IRCLogs • ArmaNelgTron.tk

Searching from 2023-08-19 00:00:00 to 2023-08-19 23:59:59.999999.
Query completed in 1.78 seconds
[2023-08-19 00:02:36] <Lucifer_arma> some quick research on the subject tells me that OpenCL is the cross-platform approach here, and that the simulation has to be possible with linear algebra.  Basically, the more of it that can be done with matrixes, the more benefit there is to be had, but there's a minimum amount that has to be possible or else there's no benefit and a lot of cost
[2023-08-19 00:02:36] <armagetronbridge> 10irc:Lucifer_arma| some quick research on the subject tells me that OpenCL is the cross-platform approach here, and that the simulation has to be possible with linear algebra.  Basically, the more of it that can be done with matrixes, the more benefit there is to be had, but there's a minimum amount that has to be possible or else there's no benefit and a lot of cost
[2023-08-19 00:03:20] <Lucifer_arma> so the question is, how much can latency be reduced by making the simulation itself faster?  Is it 5%, or 25%?
[2023-08-19 00:03:20] <armagetronbridge> 10irc:Lucifer_arma| so the question is, how much can latency be reduced by making the simulation itself faster?  Is it 5%, or 25%?
[2023-08-19 00:04:04] <Lucifer_arma> since I don't have any linear algebra skills, let alone the ability to simulate light cycles with it, this is where I have to bug off :)
[2023-08-19 00:04:04] <armagetronbridge> 10irc:Lucifer_arma| since I don't have any linear algebra skills, let alone the ability to simulate light cycles with it, this is where I have to bug off :)
[2023-08-19 02:00:13] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 02:05:04] <-- Armanelgtron has quit (No Ping reply in 270 seconds.)
[2023-08-19 02:07:31] --> Armanelgtron has joined the channel
[2023-08-19 02:07:32] -!- cadmium.libera.chat set mode #armagetron +nt
[2023-08-19 02:07:32] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 02:07:40] --> Armanelgtron has joined the channel
[2023-08-19 02:07:40] -!- Topic for #armagetron is "Armagetron Advanced | http://www.armagetronad.org/ | Welcome to IRC"
[2023-08-19 02:07:40] -!- Topic set by ChanServ!services@services.oftc.net on 2022-12-21 00:36:08 UTC
[2023-08-19 02:07:42] -!- weber.oftc.net set mode #armagetron +nt
[2023-08-19 02:07:42] -!- Channel #armagetron created on 2021-04-20 19:56:37 UTC
[2023-08-19 03:37:19] <armagetron-bridge> 14discord:olive| @ppotter just scoop the pop once a day
[2023-08-19 03:37:20] <armagetronbridge> 14discord:olive| @ppotter just scoop the pop once a day
[2023-08-19 06:14:45] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 06:18:05] <-- Armanelgtron has quit (No Ping reply in 240 seconds.)
[2023-08-19 06:19:18] --> Armanelgtron has joined the channel
[2023-08-19 06:19:18] -!- Topic for #armagetron is "Armagetron Advanced | http://www.armagetronad.org/ | Welcome to IRC"
[2023-08-19 06:19:18] -!- Topic set by ChanServ!services@services.oftc.net on 2022-12-21 00:36:08 UTC
[2023-08-19 06:19:18] -!- weber.oftc.net set mode #armagetron +nt
[2023-08-19 06:19:18] -!- Channel #armagetron created on 2021-04-20 19:56:37 UTC
[2023-08-19 06:19:23] --> Armanelgtron has joined the channel
[2023-08-19 06:19:24] -!- osmium.libera.chat set mode #armagetron +nt
[2023-08-19 06:19:24] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 08:11:09] --> monr0e has joined the channel
[2023-08-19 08:19:14] <armagetronbridge> 08discord:delinquent| Actually would too. My passport expired last year and it's a pisstake to get it renewed rn
[2023-08-19 08:19:14] <armagetron-bridge> 08discord:delinquent| Actually would too. My passport expired last year and it's a pisstake to get it renewed rn
[2023-08-19 09:50:58] <armagetronbridge> 14discord:olive| what did they do?
[2023-08-19 09:50:58] <armagetron-bridge> 14discord:olive| what did they do?
