466[20:07:06] <luna​r_sam> ULiL let's go
L467[20:07:09] <Ne​uro> have fun 🙂
L468[20:08:30] <Ne​uro> second bank validated https://tinyurl.com/yzrnlaqd
L469[20:08:40] <Ne​uro> I decided 2 banks is enough for now or my laptop is probably gonna implode
L470[20:08:46] <Ne​uro> so now time to make the memory controller
L471[20:09:15] <Ne​uro> given that there are 13 things that can access memory, gonna be interesting
L472[20:10:59] <dequbed> Reminds me when we wrote an ARM chip IP at Uni. Skipped most of the fun part sadly (that is, the cache controller).
L473[20:11:32] <Ne​uro> yeah this isnt nearly as interesting than that
L474[20:11:42] <Ne​uro> just a bunch of simple memory pipelines
L475[20:11:56] <Ne​uro> I know CPU cache stuff gets a lot more complicated
L476[20:12:08] <Kristo​pher38> I'd run those tests at least several times, i found that for some reason it doesn't do reads/writes(?) fast enough
L477[20:12:23] <luna​r_sam> EE programming time B)
L478[20:12:33] <Kristo​pher38> well depends on what your delays are
L479[20:12:42] <Ne​uro> I have comfy delays set for now yeah
L480[20:12:45] <Ne​uro> I will fiddle with timing later
L481[20:12:48] <Kristo​pher38> actually that might be an OC thing not present in full-redstone circuits
L482[20:12:57] <Ne​uro> for now this is a basic tier "is the wiring good" test
L483[20:13:38] <Ne​uro> currently it only does 1 write a second
L484[20:13:43] <Ne​uro> and a read every 0.4 seconds I think
L485[20:14:02] <Ne​uro> I know both could be a lot faster but again this is just testing whether it works
L486[20:15:24] <Ne​uro> wow I did actually get an unexpected mismatch
L487[20:15:26] <Ne​uro> heh
L488[20:16:06] <Ne​uro> probably a small section didnt copy over quite right
L489[20:16:46] <Ne​uro> it could also be the lua, the lua code that runs this is majorly weird
L490[20:17:00] <Ne​uro> unfortunately I didnt know lua before making this so yeah
L491[20:17:40] <Ne​uro> https://tinyurl.com/ygp7ekz6
L492[20:21:05] <Ne​uro> nvm LOL
L493[20:21:07] <Ne​uro> i misplaced a cable
L494[20:21:10] <Ne​uro> was jumping some weird shit
L495[20:21:46] <Ne​uro> that will almost certainly have been it
L496[20:24:05] <Ne​uro> this was the old memory controller https://tinyurl.com/yg28bwzj
L497[20:24:16] <Ne​uro> I don't really fully know how it works, it sorta half came to be in an epiphany
L498[20:24:25] <Ne​uro> but still very simple
L499[20:24:34] <Ne​uro> new one will probably be just this but looping thru more control signals
L500[20:24:40] <Ne​uro> I don't really fully know how it works, it sorta half came to me in an epiphany [Edited]
L501[20:25:20] <Kristo​pher38> uh
L502[20:25:25] <Kristo​pher38> what is its job?
L503[20:25:34] <Ne​uro> the old design had 4 cores
L504[20:25:41] <Ne​uro> this would round robin between requests from each core
L505[20:25:47] <Kristo​pher38> oh i see
L506[20:25:48] <Ne​uro> setting the appropriate gates and wires live
L507[20:26:07] <Ne​uro> essentially if multiple lines are live
L508[20:26:23] <Ne​uro> it will activate the first live line going left to right in a loop
L509[20:26:34] <Ne​uro> but the smart part is, it starts looking from the last line that was activated
L510[20:26:53] <Kristo​pher38> so it's like a spicy priority encoder?
L511[20:26:56] <Ne​uro> essentially making it a somewhat "fair" round robin allocation
L512[20:27:00] <Ne​uro> yeah
L513[20:27:05] <Ne​uro> again, I dont super know how it works
L514[20:27:14] <Ne​uro> but if you set all inputs at the top high
L515[20:27:23] <Ne​uro> and repeatedly pulse the bottom left memory pulse input
L516[20:27:31] <Ne​uro> it will cycle thru all 4 in order
L517[20:27:47] <Ne​uro> remove one and it will cycle thru the remaining 3, etc
L518[20:28:10] <Ne​uro> obviously in the real thing once a request has been satisfied the core would set the input back to low
L519[20:28:12] <Ocawes​ome101> it's always fun when things just come to you like that
L520[20:28:26] <Ne​uro> I really am quite proud of it, for something stupidly simple
L521[20:29:06] <Ne​uro> the best part is it doesn't use any pulse repeaters
L522[20:29:09] <Ne​uro> like a timer or whatever
L523[20:29:12] <Ne​uro> to iterate through the lines
L524[20:29:18] <Ne​uro> just pure direct bit logic on one pulse
L525[20:29:47] <Ne​uro> well, it has one repeater so two pulses
L526[20:30:09] <Ne​uro> don't 100% remember why I needed that
L527[20:31:04] <Ne​uro> obviously once you have line specific control signals
L528[20:31:21] <Ne​uro> all you need to do is open gates to the memory itself and just let it work
L529[20:31:27] <Ne​uro> so that part of it is pretty simple
L530[20:37:46] <Ne​uro> honestly, personally, I don't really like this way of doing it
L531[20:37:56] <Ne​uro> not because of any specific reason, just more inherently, it feels off
L532[20:38:01] <Ne​uro> but I sure cant think of a better way so hey
L533[20:38:07] <Ne​uro> it does work
L534[20:55:52] <dequbed> Neuro the fun part about engineering is that so often systems are latent emergent from really simple beginnings.
L535[20:55:55] <dequbed> Stupid gates aren't useful, but string them together and suddenly you have a computer revolution. Software is just setting single bits to one or zero but do it enough and you get interactive art letting people express their dreams in unseen before detail and depth.
