| --- Log opened Mon May 14 00:00:37 2012 |
| 00:01 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 00:32 | | You're now known as TheWatcher[T-2] |
| 00:35 | | You're now known as TheWatcher[zZzZ] |
| 00:39 | | * McMartin files many issues against himself on Github. |
| 00:39 | | * McMartin clearly has issues. |
| 00:58 | | jeroid [jerith@EB02D1.779B91.7C1BA3.E03A48] has joined #code |
| 01:06 | | Derakon is now known as Derakon[AFK] |
| 01:08 | | himi [fow035@D741F1.243F35.CADC30.81D435] has joined #code |
| 01:08 | | mode/#code [+o himi] by ChanServ |
| 01:18 | | Noah [maoranma@Nightstar-ec7f96f9.pools.spcsdns.net] has joined #code |
| 01:24 | | cpux|2 is now known as cpux |
| 01:24 | | jeroid [jerith@EB02D1.779B91.7C1BA3.E03A48] has quit [Ping timeout: 121 seconds] |
| 01:25 | | jeroid [jerith@EB02D1.779B91.7C1BA3.E03A48] has joined #code |
| 01:26 | | jeroid [jerith@EB02D1.779B91.7C1BA3.E03A48] has quit [[NS] Quit: Bye] |
| 01:26 | | jeroid [jerith@Nightstar-ceb4eaff.tmodns.net] has joined #code |
| 01:44 | | jeroid [jerith@Nightstar-ceb4eaff.tmodns.net] has quit [[NS] Quit: Bye] |
| 02:01 | | maoranma [maoranma@490720.C448F4.E7801D.4B8709] has joined #code |
| 02:02 | | Noah [maoranma@Nightstar-ec7f96f9.pools.spcsdns.net] has quit [NickServ (GHOST command used by maoranma)] |
| 02:02 | | maoranma is now known as Noah |
| 02:10 | | Attilla [Obsolete@Nightstar-02604040.as43234.net] has quit [Ping timeout: 121 seconds] |
| 02:12 | | Kindamoody[zZz] is now known as Kindamoody |
| 02:51 | | Kindamoody is now known as Kindamoody|afk |
| 04:35 | | Kindamoody|afk is now known as Kindamoody |
| 04:47 | < Rhamphoryncus> | rms: I meant an img tag and all the other places html has images |
| 04:58 | | himi [fow035@D741F1.243F35.CADC30.81D435] has quit [Ping timeout: 121 seconds] |
| 05:18 | | Derakon[AFK] is now known as Derakon |
| 05:33 | | * Derakon eyes this path: Source/Core/Core/Src/IPC_HLE/WII_IPC_HLE_Device_usb.cpp |
| 05:33 | <&Derakon> | (As seen on the TASVideos forums, discussing the Dolphin Wii/Gamecube emulator) |
| 05:34 | < Noah> | Shoot the core! |
| 05:34 | < Noah> | No, not that core, the other one, near the src |
| 05:44 | | Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has quit [Client exited] |
| 05:54 | | ErikMesoy|sleep is now known as ErikMesoy |
| 06:29 | | Derakon is now known as Derakon[AFK] |
| 07:02 | <&McMartin> | Awesome. Light Table reached its goal! |
| 07:14 | <&McMartin> | Also, it looks like I should really learn Clojure. |
| 07:18 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 07:18 | | mode/#code [+o himi] by ChanServ |
| 07:41 | | eckse [eckse@Nightstar-3b021371.dsl.sentex.ca] has quit [Client closed the connection] |
| 07:55 | <&McMartin> | http://clojure.org/jvm_hosted |
| 07:56 | <&McMartin> | This is nuts |
| 08:24 | < froztbyte> | haha |
| 08:24 | < froztbyte> | welcome to the world of not-perl, where things are crazyawesome |
| 08:25 | <&McMartin> | Well |
| 08:25 | <&McMartin> | It's still Swing |
| 08:25 | <&McMartin> | So there is a certain level of D: |
| 08:26 | < froztbyte> | sure |
| 08:26 | < froztbyte> | but I meant that more generally |
| 08:27 | < froztbyte> | there's all sorts of shiny coming up in newer languages and such |
| 08:27 | < froztbyte> | like have you seen Julia? |
| 08:27 | <&McMartin> | I do wonder if Clojure would have the same problem I had with Scala though |
| 08:27 | <&McMartin> | In that "Just In Time" was more "Maybe After My Nap" |
| 08:28 | < froztbyte> | http://julialang.org/ |
| 08:28 | < froztbyte> | McMartin: how do you mean? |
| 08:28 | <&McMartin> | I would get things like five minute process spinup times for things like "Hello World" |
| 08:28 | <&McMartin> | Woo, LLVM |
| 08:30 | <&McMartin> | Looks like it's competing with R and friends. |