[2023-08-19 09:51:33] <armagetronbridge> 14discord:olive| is there no fast track to squeeze you out? i'll cover it
[2023-08-19 09:51:34] <armagetron-bridge> 14discord:olive| is there no fast track to squeeze you out? i'll cover it
[2023-08-19 09:53:47] <armagetronbridge> 08discord:delinquent| Because it's been so long, and because I have some other ID issues and no family to verify my identity, I have to go through a rather long interviewing and vetting process to ensure I actually *am* Mr Doll, and not Mr Reeves pretending to be Mr Doll
[2023-08-19 09:53:48] <armagetron-bridge> 08discord:delinquent| Because it's been so long, and because I have some other ID issues and no family to verify my identity, I have to go through a rather long interviewing and vetting process to ensure I actually *am* Mr Doll, and not Mr Reeves pretending to be Mr Doll
[2023-08-19 09:59:04] <armagetronbridge> 14discord:olive| don't you have a social security number or something that is tied to your identity
[2023-08-19 09:59:04] <armagetron-bridge> 14discord:olive| don't you have a social security number or something that is tied to your identity
[2023-08-19 10:00:55] <armagetronbridge> 14discord:olive| there must be some kind of paper trail leading back to your birth, not sure how it works in England - are they that overly bureaucratic?
[2023-08-19 10:00:57] <armagetron-bridge> 14discord:olive| there must be some kind of paper trail leading back to your birth, not sure how it works in England - are they that overly bureaucratic?
[2023-08-19 10:03:21] <armagetronbridge> 08discord:delinquent| In the UK, your NI number isn't proof of identity
[2023-08-19 10:03:21] <armagetron-bridge> 08discord:delinquent| In the UK, your NI number isn't proof of identity
[2023-08-19 10:04:27] <armagetronbridge> 03discord:ninjapotato1080| does a drivers license work if you have one?
[2023-08-19 10:04:28] <armagetron-bridge> 03discord:ninjapotato1080| does a drivers license work if you have one?
[2023-08-19 10:04:35] <armagetronbridge> 08discord:delinquent| The problem stems from being on my mother's passport when I was younger. Since hers was a German passport, it makes life very difficult if I let it lapse. It's a pain, but there it is
[2023-08-19 10:04:36] <armagetron-bridge> 08discord:delinquent| The problem stems from being on my mother's passport when I was younger. Since hers was a German passport, it makes life very difficult if I let it lapse. It's a pain, but there it is
[2023-08-19 10:05:01] <armagetronbridge> 08discord:delinquent| that's in spite of the fact that I never *had* german citizenship too, I never understood how she got aroudn that
[2023-08-19 10:05:02] <armagetron-bridge> 08discord:delinquent| that's in spite of the fact that I never *had* german citizenship too, I never understood how she got aroudn that
[2023-08-19 10:10:14] <armagetronbridge> 08discord:delinquent| > arma already does a bit of firehosing
[2023-08-19 10:10:14] <armagetron-bridge> 08discord:delinquent| > arma already does a bit of firehosing
[2023-08-19 10:10:15] <armagetronbridge> 08discord:delinquent| Hm. I think I was more suggesting that every packet be retransmitted multiple times regardless of it's ACK status. With the bandwidth behind the modern internet, and the possibility of dscp values now in play, that alone migh tbe enough to mitigate the worst of the issues we've been facing.
[2023-08-19 10:10:15] <armagetron-bridge> 08discord:delinquent| Hm. I think I was more suggesting that every packet be retransmitted multiple times regardless of it's ACK status. With the bandwidth behind the modern internet, and the possibility of dscp values now in play, that alone migh tbe enough to mitigate the worst of the issues we've been facing.
[2023-08-19 10:10:16] <armagetronbridge> 08discord:delinquent| Speaking of packet stuff, I've been setting MAX_IN_RATE and MAX_OUT_RATE on my machines, either to 64 or 128, I forget which. My primary concern behind this was avoidance of UDP fragmentation, but I'm also aware that I was basing this off a very old interpretation of where and when fragmentation was likely to occur. At any rate, in theory this should restrict packet size (althoug <clipped message>
[2023-08-19 10:10:17] <armagetron-bridge> 08discord:delinquent| Speaking of packet stuff, I've been setting MAX_IN_RATE and MAX_OUT_RATE on my machines, either to 64 or 128, I forget which. My primary concern behind this was avoidance of UDP fragmentation, but I'm also aware that I was basing this off a very old interpretation of where and when fragmentation was likely to occur. At any rate, in theory this should restrict packet size (althoug <clipped message>
[2023-08-19 10:10:17] <armagetronbridge> 08discord:delinquent| h I question whether it might actually force fragmentation instead...)
[2023-08-19 10:10:18] <armagetron-bridge> 08discord:delinquent| h I question whether it might actually force fragmentation instead...)