L536[20:56:40] <Ne​uro> haha
L537[20:56:42] <Ne​uro> I am very aware
L538[20:57:10] <dequbed> And (unless you *start* with particle physics) you can spend years on end just drilling down into further and further detail, making small optimizations to smaller and smaller parts on your way, growing quite old before you hit the edge of human understanding.
L539[20:57:27] <Ne​uro> infact a phrase I coined pretty much as a result of all the learning I did about CPUs during the CPU project, was "Everything complex is just something simple done over and over"
L540[20:57:57] <Ne​uro> so yes, I am aware that design is simple and perfectly functional, which normally is good
L541[20:57:58] <dequbed> I wouldn't quite agree but at the right scale there is some truth behind that statement
L542[20:58:21] <Ne​uro> it's just that, as said, there is some part of me that just doesnt particularly like it for some reason
L543[20:58:42] <Ne​uro> >dequbed: I wouldn't quite agree but at the rig…
L544[20:58:42] <Ne​uro> well of course, one can hardly cram the truth of the world into ten words. but it applies to a lot of things, that's for sure
L545[20:58:47] <dequbed> QFT does qualify as "complex" in my books and I don't think we can say quite yet that it's just very simple things done many times ;)
L546[20:59:02] <Ne​uro> thats just because we haven't figured out what the simple thing is 🙂
L547[20:59:05] <Ne​uro> many people are trying to tho
L548[20:59:10] <dequbed> And any engineering based on QFT or similarly complex topics does sadly inherit said complexity.
L549[20:59:43] <Ne​uro> yes, anything small enough has to deal with those ramifications whether it likes to or not
L550[21:00:20] <Ne​uro> I still find it amusing how far a lot of these technologies have come
L551[21:00:38] <Ne​uro> did you know there are public access quantum computers you could queue up jobs for and get results from like, right now?
L552[21:00:45] <dequbed> The real work of the engineering sciences is not to *be* simple but to find a set of rules that allows you to describe a complex system in simple terms. The strenght of a steel beam is an exercise in metallurgy that will span pages and pages of maths but we can model it quite well using just a number and two terms.
L553[21:00:46] <Ne​uro> I mean you probably did but it's still fascinating to me
L554[21:01:56] <Ne​uro> >Neuro: did you know there are public access …
L555[21:01:56] <Ne​uro> (or to be specific, there were like a year and a bit ago, maybe the programs been shut down. IDK)
L556[21:02:00] <dequbed> Computer science will go that way too I'm certain of it, but we have very little idea what those rules are going to look like. And in any way I will end up as a metallurgist, not a statistican
L557[21:02:41] <Ne​uro> >dequbed: The real work of the engineering scie…
L558[21:02:41] <Ne​uro> I still think said question is, in terms of actions just simple equations
L559[21:02:53] <Ne​uro> the main complexity comes from the environment being highly irregular
L560[21:02:59] <Ne​uro> at least, when looking close up on a steel beam
L561[21:03:06] <dequbed> s/statistican/structural engineer/ turns out english doesn't have the simple word :P
L562[21:03:09] <MichiBot> <dequbed> Computer science will go that way too I'm certain of it, but we have very little idea what those rules are going to look like. And in any way I will end up as a metallurgist, not a structural engineer
L563[21:03:10] <Ne​uro> from which one gets stress propagation and any other numbers of various behaviours
L564[21:03:17] <Ne​uro> >MichiBot: <dequbed> Computer science will go th…
L565[21:03:18] <Ne​uro> ?
L566[21:03:40] <dequbed> Michibot is a bot, doing sed replacements here
L567[21:03:49] <dequbed> Since IRC doesn't have message editing
L568[21:03:53] <dequbed> %pet MichiBot
L569[21:03:54] <MichiBot> dequbed is petting MichiBot with giant meteor. MichiBot regains 1d4 => 4 hit points!
L570[21:03:58] <Ne​uro> but all those systems still work on very simple stress equations
L571[21:03:59] <Ne​uro> ah
L572[21:04:01] <Ne​uro> ok
L573[21:04:18] <Ne​uro> just simple stress equations applied to, say, instead of the whole beam, a specific cross section with void size X
L574[21:04:19] <Ne​uro> etc etc
L575[21:04:48] <Ne​uro> the action is simple, the environment is not
L576[21:04:56] <dequbed> Stress equations? yes. But the changes inside the beam is based on not the Young module of this alloy of steel but on things like the grain size of the different crystals, the lattice orientation, the material of the boundaries between grains, etc.
L577[21:05:20] <Ne​uro> well yes thats the complicated environment I mention
L578[21:05:46] <Ne​uro> they still individually have simple behaviours
L579[21:05:53] <dequbed> And the changes inside crystal lattices of steel aren't equations I'd describe as "simple".
L580[21:06:41] <Ne​uro> well now one just gets into subjectivity and thats boring
L581[21:06:48] <Ne​uro> anyway enough examining some stupid phrase I made u
L582[21:06:49] <dequbed> The same way turbulent flow in the boundary region of a pipe isn't a simple equation but calculating the throughput of a given fluid through a given pipe can be modelled with a very simple equation
L583[21:06:50] <Ne​uro> anyway enough examining some stupid phrase I made up [Edited]
L584[21:09:38] <Ne​uro> thats just a specific scale/viewpoint that results in a complicated equation. What I mean more is if you take infinitesimally small quantities of air, and view how they interact with their neighbouring infinitesimally small quantities of air, those equations are very simple
L585[21:09:59] <Ne​uro> it's true that obviously to get anything useful out of that you have to scale things up, and then things may or may not become more complex
L586[21:10:14] <Ne​uro> depending on the situation and accuracy you are looking for
L587[21:10:29] <dequbed> No because if you go "infinitesimally" small you end up beyond QFT. Ask Vexatos how "simple" that is.
