<<Prev Next>> Scroll to Bottom
Stuff goes here
L1[00:37:03] ⇦ Quits: lunar_sam (c44a7f2987@jabberfr.org) (Ping timeout: 186 seconds)
L2[00:38:13] ⇨ Joins: lunar_sam (c44a7f2987@2a00:c70:1:178:170:40:189:1)
L3[01:16:42] <Izzy> @Ocawesome101 https://vt100.net/mirror/mds-199909/cd3/term/vt420rm2.pdf chapter 9
L4[01:17:29] <PwnagePineap​ple (He/Him)> %tonk
L5[01:17:29] <MichiBot> Yay! PwnagePineap​ple (He/Him)! You beat Compan​ionCube's previous record of 2 hours, 48 minutes and 4 seconds (By 59 minutes and 12 seconds)! I hope you're happy!
L6[01:17:30] <MichiBot> PwnagePineapple (He/Him)'s new record is 3 hours, 47 minutes and 17 seconds! PwnagePineapple (He/Him) also gained 0.00297 (0.00099 x 3) tonk points for stealing the tonk. Position #8. Need 0.121118 more points to pass walksanato​r. B⃢ot.!
L7[01:22:17] <Izzy> everything but DECFRA and DECCRA would be much easier with a buffering terminal emulator rather than one that talks directly to the GPU
L8[01:23:15] * Amanda teaches elfi how to forge IDs by melting down a bunch of brass from cop badges
L9[01:24:37] * Amanda lays her head on elfi and passes out
L10[01:24:42] <Amanda> Night girls
L11[01:29:28] <Izzy> How much video memory does each GPU tier have?
L12[01:50:53] ⇨ Joins: Hawk777 (~Hawk777@2607:c000:8299:e400:57bf:3df5:664e:8a77)
L13[03:14:14] ⇦ Quits: Hawk777 (~Hawk777@2607:c000:8299:e400:57bf:3df5:664e:8a77) (Quit: Leaving.)
L14[04:16:36] ⇨ Joins: Hawk777 (~Hawk777@2607:c000:8299:e400:57bf:3df5:664e:8a77)
L15[05:57:24] ⇦ Quits: dustinm` (~dustinm@static.38.6.217.95.clients.your-server.de) (Quit: Leaving)
L16[06:00:21] ⇦ Quits: glasspelican (~quassel@2607:5300:201:3100::325) (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
L17[06:00:54] ⇨ Joins: glasspelican (~quassel@2607:5300:201:3100::325)
L18[06:17:16] ⇨ Joins: dustinm` (~dustinm@static.38.6.217.95.clients.your-server.de)
L19[06:34:58] <Forec​aster> %tonk
L20[06:34:59] <MichiBot> Yeah! Forec​aster! You beat PwnagePineap​ple (He/Him)'s previous record of 3 hours, 47 minutes and 17 seconds (By 1 hour, 30 minutes and 12 seconds)! I hope you're happy!
L21[06:35:00] <MichiBot> Forecaster's new record is 5 hours, 17 minutes and 29 seconds! Forecaster also gained 0.006 (0.0015 x 4) tonk points for stealing the tonk. Position #1.
L22[06:55:12] ⇨ Joins: Vexatos (~Vexatos@p200300eaef1a5768c3ee193d72b4af83.dip0.t-ipconnect.de)
L23[06:55:12] zsh sets mode: +v on Vexatos
L24[07:16:51] ⇦ Quits: Hawk777 (~Hawk777@2607:c000:8299:e400:57bf:3df5:664e:8a77) (Quit: Leaving.)
L25[07:59:38] <Kristo​pher38> Izzy: tier number times resolution at that tier
L26[07:59:47] <Izzy> ah sweet
L27[08:00:27] <Izzy> so it'd be like, 640 bytes, 4000 bytes, and 24000 bytes?
L28[08:01:23] <Kristo​pher38> I don't remember the resolutions of off the top of my head but sounds about right
L29[08:15:08] <AR2​000> Sounds right to me.
L30[08:15:08] <AR2​000> A T3 GPU can create 3 full extra buffers
L31[08:16:04] <Izzy> so one T3 GPU could support 12 virtual terminals
L32[08:16:33] <Izzy> I wonder if you could use one GPU to round robin blit multiple screens
L33[08:35:27] <AR2​000> >Izzy: so one T3 GPU could support 12 virtual terminals
L34[08:35:28] <AR2​000> I don't think so. Unless you use a T2 resolution
L35[08:35:38] <Izzy> yup
L36[08:35:48] <Izzy> that's what I was thinking
L37[08:35:56] <stephan48> i would image performance of the updates would be the main issue?