[2023-08-19 10:10:18] <armagetronbridge> 08discord:delinquent| --- Scratch that last, I just learned that 1500b is the default, with 1397b being usable. I feel very uneducated rn. @Deso I was wrong about *all* of that lmao
[2023-08-19 10:10:19] <armagetron-bridge> 08discord:delinquent| --- Scratch that last, I just learned that 1500b is the default, with 1397b being usable. I feel very uneducated rn. @Deso I was wrong about *all* of that lmao
[2023-08-19 10:10:20] <armagetronbridge> 08discord:delinquent| --- Even worse, apparently neither max-in nor max-out configs actually *do* anything. @Z-Man any chance we can look at this at the same time as DSCP stuff? At the moment, connection speeds are entirely mandated by whether or not the user has selected dialup, isdn, or dsl in their initial setup, and the limit of that in/out rate is 64/16. I've manually set that to 1000 in my user. <clipped message>
[2023-08-19 10:10:21] <armagetron-bridge> 08discord:delinquent| --- Even worse, apparently neither max-in nor max-out configs actually *do* anything. @Z-Man any chance we can look at this at the same time as DSCP stuff? At the moment, connection speeds are entirely mandated by whether or not the user has selected dialup, isdn, or dsl in their initial setup, and the limit of that in/out rate is 64/16. I've manually set that to 1000 in my user. <clipped message>
[2023-08-19 10:10:21] <armagetronbridge> 08discord:delinquent| cfg, but either enabling the use of MAX_IN_RATE and MAX_OUT_RATE or even just an updated bandwidth limit would be helpful.
[2023-08-19 10:10:22] <armagetron-bridge> 08discord:delinquent| cfg, but either enabling the use of MAX_IN_RATE and MAX_OUT_RATE or even just an updated bandwidth limit would be helpful.
[2023-08-19 10:10:22] <armagetronbridge> 08discord:delinquent| > I'm not suggesting that bare metal...
[2023-08-19 10:10:23] <armagetron-bridge> 08discord:delinquent| > I'm not suggesting that bare metal...
[2023-08-19 10:10:24] <armagetronbridge> 08discord:delinquent| No, I know. What I was really getting at is that armagetron has such a low overhead that performance really isn't down to the underlying infrastructure anymore, at least the immediate infrastructure. I think we really just need to accept that the landscape of the internet has significantly changed, with the advent of streaming, video-conferencing, p2p distribution (which isn't ju <clipped message>
[2023-08-19 10:10:24] <armagetron-bridge> 08discord:delinquent| No, I know. What I was really getting at is that armagetron has such a low overhead that performance really isn't down to the underlying infrastructure anymore, at least the immediate infrastructure. I think we really just need to accept that the landscape of the internet has significantly changed, with the advent of streaming, video-conferencing, p2p distribution (which isn't ju <clipped message>
[2023-08-19 10:10:25] <armagetronbridge> 08discord:delinquent| st torrents anymore - a lot of game installers use this approach), even the size of regular HTTP downloads have gotten significantly higher. Core internet infrastructure has also improved, yes, but with so much traffic to schedule it is hardly surprising that we are seeing an increase in latency or loss induced incidents. That's why I put forth the idea of priority in the first p <clipped message>
[2023-08-19 10:10:26] <armagetron-bridge> 08discord:delinquent| st torrents anymore - a lot of game installers use this approach), even the size of regular HTTP downloads have gotten significantly higher. Core internet infrastructure has also improved, yes, but with so much traffic to schedule it is hardly surprising that we are seeing an increase in latency or loss induced incidents. That's why I put forth the idea of priority in the first p <clipped message>
[2023-08-19 10:10:26] <armagetronbridge> 08discord:delinquent| lace - in fact, this is something that most games incorporate now, it's no longer limited to the likes of RTC.
[2023-08-19 10:10:27] <armagetron-bridge> 08discord:delinquent| lace - in fact, this is something that most games incorporate now, it's no longer limited to the likes of RTC.
[2023-08-19 10:10:28] <armagetronbridge> 08discord:delinquent| However, there *is* a networking and thread-timing overhead with a shared-resource implementation. I can actually demonstrate this in my own DC, although other solutions have a greater impact. Hypervisors are probably the worst offenders, since they add an additional hosting layer between the runtime and the instruction execution - and worse is that there is always the vague poss <clipped message>
[2023-08-19 10:10:28] <armagetron-bridge> 08discord:delinquent| However, there *is* a networking and thread-timing overhead with a shared-resource implementation. I can actually demonstrate this in my own DC, although other solutions have a greater impact. Hypervisors are probably the worst offenders, since they add an additional hosting layer between the runtime and the instruction execution - and worse is that there is always the vague poss <clipped message>
[2023-08-19 10:10:29] <armagetronbridge> 08discord:delinquent| ibility (although I think protections are in place these days) of thread mismatching, where an expected instruction set is launched on a thread other than the one it was expecting. That's only really an issue in super-multi-threaded applications, like ones designed to run on eight or more cores, where there's also a thread controller. Those sorts of applications are admittedly ra <clipped message>
[2023-08-19 10:10:30] <armagetron-bridge> 08discord:delinquent| ibility (although I think protections are in place these days) of thread mismatching, where an expected instruction set is launched on a thread other than the one it was expecting. That's only really an issue in super-multi-threaded applications, like ones designed to run on eight or more cores, where there's also a thread controller. Those sorts of applications are admittedly ra <clipped message>
[2023-08-19 10:10:30] <armagetronbridge> 08discord:delinquent| re, though. Regardless, adding the overhead is actually noticeable in armagetron - hence my decision to use bare metal here in the UK, and go for a non-cyclical provider in the US.