L588[21:11:17] <Ne​uro> okay do we really need to get into semantics
L589[21:11:58] <Ne​uro> "infinitesimally", "very small", okay whatever
L590[21:11:59] <Vexatos> hi what
L591[21:12:00] <dequbed> This is not about semantics. This is about that there *is* no level where the *reality* is simple. We can only make simple *models* which is my central point.
L592[21:12:11] <dequbed> Vexatos: Ohi <3
L593[21:12:36] <Vexatos> qft is a scam
L594[21:12:43] <Ne​uro> oh no
L595[21:12:46] <Vexatos> they are just trying to take your hard earned braincells
L596[21:12:58] <dequbed> QFT is a scam invented by physics teachers to sell you more physics!
L597[21:13:20] <Kristo​pher38> Lmao
L598[21:14:09] <Vexatos> @Neuro I professionally work with infinitesimally small quantities of something
L599[21:14:19] <Vexatos> and the equations are mathematically unsolvable so I wouldn't consider that easy
L600[21:14:22] <dequbed> Vexatos: But yes, I have no idea about quantum chemistry and my quantum physics isn't up to speed either. At what level do things become simple? :P
L601[21:14:42] <Ne​uro> yes fine, I will suffer my poor word choice
L602[21:15:23] <Ne​uro> I think it was clear I didn't mean to enter the quantum realm, but feel free to discuss
L603[21:15:26] <Ne​uro> I find it very interesting as well
L604[21:15:36] <Ne​uro> and yes I am obviously not claiming its simple currently
L605[21:15:38] <dequbed> @Neuro this is still not about semantics. But at what scale are things easy? Single atoms? Interaction of singular molecular gases with each other? Moles? Grams?
L606[21:15:40] <Ne​uro> really really far from it
L607[21:15:45] <Ne​uro> but it's also, well, not complete
L608[21:15:48] <Ne​uro> that much is also obvious
L609[21:16:10] <Forec​aster> has anyone invented quantum bitcoins yet
L610[21:16:19] <Ne​uro> no, dont give them any ideas
L611[21:16:30] <Kristo​pher38> I was about to say the same
L612[21:16:40] <Ne​uro> >dequbed: <@321985179960868864> this is still n…
L613[21:16:40] <Ne​uro> you miss the point. the point isnt that there is some universal scale where everything ever becomes simple, universally
L614[21:16:59] <CompanionCube> who would you piss off *the most* by doing such a thing
L615[21:17:28] <Ne​uro> I mean, there can't even really be such a scale, since a lot of systems dont interact with single atoms or whatever at all
L616[21:17:33] <Ne​uro> like logical systems
L617[21:17:54] <dequbed> CC: Quantum Bitcoin? Me because I'd be asked by entirely way too many people to make their miner go fast while they have fuckall understanding what they're doing and are still orders of magnitude richer than me <.<
L618[21:18:17] <Ne​uro> the point is about the reducibility of, well, quite a lot of things, to simple roots
L619[21:18:39] <Ne​uro> and a lot of the things that you like to mention as being incredibly complex are also incredibly unfinished, and therefore might come to have simple roots in time
L620[21:18:54] <dequbed> Simple roots or as most people would call them, "models". Which is basically my point so yes I agree :P
L621[21:19:16] <Forec​aster> I have a lot of incredibly unfinished projects
L622[21:19:21] <Ne​uro> lol
L623[21:19:23] <Ne​uro> same
L624[21:19:38] <Ne​uro> >dequbed: Simple roots or as most people would …
L625[21:19:38] <Ne​uro> I was never trying to make this an argument or about points
L626[21:19:43] <Ne​uro> I even tried to stop this earlier
L627[21:19:45] <Ne​uro> but anyway
L628[21:20:22] <Forec​aster> %sip
L629[21:20:23] <MichiBot> You drink a porous oculemon potion (New!). A genie tries to turn Forecaster into a dragon, Forecaster tries to evade but is caught in the beam and transformed until they tonk. (2 vs DC 12)
L630[21:20:44] <Forec​aster> well then
L631[21:20:51] <Ne​uro> the problem is I cannot help myself replying to points made at me (If i think there is a suitable reply) and I feel you are likely the same
L632[21:20:52] <Ne​uro> heh
L633[21:21:44] <dequbed> @Neuro oh don't worry I'm not arguing with you about this. I'm just having fun sitting here and having what I said be told back to me with a prefix of "no you see .." :P
L634[21:22:00] <Ne​uro> delightfully passive aggressive
L635[21:22:52] <Ne​uro> anyway
L636[21:22:54] <Ne​uro> i'll be off
L637[21:23:19] <dequbed> take care ^^
L638[21:23:34] <Forec​aster> watch out for dragons
L639[21:23:36] <Forec​aster> oh wait
L640[21:28:23] <Kristo​pher38> https://drewdevault.com/2021/11/16/Cash-for-leftpad.html
L641[21:30:57] <dequbed> DeVault missing the entire point of dependencies in software engineering. Again. yay.
L642[21:33:34] <Ne​uro> ok cooled down
L643[21:33:44] <Ne​uro> sorry Deq
L644[21:33:51] <dequbed> for what? o.O
L645[21:33:58] <Ne​uro> >Kristopher38: https://drewdevault.com/2021/11/16/Ca���
L646[21:33:58] <Ne​uro> honestly just sounds like he wants to cause a bit of chaos
L647[21:34:08] <Ne​uro> a little trolling, if you will
L648[21:34:34] <CompanionCube> i mean
L649[21:34:42] <CompanionCube> the application link explicitly says that
L650[21:35:05] <CompanionCube> it's A Modest Proposal for JavaScript dependencies.
