Conversation with #inferno at Tue Oct 15 00:34:37 2013 on powerman@irc.freenode.net (irc)
(00:34:37) #inferno: Topic for #inferno set by mennis at 17:30:16 on 12/28/09
(00:42:32) kabbi left the room (quit: Remote host closed the connection).
(00:44:27) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(00:45:39) trsohmers: Greetings
(00:51:33) leetspete1: Greetings!
(00:52:00) raphaelsc: Greetings =)
(00:52:03) raphaelsc: from Brazil
(00:56:21) leetspete1: trsohmers, raphaelsc:  Now we're all properly greeted. ☺
(00:56:53) trsohmers: Hey guys, just discovered Inferno today, and wanted to see if the community was still alive
(00:58:05) raphaelsc: Inferno will always be alive in our respective hearts =)
(00:58:58) trsohmers: I'm really fascinated by the concept of Plan 9, and interested in the lean-ness of Inferno
(01:01:18) trsohmers: raphaelsc: how long have you been using Inferno?
(01:01:50) trsohmers: and how would you categorize your interest in it?
(01:02:21) raphaelsc: I started using it about 2 years ago with leetspete1's help.
(01:02:39) leetspete1: trsohmers: There's recent-ish growth in the community, although it's still somewhat small.
(01:02:46) raphaelsc: But it hasn't been my default system.
(01:03:58) trsohmers: I'm interested in it's potential for high performance cluster computing... specifically a new processor architecture that I am working with
(01:04:11) leetspete1: That's pretty cool!
(01:07:50) trsohmers: I'm a hardware guy though, so I'm trying to connect my low level hardware knowledge to some of the OS concepts
(01:08:34) leetspete1: Are you able to talk about the chip?
(01:08:44) trsohmers: Yea
(01:09:03) trsohmers: Well, I can talk about one revision of it
(01:09:44) trsohmers: The version that I am actually using for a system is not yet announced
(01:10:32) leetspete1: Cool.  What's the name of the one you are allowed to talk about?  ☺
(01:10:33) trsohmers: but it (*may*) be possible to prototype the OS or at least some of the functionality on this early silicon
(01:10:42) trsohmers: It is the Adapteva Epiphany III
(01:10:48) trsohmers: http://www.adapteva.com/
(01:11:42) leetspete1: Yeah.  The caveats on building it are that for new arches, you'll need to know some specifics about how the C compiler lays things out on the stack, but it's fairly simple otherwise, I believe.
(01:11:50) trsohmers: The Parallella (the development board that showcases the Epiphany III) was kickstart'd last year, and will be shipping to backers next month... I've had one for the past 2 months and doing basic work with it
(01:12:26) trsohmers: Well virtually everything about the board, the architecture, and the eSDK (Epiphany SDK) is pretty open
(01:12:37) leetspete1: Except the JIT, the JIT requires some additional hacking. ☺
(01:13:10) leetspete1: Oh, this board looks awesome.
(01:14:09) trsohmers: It's pretty fun
(01:17:57) trsohmers: The datasheet for the chip, plus the architecture and eSDK reference manuals are easy to find
(01:18:03) leetspete1: Looks pretty fascinating!  I'm in crunch time for some client work, or I'd probably be playing with an emulator already.
(01:19:15) leetspete1: Damned "need to pay bills" always messing with me. ;)
(01:19:30) trsohmers: Yea, I know that feeling
(01:20:24) trsohmers: if you have some free time in the future, I would love to talk to you about it
(01:20:57) trsohmers: One thing about the Epiphany emulator: It only emulates one core, while one of the key things of the architecture is core communication and shared memory
(01:25:02) leetspete1: What's the ISC like?  Poking through the website, a few pages into the arch reference, looks like a GPU-ish architecture?
(01:25:50) trsohmers: GPU-ish, but has a lot of nifty features
(01:25:50) trsohmers: Each core has its own L1 Cache, but any core can access any other cores memory
(01:26:10) trsohmers: unlike a GPU, in which all the cores must be working on the same process at the same time, the Epiphany can have different cores working on different processes
(01:28:42) leetspete1: Seriously cool stuff.
(01:33:52) trsohmers: yea
(01:34:04) trsohmers: Are you typically pretty busy?
(01:39:03) leetspete1: Well, I've only been doing this company since the end of last year, so there are spells of "incredibly busy" and "incredibly bored", haha.  No steady clients yet, except for one.
(01:41:27) trsohmers: awesome, I'm working on this cluster computing stuff as a startup as well
(01:50:53) leetspete1: Yeah, unfortunately, I don't get enough time to play with cluster stuff for clients (Los Angeles...the clients all want "social web apps" :/ ).  But we're boot-strapping some cluster stuff with the client money (using Inferno, of course).
(01:51:31) leetspete1: The Epiphany looks a lot more usable than most GPU computing.  It's pretty fascinating.
(01:52:18) trsohmers: That's why I started using it
(01:52:38) trsohmers: Yea, I know the web app crap... I just moved to the SF Bay Area from Boston
(01:53:29) trsohmers: Luckily I have enough money to try to push through working on this hardware. Kinda sad that there isn't more real hardware or low level software development happening in "Silicon" Valley
(01:59:04) leetspete1: It's definitely going to have to happen at some point, the same as getting serious about cluster computing.  It's not too heavy yet, but we're due for an architectural overhaul if the lights dim when someone turns on their desktop. ☺
(01:59:36) raphaelsc: Does Inferno has compressed in-memory swap such as zswap?