[2023-08-19 10:10:31] <armagetron-bridge> 08discord:delinquent| re, though. Regardless, adding the overhead is actually noticeable in armagetron - hence my decision to use bare metal here in the UK, and go for a non-cyclical provider in the US.
[2023-08-19 10:10:31] <armagetronbridge> 08discord:delinquent| > async, "Is it even worth it"
[2023-08-19 10:10:32] <armagetron-bridge> 08discord:delinquent| > async, "Is it even worth it"
[2023-08-19 10:10:33] <armagetronbridge> 08discord:delinquent| I'd have to guess no, there's not really any benefit given the processing power of a modern machine. Given that we can and do run several servers on one machine, also, I would argue that taking this route could potentially negatively impact performance too. That is, of course, operating on the assumption that the different running processes are given specific thread assignments a <clipped message>
[2023-08-19 10:10:34] <armagetron-bridge> 08discord:delinquent| I'd have to guess no, there's not really any benefit given the processing power of a modern machine. Given that we can and do run several servers on one machine, also, I would argue that taking this route could potentially negatively impact performance too. That is, of course, operating on the assumption that the different running processes are given specific thread assignments a <clipped message>
[2023-08-19 10:10:34] <armagetronbridge> 08discord:delinquent| nd that armagetron operates as one, single, continuous task. I have to confess, this is getting into territory I can only vaguely speculate on and I'm doing a *lot* of guesswork. 
[2023-08-19 10:10:35] <armagetron-bridge> 08discord:delinquent| nd that armagetron operates as one, single, continuous task. I have to confess, this is getting into territory I can only vaguely speculate on and I'm doing a *lot* of guesswork. 
[2023-08-19 10:10:35] <armagetronbridge> 08discord:delinquent| > AI
[2023-08-19 10:10:36] <armagetron-bridge> 08discord:delinquent| > AI
[2023-08-19 10:10:37] <armagetronbridge> 08discord:delinquent| crow_of_judgement.png
[2023-08-19 10:10:37] <armagetron-bridge> 08discord:delinquent| crow_of_judgement.png
[2023-08-19 10:10:38] <armagetronbridge> 08discord:delinquent| Ok, conceivably. That said, acceleration has been in and out of favour many times over the years. I question whether it's worth the effort to try and incorporate it - especially when the cost of hosting would jump considerably. At present, I pay abotu forty quid a month for the US host. If I added AI provision, that cost would quadruple - even if I only wanted the hardware. Even  <clipped message>
[2023-08-19 10:10:39] <armagetron-bridge> 08discord:delinquent| Ok, conceivably. That said, acceleration has been in and out of favour many times over the years. I question whether it's worth the effort to try and incorporate it - especially when the cost of hosting would jump considerably. At present, I pay abotu forty quid a month for the US host. If I added AI provision, that cost would quadruple - even if I only wanted the hardware. Even  <clipped message>
[2023-08-19 10:10:39] <armagetronbridge> 08discord:delinquent| adding a graphics card to a vps is somewhat cost-prohibitive - and now I'm going to be managing a whole heap of other servers for another community member who is bugging out for a bit, that cost is *ouch*. It might be better instead to focus on better utilisation of modern CPU capabilities - they're certainly leaps and bounds ahead of what we had twenty years ago. With ARM coming <clipped message>
[2023-08-19 10:10:40] <armagetron-bridge> 08discord:delinquent| adding a graphics card to a vps is somewhat cost-prohibitive - and now I'm going to be managing a whole heap of other servers for another community member who is bugging out for a bit, that cost is *ouch*. It might be better instead to focus on better utilisation of modern CPU capabilities - they're certainly leaps and bounds ahead of what we had twenty years ago. With ARM coming <clipped message>
[2023-08-19 10:10:41] <armagetronbridge> 08discord:delinquent|  into play too, which is what all my hosts run on, there's even more scope for improvement.
[2023-08-19 10:10:41] <armagetron-bridge> 08discord:delinquent|  into play too, which is what all my hosts run on, there's even more scope for improvement.