L651[21:35:43] <Kristo​pher38> @Neuro you'll get used to dequbed, I disliked him at first too, then I accepted he's just way more knowledgeable than me
L652[21:35:51] <Ne​uro> I don't dislike him
L653[21:36:01] <Ne​uro> I just really really cannot take being made fun of
L654[21:36:04] <Ne​uro> one of my many flaws
L655[21:36:18] <Kristo​pher38> Oh you get what I mean
L656[21:36:21] <Ne​uro> had to pull myself out for a little bit or I probably would have been horrible
L657[21:36:33] <Ne​uro> I don't dislike her [Edited]
L658[21:36:51] <Ne​uro> people are just people, but it gets oddly hard to remember that at times
L659[21:36:59] <Ne​uro> probably because of the text and avatars and stuff
L660[21:37:07] <dequbed> I have an avatar? o.O
L661[21:37:11] <Ne​uro> im sure if I met them irl, and even had an exact carbon copy of that convo line for line
L662[21:37:16] <Ne​uro> wouldnt have felt bad at all
L663[21:37:19] <dequbed> Wait what profile picture does Corded give me?
L664[21:37:27] <Ne​uro> https://tinyurl.com/yezb99mn
L665[21:37:32] <Ne​uro> i was speaking more generally but you do
L666[21:37:33] <Forec​aster> Corded generates letter avatars unless manually overriden
L667[21:37:38] <dequbed> oh okay good
L668[21:37:43] <dequbed> was worried there for a second
L669[21:38:24] <Ne​uro> well
L670[21:38:27] <Ne​uro> perils of the internet
L671[21:38:59] <Ne​uro> its like the asynchronous nature removes retractability
L672[21:39:01] <Ne​uro> stuff just sorta goes on
L673[21:39:08] <dequbed> Also Neuro I didn't intend to make fun of you so sorry if it sounded like that.
L674[21:39:11] <Ne​uro> if you want to just cut something off IRL you can get it done 9/10 times
L675[21:39:21] <Ne​uro> doesnt really work like that on the internet
L676[21:39:22] <Ne​uro> irritating
L677[21:39:51] <Ne​uro> >dequbed: Also Neuro I didn't intend to make fu…
L678[21:39:51] <Ne​uro> Thanks. I never thought you did. but well, emotions are emotions, and not bound by silly things like logic
L679[21:41:30] <Ne​uro> anyway
L680[21:41:38] <Ne​uro> for something on OC . . for once
L681[21:41:46] <Ne​uro> how do you get the filesize of something?
L682[21:42:10] <Ne​uro> filesystem.size has two problems: it first runs the file, which is unnecessary, and then it doesnt work and spits out some error
L683[21:42:36] <Forec​aster> you can't talk about oc in here, that's illegal
L684[21:42:43] <Ne​uro> lol
L685[21:42:44] <Forec​aster> also what's "some error"
L686[21:42:48] <Ne​uro> should we get back onto quantum field dynamics?
L687[21:42:56] <Ne​uro> >Forecaster: also what's "some error"
L688[21:42:56] <Ne​uro> let me grab it, sec
L689[21:43:26] <Ne​uro> regardless of the error being able to find the size of a lua file without it running first would be nice
L690[21:44:27] <Ne​uro> https://tinyurl.com/yetqurue
L691[21:46:33] <Ne​uro> you can see via the print statements that it runs the file first https://tinyurl.com/yzza9zgh
L692[21:47:02] <Ne​uro> its possible this is a bug that has been fixed because my version is a little old
L693[21:47:21] <Forec​aster> well then you should update it
L694[21:48:13] <Ne​uro> yeah, I dont fully remember why I didnt last time
L695[21:48:21] <Ne​uro> I think another mod I was using with it forced it to be that low
L696[21:48:24] <Ne​uro> maybe openglasses??
L697[21:49:00] <Ne​uro> regardless its rly only barely out of date
L698[21:49:02] <Ne​uro> I will update it and see
L699[21:49:20] <Forec​aster> if you're looking on curseforge that's not the latest version
L700[21:49:41] <Ne​uro> oh well where is it then
L701[21:49:48] <Forec​aster> on the build server
L702[21:50:00] <Ne​uro> altho message here says latest version 1.7.5 which is what curseforge has
L703[21:50:21] <Forec​aster> yeah that's the latest release
L704[21:50:39] <Forec​aster> but the dev builds are more recent
L705[21:50:45] <Ne​uro> if I download the latest artifact jar and just slot it in will it work?
L706[21:50:53] <Kristo​pher38> filesystem.size("memtest.lua")
L707[21:50:53] <Forec​aster> yes
L708[21:51:01] <Ne​uro> ok
L709[21:51:11] <Ne​uro> >Kristopher38: filesystem.size("memtest.lua")
L710[21:51:12] <Ne​uro> ah thats probably it
L711[21:51:17] <Kristo​pher38> @Neuro latest dev build has a non-working redstone io block
L712[21:51:28] <Ne​uro> you mean redstone cards?
L713[21:51:32] <Ne​uro> or the specific block
L714[21:51:50] <Kristo​pher38> Not sure if only the block or the redstone card as well
L715[21:52:06] <Ne​uro> hm well I cant work with that unfortunately
L716[21:52:08] <Kristo​pher38> My guess would be both
L717[21:52:16] <Ne​uro> im gonna test whether what you say works
L718[21:52:21] <Ne​uro> if it does im gonna hold off for now
L719[21:52:27] <Ne​uro> theres also the question of openglasses as said
L720[21:52:38] <Ne​uro> I like it quite a bit and dont want to have to ditch it if it cant work with a newer version
L721[21:53:43] <Kristo​pher38> It should work
L722[21:53:43] <Ash​irg> %sip
L723[21:53:44] <MichiBot> You drink a ripe pear potion (New!). A giant boulder is rolling towards Ashirg! They fail to evade it with a 12 vs DC 15 and takes 1d​6 => 4 damage.