| 08:30 | <&McMartin> | I can't help but notice how much ass JavaScript is kicking in that benchmark though o_O |
| 08:30 | < froztbyte> | oh, that |
| 08:30 | < froztbyte> | yes, that's a thing some people are working on as well |
| 08:30 | < froztbyte> | (in various places) |
| 08:31 | < froztbyte> | McMartin: yeah but it's specifically v8 |
| 08:31 | < froztbyte> | which also makes a shitton of difference |
| 08:31 | <&McMartin> | That's the one with type inference, right? |
| 08:31 | < froztbyte> | think so |
| 09:13 | | You're now known as TheWatcher |
| 09:56 | | Attilla [Obsolete@Nightstar-02604040.as43234.net] has joined #code |
| 10:27 | | cpux|2 [cpux@Nightstar-c5874a39.dyn.optonline.net] has joined #code |
| 10:30 | | cpux [cpux@Nightstar-c5874a39.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 10:57 | | Reiver [reiverta@5B433A.F67240.ED37F3.EC38A4] has joined #code |
| 11:10 | | Kindamoody is now known as Kindamoody|out |
| 11:23 | | Reiver [reiverta@5B433A.F67240.ED37F3.EC38A4] has quit [Ping timeout: 121 seconds] |
| 11:24 | | Reiver [reiverta@5B433A.F67240.ED37F3.EC38A4] has joined #code |
| 12:00 | | ErikMesoy1 [Erik_Mesoy@A08927.B4421D.B81A91.464BAB] has joined #code |
| 12:01 | | ErikMesoy [Erik_Mesoy@Nightstar-bd2f5f93.80-203-16.nextgentel.com] has quit [Ping timeout: 121 seconds] |
| 12:04 | < Tarinaky> | Does anyone know if there's a way to change chromium's memory profile to 'low' without restarting it with switches? I'd like to to reap some of its unused working set. |
| 12:20 | <@rms> | Did you mean model? |
| 12:20 | <@rms> | --memory-profile loads an accounting DLL |
| 12:20 | < Tarinaky> | I can't remember what chromium called it. |
| 12:21 | < Tarinaky> | I just want chrome to downsize some of the RAM it's allocated itself. |
| 12:21 | < Tarinaky> | Since I'm no longer looking at an embedded youtube video there's no reason to pretend you'll need the memory for it >.> |
| 12:21 | <@rms> | Yeah it's memory model |
| 12:26 | <@rms> | Can't find anything, sorry |
| 12:34 | | * Tarinaky sighs. |
| 13:59 | <@ToxicFrog> | Close the tab it was in and it might release the memory. |
| 13:59 | <@ToxicFrog> | Honestly I'd put bigger odds on this being a memory leak in the Flash plugin |
| 14:00 | <@TheWatcher> | always a safe bet, that one |
| 14:00 | <@TheWatcher> | Adobe software, how does it work? |
| 14:01 | <@Tamber> | Luck and prayers. |
| 14:19 | | Reiv [reiverta@5B433A.F67240.ED37F3.EC38A4] has joined #code |
| 14:20 | | Reiver [reiverta@5B433A.F67240.ED37F3.EC38A4] has quit [Ping timeout: 121 seconds] |
| 14:26 | < gnolam> | TheWatcher: Adobe software. Proof that you /can/ pay your programmers in bad crack. |
| 14:26 | <@TheWatcher> | No, that's oracle >.> Adobe gives them LSD, more arty~ |
| 14:27 | | * TheWatcher also stabs the DBD::mysql documentation |
| 14:33 | < froztbyte> | haha |
| 15:05 | | RichyB [MyCatVerbs@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 15:44 | | Reiv [reiverta@5B433A.F67240.ED37F3.EC38A4] has quit [Client closed the connection] |
| 16:21 | | ErikMesoy1 is now known as ErikMesoy |
| 16:23 | | ErikMesoy [Erik_Mesoy@A08927.B4421D.B81A91.464BAB] has quit [[NS] Quit: reboot] |
| 16:29 | | ErikMesoy [Erik_Mesoy@A08927.B4421D.B81A91.464BAB] has joined #code |
| 17:43 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Reboot] |
| 17:47 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has joined #code |
| 18:04 | | eckse [eckse@Nightstar-3b021371.dsl.sentex.ca] has joined #code |
| 18:04 | | mode/#code [+o eckse] by ChanServ |
| 18:08 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has quit [[NS] Quit: Reboot again] |
| 18:12 | | Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has joined #code |