[2023-08-19 10:35:25] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 10:36:06] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 10:39:01] --> Armanelgtron has joined the channel
[2023-08-19 10:39:03] -!- cadmium.libera.chat set mode #armagetron +nt
[2023-08-19 10:39:03] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 10:44:15] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 10:47:25] --> Armanelgtron has joined the channel
[2023-08-19 10:47:25] -!- Topic for #armagetron is "Armagetron Advanced | http://www.armagetronad.org/ | Welcome to IRC"
[2023-08-19 10:47:25] -!- Topic set by ChanServ!services@services.oftc.net on 2022-12-21 00:36:08 UTC
[2023-08-19 10:47:26] -!- weber.oftc.net set mode #armagetron +nt
[2023-08-19 10:47:26] -!- Channel #armagetron created on 2021-04-20 19:56:37 UTC
[2023-08-19 10:47:27] --> Armanelgtron has joined the channel
[2023-08-19 10:47:29] -!- tungsten.libera.chat set mode #armagetron +nt
[2023-08-19 10:47:29] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 11:03:32] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 11:07:07] <-- Armanelgtron has quit (No Ping reply in 270 seconds.)
[2023-08-19 11:08:04] --> Armanelgtron has joined the channel
[2023-08-19 11:08:04] -!- Topic for #armagetron is "Armagetron Advanced | http://www.armagetronad.org/ | Welcome to IRC"
[2023-08-19 11:08:04] -!- Topic set by ChanServ!services@services.oftc.net on 2022-12-21 00:36:08 UTC
[2023-08-19 11:08:05] -!- weber.oftc.net set mode #armagetron +nt
[2023-08-19 11:08:05] -!- Channel #armagetron created on 2021-04-20 19:56:37 UTC
[2023-08-19 11:08:45] --> Armanelgtron has joined the channel
[2023-08-19 11:08:46] -!- silver.libera.chat set mode #armagetron +nt
[2023-08-19 11:08:46] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 11:16:00] <armagetron-bridge> 14discord:olive| haha, i'll watch your cat instead
[2023-08-19 11:16:00] <armagetronbridge> 14discord:olive| haha, i'll watch your cat instead
[2023-08-19 11:17:04] <armagetron-bridge> 14discord:olive| I would ask my younger brother but I'm sure he will give Mila enough attention and play games instead, more than me
[2023-08-19 11:17:05] <armagetronbridge> 14discord:olive| I would ask my younger brother but I'm sure he will give Mila enough attention and play games instead, more than me
[2023-08-19 11:17:19] <armagetron-bridge> 14discord:olive| she's a sensitive cat, I'd rather put faith in tronners
[2023-08-19 11:17:20] <armagetronbridge> 14discord:olive| she's a sensitive cat, I'd rather put faith in tronners
[2023-08-19 12:42:12] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 12:44:42] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 12:45:41] --> Armanelgtron_ has joined the channel
[2023-08-19 12:45:41] -!- Topic for #armagetron is "Armagetron Advanced | http://www.armagetronad.org/ | Welcome to IRC"
[2023-08-19 12:45:41] -!- Topic set by ChanServ!services@services.oftc.net on 2022-12-21 00:36:08 UTC
[2023-08-19 12:45:41] -!- Armanelgtron changed nick to Guest9381
[2023-08-19 12:45:41] -!- Armanelgtron changed nick to Armanelgtron
[2023-08-19 12:45:42] -!- weber.oftc.net set mode #armagetron +nt
[2023-08-19 12:45:42] -!- Channel #armagetron created on 2021-04-20 19:56:37 UTC
[2023-08-19 12:45:56] <-- Guest9381 has quit (Ping timeout: 480 seconds)
[2023-08-19 12:45:58] --> Armanelgtron has joined the channel
[2023-08-19 12:45:59] -!- mercury.libera.chat set mode #armagetron +nt
[2023-08-19 12:45:59] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 13:24:43] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 13:24:43] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 13:27:32] --> Armanelgtron_ has joined the channel
[2023-08-19 13:27:32] -!- Topic for #armagetron is "Armagetron Advanced | http://www.armagetronad.org/ | Welcome to IRC"
[2023-08-19 13:27:32] -!- Topic set by ChanServ!services@services.oftc.net on 2022-12-21 00:36:08 UTC
[2023-08-19 13:27:33] -!- Armanelgtron changed nick to Guest9385
[2023-08-19 13:27:33] -!- Armanelgtron changed nick to Armanelgtron
[2023-08-19 13:27:34] -!- weber.oftc.net set mode #armagetron +nt
[2023-08-19 13:27:34] -!- Channel #armagetron created on 2021-04-20 19:56:37 UTC
[2023-08-19 13:27:46] <-- Guest9385 has quit (Ping timeout: 480 seconds)
[2023-08-19 13:28:36] --> Armanelgtron has joined the channel
[2023-08-19 13:28:37] -!- osmium.libera.chat set mode #armagetron +nt
[2023-08-19 13:28:37] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 15:14:25] <-- Armanelgtron has quit (No Ping reply in 180 seconds.)