L724[21:53:50] <Ash​irg> Ouch
L725[21:54:25] <Ne​uro> obviously if that fabrication update had come through
L726[21:54:36] <Ne​uro> I probably would have put decent effort into updating everything
L727[21:54:42] <Ne​uro> possibly adding/ditching things
L728[21:54:49] <Ne​uro> but without it, eh
L729[21:55:16] <Ne​uro> this is an old pack lol
L730[22:00:50] <Ne​uro> @Kristopher38 unfortunately that has just exposed a second, equally problematic problem https://tinyurl.com/yh9kjwy8
L731[22:01:32] <Ne​uro> ok it fixed with fullpath but I am surprised it cannot work on relative path
L732[22:02:43] <Ne​uro> the confusing thing is the size it gives is a solid 1k below the size of a floppy written with only that file
L733[22:03:16] <Ne​uro> im trying to stay within EEPROM size so which one applies there is the important one to keep note of
L734[22:03:50] <Ne​uro> it says 1860 while a floppy with only that file on it says 2884
L735[22:04:36] <Kristo​pher38> the former is the correct file size
L736[22:04:57] <Ne​uro> does that mean as long as the former is below 4k I can put it on an EEPROM?
L737[22:07:23] <Kristo​pher38> yeah
L738[22:07:27] <Kristo​pher38> also minifiers exist
L739[22:07:42] <Kristo​pher38> also you need to be wary of what's available on an eeprom
L740[22:08:49] <Ne​uro> or rather, what's not
L741[22:08:57] <Ne​uro> what is not available?
L742[22:09:16] <Ne​uro> I only really need redstone component, modem component, and well, lua
L743[22:09:26] <Kristo​pher38> openOS APIs
L744[22:09:37] <Kristo​pher38> so basically no require
L745[22:10:01] <CompanionCube> Izaya: https://github.blog/2021-11-15-githubs-commitment-to-npm-ecosystem-security/ talk about buying the lede...
L746[22:10:05] <Kristo​pher38> and anything that comes with it
L747[22:10:34] <Ne​uro> oh well I only used require for components
L748[22:10:42] <Ne​uro> and I know theres a different way to do that on microcontrollers
L749[22:10:58] <Ne​uro> so thats nbd
L750[22:11:49] <Forec​aster> more accurately, no OpenOS libraries unless you include them in the eeprom code
L751[22:12:56] <Ne​uro> yeah, all this code is super simple tho
L752[22:13:18] <Ne​uro> as said, as long as I can still use redstone component and moden component
L753[22:13:24] <Ne​uro> all good
L754[22:17:08] <Ne​uro> not gonna fully implement that yet anyway
L755[22:17:19] <Co​mpu> we may have found my problem
L756[22:17:52] <Co​mpu> @MGR was having me switch from minitel to GERT and found there were duplicate messages on the wired network
L757[22:18:16] <Co​mpu> we found it was because the server rack's integrated network switch was turned on
L758[22:18:36] <Kristo​pher38> so you got indoctrinated
L759[22:18:45] <Ne​uro> cool
L760[22:18:49] <Co​mpu> yeah but i'm not sure if i wanna stick to GERT
L761[22:18:59] <Co​mpu> minitel has that useful frequest utility
L762[22:19:10] <Co​mpu> and the ability to find computers by hostname
L763[22:19:19] <Kristo​pher38> iirc minitel guide explicitly mentions to turn off the rack's integrated switch
L764[22:19:30] <Co​mpu> what u guys think?
L765[22:19:33] <Co​mpu> gert or minitel?
L766[22:19:44] <Forec​aster> I've used neither, so I say use both!
L767[22:19:54] <Kristo​pher38> what was the command that amanda frequently uses
L768[22:20:08] <Forec​aster> choose?
L769[22:20:13] <Kristo​pher38> %choose gert or minitel
L770[22:20:16] <MichiBot> Kristo​pher38: Once you get a taste of "gert" you can't stop.
L771[22:20:21] <Ne​uro> LOL
L772[22:21:24] <Brisingr​Aerowing> https://www.youtube.com/watch?v=6aNm3qWkW8A
L773[22:21:28] <MichiBot> [545] Master Lock No. 3 vs. 900 Amps! | length: 3m 24s | Likes: 38,536 Dislikes: 780 Views: 1,516,207 | by LockPickingLawyer | Published On 1/10/2017
L774[22:22:03] <Ne​uro> oh god what has he done now
L775[22:22:08] <Brisingr​Aerowing> Old video
L776[22:22:28] <Ne​uro> oh yeah right hes past like 1000 isnt he
L777[22:22:47] <Brisingr​Aerowing> 1382 is the latest
L778[22:22:57] <Ne​uro> thats really high
L779[22:23:00] <Ne​uro> guess his pace has accelerated
L780[22:23:18] <Ne​uro> yeah it deffo has
L781[22:23:34] <Ne​uro> maybe hes leaning more towards lockpicking and less towards lawyer
L782[22:23:34] <Ne​uro> heh
L783[22:25:45] <Co​mpu> what is the rack's integrated switch for anyways?
L784[22:25:50] <Co​mpu> it just seems to cause problems
L785[22:25:53] <Co​mpu> why does it exist?
L786[22:26:52] <Kristo​pher38> to allow passing messages between multiple servers while they're connected to different rack sides presumably
L787[22:27:02] <Co​mpu> ah
L788[22:27:19] <Co​mpu> i have a single server with multiple network cards
L789[22:27:24] <Forec​aster> yep, it just links all the sides network connections together
L790[22:27:35] <Co​mpu> btw speaking of multiple cards
L791[22:27:40] <Forec​aster> shown by a line that appears at the bottom of the gui
L792[22:27:44] <Co​mpu> i can't seem to change the side for the last network card
L793[22:27:49] <Co​mpu> it's stuck on "bottom"
L794[22:28:12] <Ne​uro> this mod has a lot of weird problems with sides lol
L795[22:28:33] <Ne​uro> my code had to swap left and right to work for . . some reason
L796[22:28:57] <Forec​aster> left and right are always the facing of the block
L797[22:29:01] <Ne​uro> and its not like the computer just is oriented in a different way because front and back are as you would expect
L798[22:29:10] <Forec​aster> and in the case of a server rack it's facing in the direction of the opening
L799[22:29:18] <Ne​uro> what about a computer?
L800[22:29:33] <Forec​aster> the front should be the side the light is on
L801[22:29:49] <Ne​uro> not here
L802[22:30:05] <Ne​uro> or unless, does it mean left and right as if you were standing in the computer, facing at the light?