(02:02:10) trsohmers: That's why I'm focusing on energy efficiency. I was originally looking at the server/web hosting markets, but surprisingly they are less friendly to newcomers than the government and HPC giants
(02:09:28) leetspete1: I don't believe it handles swap at all, although you're unlikely to need it (the garbage collector is very good).  My current system on a 1080p display desktop system with full GUI and everything running (the code I'm working on, IRC client, etc.) rarely passes 10MB.
(02:09:59) leetspete1: Memory's pretty well-segmented; swap ought to be simple to implement.
(03:58:45) mortdeus left the room (quit: Ping timeout: 245 seconds).
(04:43:49) anth_r [none@root.9srv.net] entered the room.
(05:47:14) trsohmers left the room (quit: Remote host closed the connection).
(05:47:41) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(05:52:41) trsohmers left the room (quit: Ping timeout: 272 seconds).
(06:48:56) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(08:33:58) trsohmers left the room (quit: Remote host closed the connection).
(08:34:24) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(08:38:38) trsohmers left the room (quit: Ping timeout: 240 seconds).
(08:42:57) Gegemon [~ynv@mx1.airis.ru] entered the room.
(08:47:01) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(08:51:00) trsohmers left the room (quit: Remote host closed the connection).
(08:51:27) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(08:55:48) trsohmers left the room (quit: Ping timeout: 240 seconds).
(09:45:12) Gegemon left the room (quit: Remote host closed the connection).
(09:53:29) qrstuv: is there a good reason the registry can't have an entry with two values sharing the same key, or is it just because of lazyness?
(09:56:30) qrstuv: i'd like to have entries like tcp!10.3.2.1!12345 resource http host example.com host www.example.com host www.www.example.com
(10:46:15) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(10:46:49) trsohmers left the room (quit: Remote host closed the connection).
(10:47:14) trsohmers [~trsohmers@2601:9:4180:23b:60b6:9439:bbd3:f0cc] entered the room.
(10:47:52) trsohmer_ [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(10:48:23) trsohmer_ left the room (quit: Remote host closed the connection).
(10:48:49) trsohmer_ [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(10:49:23) trsohme__ [~trsohmers@2601:9:4180:23b:a502:4ed7:267b:213e] entered the room.
(10:49:23) trsohmer_ left the room (quit: Read error: Connection reset by peer).
(10:50:28) trsohme__ left the room (quit: Remote host closed the connection).
(10:50:55) trsohmer_ [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(10:51:26) trsohmers left the room (quit: Ping timeout: 240 seconds).
(10:51:27) trsohmer_ left the room (quit: Read error: Connection reset by peer).
(10:51:30) trsohme__ [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(10:52:07) trsohme__ left the room (quit: Remote host closed the connection).
(10:52:34) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(10:53:26) trsohmer_ [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(10:53:27) trsohmers left the room (quit: Read error: Connection reset by peer).
(10:54:01) trsohmer_ left the room (quit: Remote host closed the connection).
(10:54:31) trsohmers [~trsohmers@2601:9:4180:23b:a50c:21d7:a501:c12f] entered the room.
(10:58:38) trsohmers left the room (quit: Ping timeout: 240 seconds).
(11:18:50) powerman: qrstuv: not sure what you mean as "good reason", but I suppose it's just to simplify management of attributes
(11:19:32) qrstuv: i mean would i make life hard for myself if i tried to change it
(11:20:40) powerman: yes, you do. registries(2) implement Attributes adt to work on string pairs, and it used a lot of many apps
(11:21:17) qrstuv: yes i do what?
(11:21:44) qrstuv: obviously Registires->Attributes would need to change a little, but that's not a big deal
(11:22:02) powerman: if you need to store several values in single attribute you should just pack these values in one string using any format you like - from string(2) quoted()/unquoted() to JSON
(11:22:19) powerman: "you do" make life hard for yourself :)
(11:22:22) qrstuv: yeah, i guess that would work
(11:23:16) qrstuv: it is the laziest solution, which to me makes it the best solution
(11:23:55) baux [~jircii@out-pix.zucchetti.com] entered the room.
(11:46:53) lamer13818267989 [~jircii@out-pix.zucchetti.com] entered the room.
(11:50:25) baux left the room (quit: Ping timeout: 268 seconds).
(12:38:35) jbrhee left the room (quit: Ping timeout: 272 seconds).
(12:38:35) leetspete1 left the room (quit: Ping timeout: 272 seconds).
(12:38:35) raphaelsc left the room (quit: Ping timeout: 272 seconds).
(12:40:28) raphaelsc [~rsc@177-177-196-122.user.veloxzone.com.br] entered the room.
(12:40:46) jbrhee [~jess@76.91.33.110] entered the room.
(13:54:58) trsohmers [~trsohmers@c-98-234-55-146.hsd1.ca.comcast.net] entered the room.
(13:59:10) trsohmers left the room (quit: Ping timeout: 245 seconds).
(14:45:17) msingle [~Thunderbi@wsip-174-77-94-2.hr.hr.cox.net] entered the room.
(18:13:41) raphaelsc left the room (quit: Ping timeout: 272 seconds).