[2023-08-19 15:15:45] <-- Armanelgtron has quit (No Ping reply in 210 seconds.)
[2023-08-19 15:18:26] --> Armanelgtron has joined the channel
[2023-08-19 15:18:27] -!- platinum.libera.chat set mode #armagetron +nt
[2023-08-19 15:18:27] -!- Channel #armagetron created on 2021-05-20 17:23:14 UTC
[2023-08-19 15:19:23] --> Armanelgtron has joined the channel
[2023-08-19 15:19:23] -!- Topic for #armagetron is "Armagetron Advanced | http://www.armagetronad.org/ | Welcome to IRC"
[2023-08-19 15:19:23] -!- Topic set by ChanServ!services@services.oftc.net on 2022-12-21 00:36:08 UTC
[2023-08-19 15:19:25] -!- weber.oftc.net set mode #armagetron +nt
[2023-08-19 15:19:25] -!- Channel #armagetron created on 2021-04-20 19:56:37 UTC
[2023-08-19 16:27:29] <armagetronbridge> 15discord:Nélg| @northernscrub 
[2023-08-19 16:27:30] <armagetron-bridge> 15discord:Nélg| @northernscrub 
[2023-08-19 16:27:30] <armagetron-bridge> 15discord:Nélg| > Even worse, apparently neither max-in nor max-out configs actually do anything.
[2023-08-19 16:27:30] <armagetronbridge> 15discord:Nélg| > Even worse, apparently neither max-in nor max-out configs actually do anything.
[2023-08-19 16:27:31] <armagetron-bridge> 15discord:Nélg| max in rate and max out rate definitely does *something*. Try setting `MAX_OUT_RATE` to, say, 1, on a server, and then joining it. You'll notice you'll receive the settings ("CYCLE_ACCEL changed from 10 to 20 on server order") at a much slower rate. Or `MAX_IN_RATE` to 1 on a client. Same effect.
[2023-08-19 16:27:31] <armagetronbridge> 15discord:Nélg| max in rate and max out rate definitely does *something*. Try setting `MAX_OUT_RATE` to, say, 1, on a server, and then joining it. You'll notice you'll receive the settings ("CYCLE_ACCEL changed from 10 to 20 on server order") at a much slower rate. Or `MAX_IN_RATE` to 1 on a client. Same effect.
[2023-08-19 16:28:03] <armagetronbridge> 15discord:Nélg| and that extends to receiving gameplay objects once joining, too
[2023-08-19 16:28:03] <armagetron-bridge> 15discord:Nélg| and that extends to receiving gameplay objects once joining, too
[2023-08-19 16:34:17] <-- monr0e has quit (Ping timeout: 258 seconds)
[2023-08-19 22:36:23] <Lucifer_arma> I definitely think marking the packets high priority (or whatever it is, I know we talked about it back in the day when it was first appearing, and iirc the decision was to wait and see what happened, because we're idealogically opposed to preferential treatment for network traffic, but the world is what it is, and preferential treatment happens, we deserve it :)  )
[2023-08-19 22:36:23] <armagetronbridge> 10irc:Lucifer_arma| I definitely think marking the packets high priority (or whatever it is, I know we talked about it back in the day when it was first appearing, and iirc the decision was to wait and see what happened, because we're idealogically opposed to preferential treatment for network traffic, but the world is what it is, and preferential treatment happens, we deserve it :)  )
[2023-08-19 22:36:34] <Lucifer_arma> is the Right Thing To Do now.
[2023-08-19 22:36:35] <armagetronbridge> 10irc:Lucifer_arma| is the Right Thing To Do now.
[2023-08-19 22:37:06] <Lucifer_arma> I'm not sure what you mean by fragmentation.  I'm seeing it as two possible situations:
[2023-08-19 22:37:06] <armagetronbridge> 10irc:Lucifer_arma| I'm not sure what you mean by fragmentation.  I'm seeing it as two possible situations:
[2023-08-19 22:37:41] <Lucifer_arma> 1)  Arma deliberately breaks each message into multiple packets.  The other side needs all the different packets to reassemble the message, and if it's missing one, it can't.
[2023-08-19 22:37:42] <armagetronbridge> 10irc:Lucifer_arma| 1)  Arma deliberately breaks each message into multiple packets.  The other side needs all the different packets to reassemble the message, and if it's missing one, it can't.