L38[08:37:34] <Izzy> apparently drawing to VRAM is pretty fast and basically free power-wise
L39[08:37:44] <Izzy> so as long as you didn't have too many updating at the same time...
L40[08:39:52] <AR2​000> Is is.
L41[08:39:52] <AR2​000> When drawing a image pixel per pixel it is instant in a frame buffer, while directly on screen let you see each pixel being drawn at a time
L42[08:40:31] <Izzy> it may even result in better performance for stuff written like the PsychOS terminal emulator, then
L43[08:42:35] <Izzy> it could draw to the framebuffer then mark it as dirty, and then put it in the queue to be blit to the screen
L44[08:51:06] <stephan48> "blit"? i thought the first time that was a writing mistake, but now that you keep repreating it i have to ask... what is the meaning?
L45[09:11:41] <Izzy> > Bit blit (also written BITBLT, BIT BLT, BitBLT, Bit BLT, Bit Blt etc., which stands for bit block transfer) is a data operation commonly used in computer graphics in which several bitmaps are combined into one using a boolean function.
L46[09:11:48] <Izzy> not entirely correct here but close enough
L47[09:30:32] <stephan48> thanks
L48[10:05:14] <Kristo​pher38> Message contained 4 or more newlines and was pastebined https://paste.pc-logix.com/oyelasiwux
L49[10:09:06] <Izzy> I'm thinking the idea is that the terminal emulator would render into a buffer, then put the buffer number (and screen address) into a queue. Once it had processed all the input it would bind and blit to each screen that needs it.
L50[10:10:34] <Izzy> Another option would be a compositing stacking windowing system, which could be interesting. Dividing up the screen space would be another good use for buffers.
L51[10:12:03] <Izzy> You could cleanly do terminal tiling by just allocating multiple buffers per display and having an offset.
L52[10:17:31] <Kristo​pher38> >Izzy: Another option would be a compositing stacking windowing system, which could be…
L53[10:17:31] <Kristo​pher38> I was writing something like this but went overambitious route and wanted to have moving windows
L54[10:17:51] <Elfi> Too bad you can't really get away with reduced colour depth in exchange for greater resolution
L55[10:18:12] <Kristo​pher38> so I'd need to redraw what was under the moved window
L56[10:18:53] <Elfi> Imagine a high-res 16-colour or even monochrome display
L57[10:19:13] <Izzy> if I could trade colour for more text cells I 200% would
L58[10:20:18] <Elfi> I wonder if you could recompile ZZT for openOS lua
L59[10:20:40] <Elfi> Maybe if there was an llvm output for it
L60[11:32:57] <DiamondC​reeper_6> Does anybody know any mods that adds circuits with this tag? https://tinyurl.com/2hnuhyvr
L61[11:47:49] <S​ky> Could that possibly be mekanism ultimate control circuits?
L62[12:47:21] <Ocawes​ome101> izzy: the issue with gpu buffered terminals is they're actually slower than unbuffered ones by a noticeable amount. i tried buffering Monolith's and it was noticeably slower.
L63[12:54:15] <Izzy> buffered in GPU memory?
L64[12:55:45] <Ocawes​ome101> yes
L65[12:55:53] <Izzy> surprising
L66[12:56:13] <Izzy> I guess it'd probably be better for overall throughput but worse for interactive use
L67[12:56:18] <Izzy> might be worth doing for buffering ig
L68[12:56:38] <Izzy> uh, multiplexing, not buffering
L69[12:57:07] <Ocawes​ome101> iirc the reason buffers are slow to draw is they have to sync between client and server. the component call cost is based on how big the buffer is.
L70[12:57:29] <Ocawes​ome101> but while just modifying the buffer, gpu calls are virtually instant
L71[12:57:53] <Amanda> You can sync partially though can't you?