| 18:29 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 18:44 | < Noah> | Are there any IRCDs that offer built in channel logging? |
| 18:44 | < Noah> | Or servicebots? |
| 18:45 | <@rms> | Not to my knowledge |
| 18:45 | <@rms> | If you want to watch for botnets, Ourmon works. |
| 18:49 | < Noah> | Nah, I'm just thinking of a logging service that breaks logs down by day, month, year and hosts them on the network. Sort of like what TheWatcher does in #mf0, but something built in to the IRCD or a services bot |
| 18:49 | < froztbyte> | Why not just use an irssi instance? |
| 18:50 | < froztbyte> | It can autosplit things using what it calls tags |
| 18:51 | < Noah> | I'm pretty sure that's what TheWatcher does |
| 18:51 | < Noah> | {05:40} * TheShadow (theshadow@Nightstar-3762b576.co.uk) has joined #mf0 |
| 18:51 | < Noah> | {05:40} * ChanServ sets mode: +o TheShadow |
| 18:51 | < Noah> | {05:41} * @TheShadow swears, wonders why irssi didn't reconnect here |
| 18:51 | < froztbyte> | I'm on tablet right now so linking isn't easy :) |
| 18:54 | < Noah> | Shoot, my failproof investigative techniques are stiffled by your irssi session |
| 18:56 | | Kindamoody|out is now known as Kindamoody |
| 19:04 | | Kindamoody is now known as Kindamoody[zZz] |
| 19:08 | < froztbyte> | Haha |
| 19:08 | < froztbyte> | Well |
| 19:09 | < froztbyte> | QuasselDroid on top of irssi |
| 19:15 | < RichyB> | Noah, you could pick an IRCd that has a sane API and write an add-on module for it. |
| 19:15 | < RichyB> | Inspircd is supposed to have a pretty good API for modules. |
| 19:16 | < RichyB> | Don't know about the rest. |
| 19:16 | <@rms> | Anope has an API system, you might be able to use that |
| 19:17 | <@rms> | Note: a server doesn't get all traffic if it has no users on that channel. |
| 19:17 | <@rms> | s/that/a/ |
| 19:24 | <@TheWatcher> | Noah: I found the reason why irssi didn't reconnect me to #mf0 is that I'd forgotten to set the autojoin >.> |
| 20:07 | | gnolam [lenin@Nightstar-202a5047.priv.bahnhof.se] has joined #code |
| 20:07 | | Vornucopia [NSwebIRC@C888DE.7F9621.4A1301.BBBE7B] has joined #code |
| 21:39 | | Vornicus is now known as Vash |
| 21:54 | | Vornucopia [NSwebIRC@C888DE.7F9621.4A1301.BBBE7B] has quit [[NS] Quit: Page closed] |
| 21:55 | | Vash is now known as Vornicus |
| 22:19 | | maoranma [maoranma@490720.C448F4.E7801D.4B8709] has joined #code |
| 22:20 | | Noah [maoranma@490720.C448F4.E7801D.4B8709] has quit [Ping timeout: 121 seconds] |
| 22:31 | | ErikMesoy is now known as ErikMesoy|sleep |
| 22:33 | | Cody [friend@EB02D1.BD813D.88B0F4.BB1851] has joined #code |
| 22:33 | < Cody> | i have one erection and no bababhiat to expelliarmus |
| 22:34 | <&McMartin> | Clojure: Not having the JIT problems I had with Scala |
| 22:34 | < RichyB> | That probably isn't on-topic. |
| 22:34 | < Cody> | after eating a grand total of 7 hot wings my anus is prince's next hit single |
| 22:34 | < RichyB> | McMartin, interesting. Seems counter-intuitive, given that Clojure is dynamically-typed. |
| 22:34 | < Cody> | this is what it sounds like when butts cry |
| 22:34 | | Cody [friend@EB02D1.BD813D.88B0F4.BB1851] has left #code ["Friendscafe.org"] |
| 22:35 | < RichyB> | OTOH java's bytecode doesn't actually have much tpe information in it except for primitive types, AIUI? Something weird going on there. |
| 22:35 | <&McMartin> | No, it's actually quite detailed |
| 22:35 | <&McMartin> | It's also possible that my JDK is better now, I guess |
| 22:36 | <&McMartin> | Alternately, maybe the Clojure REPL is more interpret-y |
| 22:36 | < RichyB> | What kind of problems were you having with Scala? |
| 22:38 | <&McMartin> | You'd type "Hello world" at the REPL prompt and it would buzz for several minutes before handing you the value - "Hello world" - of that expression |