[2023-08-19 22:39:14] <Lucifer_arma> 2)  Where I was trying to go, at levels lower than sockets themselves, a UDP message can get broken up at the hardware layer.  Say a packet-switching device, or a wifi router that's receiving signals from more than one device at a time.  In this case, each device waits until it's received the whole datagram before it passes it on, but if something happens where some of the datagram is lost, the entire thing is discarded
[2023-08-19 22:39:15] <armagetronbridge> 10irc:Lucifer_arma| 2)  Where I was trying to go, at levels lower than sockets themselves, a UDP message can get broken up at the hardware layer.  Say a packet-switching device, or a wifi router that's receiving signals from more than one device at a time.  In this case, each device waits until it's received the whole datagram before it passes it on, but if something happens where some of the datagr <clipped message>
[2023-08-19 22:39:16] <armagetronbridge> 10irc:Lucifer_arma| am is lost, the entire thing is discarded
[2023-08-19 22:40:07] <Lucifer_arma> Each of these possibilities has opposite solutions.  For 1), make the datagrams bigger, and for 2) make them smaller.  But this also gets at what I was talking about about infrastructure and the importance of wired connections.
[2023-08-19 22:40:07] <armagetronbridge> 10irc:Lucifer_arma| Each of these possibilities has opposite solutions.  For 1), make the datagrams bigger, and for 2) make them smaller.  But this also gets at what I was talking about about infrastructure and the importance of wired connections.
[2023-08-19 22:40:45] <Lucifer_arma> Wifi always has more packet loss, no matter how fast it is.  It's the nature of the thing.  (Yes, I realize that physically, there's not a lot of difference between wireless and wired, just the medium, but let's ignore that)
[2023-08-19 22:40:46] <armagetronbridge> 10irc:Lucifer_arma| Wifi always has more packet loss, no matter how fast it is.  It's the nature of the thing.  (Yes, I realize that physically, there's not a lot of difference between wireless and wired, just the medium, but let's ignore that)
[2023-08-19 22:42:00] <Lucifer_arma> The main differences between the various wifi versions involve error correcting and speed, so 802.11N, for example, besides being gigabit, also includes superior error correcting algorithms, so there's less packet loss
[2023-08-19 22:42:00] <armagetronbridge> 10irc:Lucifer_arma| The main differences between the various wifi versions involve error correcting and speed, so 802.11N, for example, besides being gigabit, also includes superior error correcting algorithms, so there's less packet loss
[2023-08-19 22:42:48] <Lucifer_arma> Cable internet is shared with your neighborhood, so the protocol has to allow traffic flowing from multiple computers over the same shared physical cable.  This is why you notice it behaving in little spurts, I believe.
[2023-08-19 22:42:49] <armagetronbridge> 10irc:Lucifer_arma| Cable internet is shared with your neighborhood, so the protocol has to allow traffic flowing from multiple computers over the same shared physical cable.  This is why you notice it behaving in little spurts, I believe.
[2023-08-19 22:43:34] <Lucifer_arma> Everything below the socket layer, we have no control over in source code, obviously, but individuals do.  That's why locating your server in a datacenter, even on a VPS, is probably better than hosting it off your bare metal cable internet.
[2023-08-19 22:43:35] <armagetronbridge> 10irc:Lucifer_arma| Everything below the socket layer, we have no control over in source code, obviously, but individuals do.  That's why locating your server in a datacenter, even on a VPS, is probably better than hosting it off your bare metal cable internet.
[2023-08-19 22:44:01] <Lucifer_arma> For players, well, I want to play well over wifi, period.
[2023-08-19 22:44:01] <armagetronbridge> 10irc:Lucifer_arma| For players, well, I want to play well over wifi, period.
[2023-08-19 22:44:52] <Lucifer_arma> Encryption, especially over wifi, also adds some latency.  I think my last attempt to measure it, which wasn't terribly scientific, came in around 10-15ms.  That's by looking at a difference in ping, as reported by arma, but on a cable internet connection, so ymmv
[2023-08-19 22:44:53] <armagetronbridge> 10irc:Lucifer_arma| Encryption, especially over wifi, also adds some latency.  I think my last attempt to measure it, which wasn't terribly scientific, came in around 10-15ms.  That's by looking at a difference in ping, as reported by arma, but on a cable internet connection, so ymmv
[2023-08-19 22:45:26] <Lucifer_arma> Point being there was a noticeable difference in ping, but once you hit pings about 70 or so, that difference isn't noticeable ingame.
[2023-08-19 22:45:26] <armagetronbridge> 10irc:Lucifer_arma| Point being there was a noticeable difference in ping, but once you hit pings about 70 or so, that difference isn't noticeable ingame.
[2023-08-19 22:46:20] <Lucifer_arma> I'm generally opposed to my AI-acceleration idea, actually.  :)  I think it's be neat technology to have, you know, a game simulation that uses it and all, but I can't imagine it actually benefitting arma enough to be worth the effort.
[2023-08-19 22:46:20] <armagetronbridge> 10irc:Lucifer_arma| I'm generally opposed to my AI-acceleration idea, actually.  :)  I think it's be neat technology to have, you know, a game simulation that uses it and all, but I can't imagine it actually benefitting arma enough to be worth the effort.