L803[22:30:15] <Forec​aster> yes
L804[22:30:32] <Co​mpu> yep
L805[22:30:37] <Co​mpu> it says that on the wiki
L806[22:30:37] <Forec​aster> that's what "the front is the side with the light is" means
L807[22:30:49] <Forec​aster> that's what "the front is the side where the light is" means [Edited]
L808[22:31:03] <Forec​aster> if you're standing opposite the computer of course left and right are reversed
L809[22:31:24] <Ne​uro> >Compu: it says that on the wiki
L810[22:31:24] <Ne​uro> I did actually look up sides api to try figure it out
L811[22:31:33] <Ne​uro> its not rly spelt out there thoi
L812[22:31:37] <Ne​uro> must have missed where it was spelt out
L813[22:31:39] <Forec​aster> I tend to just use the cardinal directions
L814[22:31:41] <Co​mpu> idk i understood it
L815[22:31:45] <Ne​uro> its not rly spelt out there tho [Edited]
L816[22:31:58] <Ne​uro> I tried cardinals too
L817[22:32:00] <Ne​uro> didnt seem to work
L818[22:32:49] <Forec​aster> the sides api is just a table of numbers by the way
L819[22:33:22] <Amanda> Cardinals are just front/back/left/right with an assumption in position
L820[22:33:27] <Kristo​pher38> Let me try to imitate Forecaster's silly joke: but have you tried bishop directions?
L821[22:33:32] <Amanda> Like, literally
L822[22:33:51] <Ne​uro> yes I can see that
L823[22:33:56] <Amanda> North and I think forward are both the same values
L824[22:33:58] <Ne​uro> im not sure what didnt work
L825[22:34:16] <Ne​uro> but I set the sides to what the minimap said I was facing and it didnt work
L826[22:34:22] <Amanda> In the sides API
L827[22:34:26] <Ne​uro> i mean its all solved now so w/e
L828[22:35:14] <Forec​aster> they're not the same, north is always north in the world
L829[22:37:17] <Amanda> No, it's not, it depends on what you're calling that it will be north or back: https://github.com/MightyPirates/OpenComputers/blob/master-MC1.7.10/src/main/resources/assets/opencomputers/loot/openos/lib/sides.lua#L10
L830[22:38:35] <Forec​aster> huh
L831[22:39:31] <Ne​uro> well that explains why using them didnt work lol
L832[22:39:48] <Amanda> when payo was brainstorming an OC2 I suggested that those two groups be seperated, or at least have seperate vlaues and a Java-side API to make it make sense
L833[22:40:32] <Ne​uro> yeah universal directionality would be nice
L834[22:40:44] <Amanda> so I guess, @Sangar ^ as a comment before OC2 leaves alpha
L835[22:41:10] <Forec​aster> I've never had an issue with the cardinal directions not doing what I expected
L836[22:41:33] <Amanda> Then everything you've placed that accepts front/back/left/right happen to be facing south
L837[22:41:59] <Forec​aster> that seems unlikely
L838[22:42:25] <Amanda> I mean, those are the ints that get passed to the Java-side APIs, so it's not like it's some magic happening
L839[22:43:19] <Ne​uro> does that mean I can replace every instance of sides.down with 0
L840[22:43:25] <Forec​aster> yes
L841[22:43:44] <Ne​uro> guess would be a little bit of code compactification
L842[22:45:02] <Ne​uro> also the world I was building my computer in has decided to randomly corrupt
L843[22:45:04] <Ne​uro> thats great
L844[22:45:23] <Izaya> 7 pings in one night is most unusual
L845[22:45:36] <Izaya> Amanda: I'm very loud on fedi
L846[22:45:44] <Amanda> someone was actuall;y using minitel, Izaya, before @MGR kitnapped them
L847[22:46:37] <Amanda> someone other than me, I mean. :P
L848[22:47:35] <Izaya> @Compu I'm not sure if this is still an open question but EIO conduits and racks behave weirdly, I ended up doing cursed things to coerce them into working
L849[22:47:52] <Co​mpu> i haven't had problems with them
L850[22:48:10] <Co​mpu> but then i have the computers connected to relays before they go to the rack
L851[22:48:54] <Izaya> I found they tended to mix component networks
L852[22:50:35] <Izaya> ended up doing this to sanitise them https://social.shadowkat.net/xmpp/upload/aHLIFt2YUPsK4GOL/2021-09-08_22.03.55.jpg
L853[22:50:43] <Forec​aster> maybe I'm misremembered the frequency of cardinal direction use, it's more likely
L854[22:50:52] <Forec​aster> maybe I misremembered the frequency of cardinal direction use, it's more likely [Edited]
L855[22:53:45] <Mic​hiyo> "I've never had issues with EIO Conduits" "I'm having an odd issue with this and I'm using EIO Conduits" hmm
L856[22:53:48] <Mic​hiyo> Anyway
L857[22:53:52] <Mic​hiyo> back to work for me.