L72[12:57:53] <Ocawes​ome101> this is to prevent people from spamming updates and causing huge amounts of network traffic
L73[12:58:32] <Ocawes​ome101> i don't think it does that. you could split the screen into a bunch of smaller buffers tho, which would speed it up
L74[12:59:08] <Ocawes​ome101> at 50x16 or something like that on a T3 gpu it's actually usable
L75[12:59:40] <Ocawes​ome101> (ULOS's GUI uses one buffer per window and has a terminal emulator that works fairly well)
L76[12:59:51] <Izzy> 80x25 is more than double the raw size of that :|
L77[12:59:55] <Izzy> hmhmhm
L78[13:00:29] <Ocawes​ome101> split each 80x25 group into maybe 5 or 10 buffers and do dirty tracking
L79[13:00:53] <Izzy> divide it into 4 40x12 segments mayb
L80[13:00:54] <Ocawes​ome101> that way you're syncing smaller buffers and not redrawing the whole terminal every update, which is significantly faster
L81[13:01:10] <Izzy> iunno that seems like a lot of complexity
L82[13:01:12] <Izzy> hmm
L83[13:01:15] <Ocawes​ome101> or 10 40x5 segments
L84[13:02:05] <Ocawes​ome101> i'm working on some PoC-ish code that does dirty tracking and whatnot
L85[13:02:14] <Izzy> okay so hear me out
L86[13:02:17] <Ocawes​ome101> for use in a windowing system
L87[13:02:20] <Ocawes​ome101> go ahead
L88[13:02:48] <Izzy> have a currently working display
L89[13:02:56] <Izzy> draw straight to that while it has its timeslice
L90[13:03:24] <Izzy> copy that into a buffer
L91[13:03:33] <Izzy> move onto the next display
L92[13:03:56] <Izzy> every other non-currently-working display would be drawing into a buffer
L93[13:04:25] <Izzy> when a display is switched to, if the buffer is dirty, do one blit then do any drawing while it's the timeslice
L94[13:09:57] <Ocawes​ome101> or just do a blit regardless, because clean buffers are also virtually free
L95[13:10:09] <Izzy> ahhh
L96[13:10:32] <Ocawes​ome101> it's just dirty ones that require syncing that are expensive
L97[13:11:02] <Izzy> man it'd be sweet to be able to copy buffers between GPUs
L98[13:11:09] <Izzy> I don't know what this would be useful for, but
L99[13:16:48] <Izzy> anyway, if a terminal isn't getting updated, don't bother to try to do anything with it
L100[13:23:15] <DiamondC​reeper_6> >Sky: Could that possibly be mekanism ultimate control circuits?
L101[13:23:15] <DiamondC​reeper_6> nope
L102[13:28:26] <Saphire> I'm yet again reminded of existence of JSON-LD
L103[13:30:00] <Ocawes​ome101> amanda: you _could_ sync buffers partially but it's not implemented and according to payonel would be a "huge mess" to do https://tinyurl.com/2lvbb2sa
L104[13:46:35] <Izzy> hmm, well, switchable virtual terminals seems Very Much Possible
L105[13:47:05] <Izzy> that would just require saving and loading screen contents to VRAM
L106[13:48:01] <Izzy> you'd just page one in and one out
L107[13:48:11] <Izzy> or rather the reverse
L108[13:48:28] <Izzy> multiple screens per GPU using buffers may be doable, but ???
L109[13:53:50] <Izzy> not sure it would actually be useful to do that
L110[13:54:45] <Izzy> given re-binding the GPU doesn't erase any state on the screen, you could just switch between them I guess
L111[13:54:52] <Izzy> not great for a GUI but it might do for terminals
L112[13:56:20] <Izzy> binding to a screen takes 0.05s either way
L113[14:06:11] <Amanda> @Nadja~ Someone's messed up the phydics of thermodynamics again~ I'm swearting like a pig, but freezing
L114[14:07:02] <Amanda> I think I finally kicked the cold, but now somebody's playing with my body's thermosense
L115[14:07:07] <Saphire> ...I think schema.org doesn't have a way to define plural people
L116[15:03:06] ⇨ Joins: DBotThePony1 (~Thunderbi@31.220.170.28)
L117[15:03:39] ⇦ Quits: DBotThePony (~Thunderbi@31.220.170.28) (Read error: Connection reset by peer)
L118[15:03:39] *** DBotThePony1 is now known as DBotThePony
L119[15:06:50] ⇦ Quits: progwml6 (~progwml6@45.159.183.209) (Ping timeout: 195 seconds)
L120[15:15:01] ⇨ Joins: Hawk777 (~Hawk777@2607:c000:8299:e400:2bee:857d:6a9e:6ed4)
L121[15:25:11] ⇨ Joins: kikoabdou (~kikoabdou@105.100.105.75)
L122[15:25:27] <kikoabdou> hello
L123[15:25:35] <kikoabdou> gamgam style
L124[15:25:47] ⇦ Quits: kikoabdou (~kikoabdou@105.100.105.75) (Client Quit)
L125[15:41:22] <Na​dja> Amanda: wasn't me!