[2023-08-19 22:47:03] <Lucifer_arma> So if some google summer of code monkey wanted to do it, and there was some benefit, and it didn't result in a maintenance nightmare, I could be persuaded.  Other than that, yeah, you're right about every reason you gave for not doing it, and there's probably even more beyond that
[2023-08-19 22:47:04] <armagetronbridge> 10irc:Lucifer_arma| So if some google summer of code monkey wanted to do it, and there was some benefit, and it didn't result in a maintenance nightmare, I could be persuaded.  Other than that, yeah, you're right about every reason you gave for not doing it, and there's probably even more beyond that
[2023-08-19 22:47:12] <Lucifer_arma> Does google still do SoC?
[2023-08-19 22:47:13] <armagetronbridge> 10irc:Lucifer_arma| Does google still do SoC?
[2023-08-19 22:50:38] <Lucifer_arma> As for multiprocessing, I really think that for the sizes of the playing fields and numbers of players we're seeing right now, it's not worth it.  In the past, before multicore computers were common, but also when our player base was the biggest (I admit I haven't logged in in a few years though), it would have been worth it for some game types, like racing and sumo.
[2023-08-19 22:50:38] <armagetronbridge> 10irc:Lucifer_arma| As for multiprocessing, I really think that for the sizes of the playing fields and numbers of players we're seeing right now, it's not worth it.  In the past, before multicore computers were common, but also when our player base was the biggest (I admit I haven't logged in in a few years though), it would have been worth it for some game types, like racing and sumo.
[2023-08-19 22:51:36] <Lucifer_arma> And, of course, if we ever get to a point where we have multiple grids in the same server, like different levels, or significant separation between them, or we simply have a huge grid with lots of players, then I think going multiprocessing is the thing to do
[2023-08-19 22:51:36] <armagetronbridge> 10irc:Lucifer_arma| And, of course, if we ever get to a point where we have multiple grids in the same server, like different levels, or significant separation between them, or we simply have a huge grid with lots of players, then I think going multiprocessing is the thing to do
[2023-08-19 22:53:03] <Lucifer_arma> I've always been in favor of the most important realtime packets being transmitted like 3 times right away.  I vaguely remember z-man either already doing that, or talking about adding it, but that conversation was so long ago that I only remember that it happened, nothing else from it
[2023-08-19 22:53:03] <armagetronbridge> 10irc:Lucifer_arma| I've always been in favor of the most important realtime packets being transmitted like 3 times right away.  I vaguely remember z-man either already doing that, or talking about adding it, but that conversation was so long ago that I only remember that it happened, nothing else from it
[2023-08-19 22:54:17] <Lucifer_arma> I'm also in favor of rearranging the conversation.  Instead of running multiple server processes, rearrange the server so that one server manages several game grids.  There's a lot of reasons we can't do that right now, but in my perfect world, that's what we'd do
[2023-08-19 22:54:17] <armagetronbridge> 10irc:Lucifer_arma| I'm also in favor of rearranging the conversation.  Instead of running multiple server processes, rearrange the server so that one server manages several game grids.  There's a lot of reasons we can't do that right now, but in my perfect world, that's what we'd do
[2023-08-19 22:54:47] <Lucifer_arma> That server would fork a new process for each new game grid when players join it, and it would allow a popular server to start a new game grid instead of just being full and disallowing new players to join
[2023-08-19 22:54:47] <armagetronbridge> 10irc:Lucifer_arma| That server would fork a new process for each new game grid when players join it, and it would allow a popular server to start a new game grid instead of just being full and disallowing new players to join
[2023-08-19 22:56:14] <Lucifer_arma> ofc, that would stress the network layer quite a bit, but I can also see a situation where it's superior to hosting multiple servers on one machine.  A network layer optimized for that kind of use should outperform four separate instances of a network layer optimized for a single game instance.
[2023-08-19 22:56:14] <armagetronbridge> 10irc:Lucifer_arma| ofc, that would stress the network layer quite a bit, but I can also see a situation where it's superior to hosting multiple servers on one machine.  A network layer optimized for that kind of use should outperform four separate instances of a network layer optimized for a single game instance.
[2023-08-19 22:56:55] <Lucifer_arma> on a side note, I'm 60 days sober today :)
[2023-08-19 22:56:55] <armagetronbridge> 10irc:Lucifer_arma| on a side note, I'm 60 days sober today :)
View entire month
DISCLAIMER: These logs of public chat may contain some content which may not be appropriate for all audiences. Use at your own risk.
Logs from 2006-2009 pulled from wrtlprnft
Format changes at: 2015-08-25, 2017-02-20, and 2020-03-23. Times (2015 and later) should be Eastern.
. [About this site] [Credits]