L858[23:01:56] <lunar_sam> GOD
L859[23:01:59] <lunar_sam> my head hurts
L860[23:02:06] <lunar_sam> can't figure out this fucking model format
L861[23:02:09] <lunar_sam> pain
L862[23:02:14] <lunar_sam> what did tasofro use
L863[23:07:25] <Amanda> Magic
L864[23:07:36] <Amanda> & a sprinkling of the souls of the Old Ones
L865[23:10:11] <Amanda> %choose rain box time or wait
L866[23:10:11] <MichiBot> Ama​nda: Some "rain box time" sounds nice
L867[23:11:37] <Ne​uro> trying to think of a clever way to do memory access management but nothing is rly clicking
L868[23:11:51] <Amanda> infinite beach with crabs and rocks, clearly
L869[23:12:17] <Ne​uro> lol, like the infinite desert xkcd
L870[23:12:28] <Amanda> That was what I was referencing, yes
L871[23:12:39] <Ne​uro> there are like . . segments of clever ideas floating around
L872[23:12:39] <Amanda> %remindme 5m rain box
L873[23:12:40] <MichiBot> I'll tell you "rain box" in 5m at 11/16/2021 11:17:39 PM
L874[23:12:48] <Ne​uro> but they're all being cockblocked by various things
L875[23:13:08] <Izaya> https://social.shadowkat.net/xmpp/upload/Op0za8G1F2ov_o-c/metatables.jpg
L876[23:13:48] <Ne​uro> if you do a system as described before, it works, but its a bit slow
L877[23:14:08] <Ne​uro> specifically, 4 ticks + however long the memory bit takes
L878[23:14:30] <Ne​uro> since 2 ticks to assign access, 2 ticks for passthrough gates to open
L879[23:15:15] <Ne​uro> means total is now 8 ticks since you then ad 2 ticks for address decode and 2 for the memory chip to output
L880[23:15:37] <Ne​uro> its certainly not terrible but you feel you can squeeze more out of it . . somehow
L881[23:15:56] <M​GR> >Amanda: someone was actuall;y using minitel…
L882[23:15:56] <M​GR> No kidnapping was involved
L883[23:16:02] <Ne​uro> I was musing about making it shorter by having "open as default" gates
L884[23:16:18] <Ne​uro> in which a gate would close all others once it gets an input
L885[23:16:31] <Ne​uro> would save 2 ticks but incompatible with two memory requests coming in same tick
L886[23:16:43] <Ne​uro> unfortunate, would work without that minor caveat
L887[23:17:40] <MichiBot> Amanda REMINDER: rain box
L888[23:19:34] <lunar_sam> once TN is usable
L889[23:19:49] <lunar_sam> i'm gonna have to tell everyone to use it for networking
L890[23:19:56] <lunar_sam> except where minitel makes more sense
L891[23:19:57] <lunar_sam> :^)
L892[23:21:37] ⇦ Quits: Vexatos (~Vexatos@port-92-192-131-75.dynamic.as20676.net) (Quit: Insert quantum chemistry joke here)
L893[23:22:18] <Kristo​pher38> @Neuro it's easy, just make a L1 cache 🙃
L894[23:22:38] <Kristo​pher38> or alternatively do speculative execution on memory loads
L895[23:23:12] <Ne​uro> speculative branching is fundamentally incompatible with the system i am designing
L896[23:23:13] <Ne​uro> unfortunately
L897[23:23:33] <Ne​uro> its also not super worth it because even at 8 ticks a load
L898[23:23:52] <Ne​uro> the difference between instruction execution time and memory read time is soooo much higher (relatively) in a real system
L899[23:24:02] <Kristo​pher38> i meant speculatively predicting load values ;^)
L900[23:24:11] <Ne​uro> still incompatible
L901[23:24:27] <Ne​uro> since there would be no way to cancel/reverse it
L902[23:25:26] <Ne​uro> the basic design for current system is essentially an 8 by 8 grid of execution units
L903[23:25:29] <Kristo​pher38> i'm just messing with you
L904[23:25:47] <Ne​uro> once you pass data into the grid, there's no easy way to reverse its effects
L905[23:25:53] <Ne​uro> infact, there isnt really even a hard way
L906[23:26:20] <Ne​uro> the basic design for current system is essentially an 8 by 8 grid of execution units which data is streamed into [Edited]
L907[23:26:41] <Ne​uro> you just can't
L908[23:26:46] *** Guest9481 is now known as SinZ
L909[23:27:20] <Ne​uro> >Kristopher38: i'm just messing with you
L910[23:27:21] <Ne​uro> well I might as well treat it seriously, specex, branching, caching etc are all real
L911[23:27:56] <Ne​uro> unfortunately the only somewhat fancy optimization the old system had was instruction precaching
L912[23:28:04] <Ne​uro> since its the only one thats relatively trivial to implement
L913[23:28:07] <Kristo​pher38> stop, you can't do that
L914[23:28:12] <Kristo​pher38> treating people seriously on the internet is illegal
L915[23:28:23] <Ne​uro> well I mean
L916[23:28:38] <Ne​uro> those are ideas I have seriously thought about even without someone else suggesting them
L917[23:29:12] <Ne​uro> honestly I don't like the current design a ton because its control flow is janky asf
L918[23:29:55] <Ne​uro> in a more ideal world, you would simply stream data into the grid nonstop
L919[23:30:04] <Ne​uro> and it would figure out how to rearrange itself and process streams on the fly
L920[23:30:27] <Ne​uro> of course, something like that is much much much much easier said than done
L921[23:30:54] <Ne​uro> but not necessarily impossible at least
L922[23:30:58] <Kristo​pher38> I don't know how your grid works
L923[23:31:10] <Ne​uro> pretty simple
L924[23:31:17] <Ne​uro> 8 by 8 grid of independent execution units
L925[23:31:30] <Ne​uro> each can talk with any other on its row or column
L926[23:31:57] <Ne​uro> they each hold one instruction
L927[23:32:01] <Ne​uro> which also encodes where they send the result
L928[23:32:26] <Ne​uro> the row 0 and row 7 cells also have special properties that allow them to interact with memory
L929[23:32:32] <Ne​uro> they are the i/o, in a sense
L930[23:32:48] <Kristo​pher38> this sounds oddly familiar to classical out of order
L931[23:32:55] <Ne​uro> perhaps
L932[23:32:58] <Ne​uro> you have to understand
L933[23:33:02] <Ne​uro> I am very much self studied
L934[23:33:06] <Ne​uro> I design and name things on my own
L935[23:33:14] <Ne​uro> more likely than not people have certainly thought similar things before
L936[23:33:18] <Ne​uro> if not even made similar things
L937[23:33:31] <Kristo​pher38> is the encoding where they send the result done on the fly or by the programmer?
L938[23:33:44] <Ne​uro> not on the fly, its encoded as part of the instruction
L939[23:34:09] <Ne​uro> https://tinyurl.com/yhlf2aoh
L940[23:34:11] <Kristo​pher38> so the programmer/compiler has to allocate where it goes?