L126[15:43:40] ⇨ Joins: progwml6 (~progwml6@45.159.183.209)
L127[16:27:50] ⇨ Joins: hello123 (~hello123@ns556296.ip-54-39-28.net)
L128[16:28:47] ⇦ Quits: hello123 (~hello123@ns556296.ip-54-39-28.net) (Client Quit)
L129[16:47:14] <Va​ur> %tonkout
L130[16:47:15] <MichiBot> Fiddlesticks! Va​ur! You beat Forec​aster's previous record of 5 hours, 17 minutes and 29 seconds (By 4 hours, 54 minutes and 45 seconds)! I hope you're happy!
L131[16:47:16] <MichiBot> Va​ur has stolen the tonkout! Tonk has been reset! They gained 0.01 tonk points! plus 0.009 bonus points for consecutive hours! (Reduced to 50% because stealing) Current score: 0.76129. Position #3 Need 0.01376481 more points to pass Compan​ionCube!
L132[16:59:51] ⇦ Quits: tehbeard (~tehesper@66.228.139.228) (Ping timeout: 186 seconds)
L133[17:02:17] <Amanda> How the hell is it only 1300
L134[17:07:20] ⇨ Joins: tehbeard (~tehesper@66.228.139.228)
L135[17:09:35] <stephan48> lies. its already 19:00!
L136[17:10:40] <n​il> i am on the direwolf20 1.12.2 pack with a few added mods like mystcraft ancient warfare and a few others
L137[17:10:41] <n​il> i am inviting any one that want to play on my server (united states west coast) right now its just me and my buddy pm me for server details if you wanna play with us
L138[17:11:00] <luna​r_sam> ok
L139[17:13:46] <n​il> i am really into opencomputers interacting with as many mods as i can i all ready got a computer controlled buildcraft power set up and working on an industrial foregoing mob grinder system
L140[17:16:19] <n​il> https://tinyurl.com/2jdj2g2b
L141[17:16:38] <n​il> the engine animate when theyy are "running"
L142[17:18:26] <n​il> https://tinyurl.com/2glzyl8t
L143[20:02:55] <Amanda> Who spilled all these tireds around the place, it's only 1600
L144[20:59:32] <PwnagePineap​ple (He/Him)> %tonk
L145[20:59:32] <MichiBot> Gadsbudlikins! PwnagePineap​ple (He/Him)! You beat Va​ur's previous record of <0 (By 4 hours, 12 minutes and 17 seconds)! I hope you're happy!
L146[20:59:33] <MichiBot> PwnagePineapple (He/Him)'s new record is 4 hours, 12 minutes and 17 seconds! PwnagePineapple (He/Him) also gained 0.0042 tonk points for stealing the tonk. Position #8. Need 0.116918 more points to pass walksanato​r. B⃢ot.!
L147[23:18:33] ⇦ Quits: Vexatos (~Vexatos@p200300eaef1a5768c3ee193d72b4af83.dip0.t-ipconnect.de) (Quit: Insert quantum chemistry joke here)
L148[23:52:56] <Saphire> Rawr
L149[23:53:05] <Saphire> %tonk
L150[23:53:05] <MichiBot> Sap​hire: You should probably read this: https://michibot.pc-logix.com/tonk
L151[23:53:13] <Saphire> Hmph
L152[23:53:55] <Saphire> %tonk 528D1
L153[23:53:55] <MichiBot> I'm sorry Saphire, you were not able to beat PwnagePineapple (He/Him)'s record of 4 hours, 12 minutes and 17 seconds this time. 2 hours, 54 minutes and 22 seconds were wasted! Missed by 1 hour, 17 minutes and 54 seconds!
L154[23:54:02] <Saphire> Wait what
L155[23:54:11] <Saphire> Oh
<<Prev Next>> Scroll to Top