L941[23:34:14] <Ne​uro> yes
L942[23:34:27] <Ne​uro> you can see, that instruction includes two 4 bit addresses
L943[23:34:44] <Ne​uro> 8 on a row + 8 on a column = 16 = 4 bits
L944[23:34:52] <Ne​uro> altho two of those are redundant because they point to the sending cell
L945[23:35:29] <Ne​uro> on the fly instruction encoding/switching would be more the ideal as I described here
L946[23:35:39] <Ne​uro> >Neuro: and it would figure out how to rearra…
L947[23:35:40] <Ne​uro> here
L948[23:35:43] <Ne​uro> but obviously, thats pretty hard
L949[23:36:12] <Ne​uro> the general point is that each EXU is gonna be pretty KISS
L950[23:36:18] <Ne​uro> cause theres gonna be 64 of them after all
L951[23:36:23] <Ne​uro> dont wanna put anything too fancy in them
L952[23:36:49] <Kristo​pher38> think about it this way: each of your execution units can hold an instruction, lets assume all of them have two operands
L953[23:36:59] <Ne​uro> each has access to two 16 bit registers
L954[23:37:02] <Ne​uro> DL1 and DL2
L955[23:37:15] <Kristo​pher38> are these local to an execution unit?
L956[23:37:21] <Ne​uro> yes
L957[23:37:35] <Ne​uro> exu has 3 local registers, the one that holds the instruction and two data ones
L958[23:37:43] <Ne​uro> plus some minor buffers for send/recieve
L959[23:37:51] <Kristo​pher38> so these are essentially the operands if an instruction needs two of them?
L960[23:37:55] <Ne​uro> yes
L961[23:38:08] <Ne​uro> all instructions have access to both and can do whatever they want to them
L962[23:38:13] <Ne​uro> all instructions have access to both and can do whatever they want with them [Edited]
L963[23:39:35] <Kristo​pher38> so think about this: instead of only keeping data there, you could also add additional bits to keep a "tag" value
L964[23:40:08] <Ne​uro> go on . .
L965[23:40:24] <Kristo​pher38> this tag would be an address of the execution unit that produces this value
L966[23:40:39] <Kristo​pher38> this tag would be an address of the execution unit that produces the data [Edited]
L967[23:40:51] <Ne​uro> you mean produced
L968[23:40:58] <Ne​uro> so as in, "recieved from x"
L969[23:41:03] <Ne​uro> or no?
L970[23:41:10] <Kristo​pher38> no, it will produce at some unknown time
L971[23:41:19] <Ne​uro> sorry I really do not understand
L972[23:41:34] <Kristo​pher38> it's alright, let me paint an example
L973[23:42:34] <Kristo​pher38> imagine you have an execution unit #1 that does some arithmetic and produces value `x`
L974[23:42:49] <Ne​uro> lets use column row addresses
L975[23:42:54] <Ne​uro> so lets say unit 0/0
L976[23:43:06] <Ne​uro> but sure
L977[23:43:23] <Kristo​pher38> well whatever, this is just a general idea
L978[23:43:29] <Ne​uro> fair
L979[23:49:50] <Kristo​pher38> now another execution unit #2 is waiting for that execution unit #1 to finish its calculations because it needs that value `x` to proceed with calculations, so lets say it wants to do an add, DL1 has a value of 42, and DL2 is supposed to have that value `x` but was waiting for execution unit #1 to complete calculating it. DL2 will have no valid value, but instead a "tag", indicating which execution unit it should receive its result from. When
L980[23:49:50] <Kristo​pher38> execution unit #1 finishes its calculations, it "broadcasts" both the value `x` and its "tag" on the bus. Execution unit #2 sees that the tag it was waiting for has been broadcasted, puts the broadcasted value `x` in its DL2 register and can then start executing
L981[23:50:50] <Ne​uro> so you mean addresses, but mutable
L982[23:51:07] <Kristo​pher38> in which part i mean addresses
L983[23:51:13] <Ne​uro> well
L984[23:51:22] <Ne​uro> the way it currently works is each cell has a row/column address
L985[23:51:41] <Ne​uro> and each cell explicitly sends its output to an instruction encoded address
L986[23:51:55] <Ne​uro> this "tag" systems seems to be like that, but more as if any cell can declare its own "address" to be the tag
L987[23:52:02] <Ne​uro> and just take anything that is tagged with that tah
L988[23:52:05] <Ne​uro> and just take anything that is tagged with that tag [Edited]
L989[23:52:52] <Ne​uro> the "address" in this context is just three bits of the column/row general bus
L990[23:53:06] <Brisingr​Aerowing> https://tinyurl.com/yhk8tc4g
L991[23:53:27] <Ne​uro> each cell on the column/row just uses a simple check as to whether the address is set to their address, and if that happens when the write line flashes, it gets passed to their input buffers
L992[23:53:46] <Kristo​pher38> i don't know if I was clear enough, EU #2 has, alongside it's DL2 without valid value, a tag with the value of 1 (as in, grab this result from execution unit 1)
L993[23:54:11] <Ne​uro> reverse addresses
L994[23:54:12] <Kristo​pher38> i don't know if I was clear enough, EU #2 has, alongside it's DL2 without valid value, a tag with the value of 1 (as in, grab this result from execution unit 1 and put it in DL2 as soon as it's available) [Edited]
L995[23:54:13] <Ne​uro> hm
L996[23:54:20] <Ne​uro> instead of "SEND TO X" it's "GET FROM X"
L997[23:54:23] <Kristo​pher38> yes
L998[23:54:28] <Ne​uro> is that even majorly different?
L999[23:54:47] <Ne​uro> I can't immediately think of how it might be but I guess makes compiling easier
L1000[23:55:52] <Ne​uro> it would also require me to free one bit from the intercell talk line but I could probably wrangle that
L1001[23:55:55] <Ne​uro> or just widen it
<<Prev Next>> Scroll to Top