| 22:38 | < RichyB> | *Minutes*!? |
| 22:38 | <&McMartin> | Minutes. |
| 22:38 | < RichyB> | Wow. |
| 22:38 | <&McMartin> | Not more than three, but more than one. |
| 22:38 | < RichyB> | I bet Clojure doesn't do *nearly* as much up-front analysis. |
| 22:39 | <&McMartin> | Yeah, quite. |
| 22:39 | < RichyB> | Scala's type system is AIUI a lot more complicated than Haskell's, with loads of different defaulting rules and such. |
| 22:39 | <&McMartin> | But at that level - and that none of the Scala tutorials warned about this - I must assume that something about my environment was severely fucking up compilationt ime. |
| 22:39 | < RichyB> | Maybe they're churning through slow code trying to match every possible multimethod against every subexpression? |
| 22:40 | <&McMartin> | No idae |
| 22:40 | <&McMartin> | *idea |
| 22:40 | <&McMartin> | TF has been far more successful with it than I, though |
| 22:41 | < RichyB> | What OS are you on? |
| 22:41 | <&McMartin> | Linux |
| 22:41 | <&McMartin> | I don't recall the JDK I was using at the time |
| 22:43 | <&McMartin> | The current clojure case is clojure 1.3 on OpenJDK 1.3 |
| 22:43 | <&McMartin> | Er |
| 22:43 | <&McMartin> | 1.6 |
| 22:44 | < RichyB> | Okay, so I've got the sun JVM installed here... |
| 22:44 | | * ToxicFrog upreads |
| 22:44 | < RichyB> | Sun JDK 1.6, on Ubuntu 10.04, amd64, slowish disk but gobs of RAM. |
| 22:44 | <&McMartin> | I should probably reinstall scala to make sure that apples and apples are compared. |
| 22:45 | <@ToxicFrog> | In my experience, the scala REPL takes a while to warm up (by which I mean "10-15 seconds", which is glacial compared to, say, python) but responds to commands in less than a second. |
| 22:45 | <@ToxicFrog> | Unless it's something that requires loading All Of The Jars from disk or something. |
| 22:45 | < RichyB> | I just pulled the Scala 2.9.2 tarball off the website. bin/scala, type "hello world" directly at the repl, takes on the close order of ~2000ms. |
| 22:46 | <@ToxicFrog> | That said, I think the -earliest- version I've used is Scala 2.7/JRE 1.6 |
| 22:46 | <&McMartin> | TF: Yeah, I was having excruciatingly huge delays, in excess of 60 seconds, on every single command |
| 22:47 | | * McMartin pulls down Scala 2.8.2 from Fedora's repos, as this is the most likely to play well with the stock JDK. |
| 22:48 | <@ToxicFrog> | Yeah, that sounds like either an environmental issue, or like they had some kind of terrible problem in the repl that was fixed before I tried it |
| 22:48 | | * ToxicFrog breaks out time(1) |
| 22:48 | <&McMartin> | I'm pretty sure I tried it at your urging, so the former has been my working theory for some time. |
| 22:48 | < RichyB> | time bin/scala -J-server -e 'java.lang.System.out.format("hello world\n")' # 0.86 seconds on my laptop. |
| 22:48 | <@ToxicFrog> | Hello world: 4.3 seconds, with 4.1 seconds of that being warmup time. |
| 22:48 | < RichyB> | Just to check the obvious, neither of you have a 486 by any chance? ;) |
| 22:49 | <@ToxicFrog> | Well, I'm testing this on a fairly old laptop |
| 22:49 | < RichyB> | Sorry, just being silly. |
| 22:49 | <@ToxicFrog> | I should try it on Orias |
| 22:49 | <&McMartin> | o_O |
| 22:49 | <&McMartin> | OK |
| 22:49 | <&McMartin> | scala |
| 22:49 | <&McMartin> | REPL: less than a second |
| 22:49 | <&McMartin> | "Hello world" |
| 22:49 | <&McMartin> | 15 seconds |
| 22:49 | <@ToxicFrog> | o.O |
| 22:49 | <&McMartin> | res0: java.lang.String = Hello world |
| 22:49 | <@ToxicFrog> | What if you do "hello world" a second time? |
| 22:49 | < RichyB> | McMartin, try scala -e 'java.lang.System.out.format("hello world\n")' # ? |
| 22:50 | < RichyB> | oh, try ToxicFrog's idea first. |
| 22:50 | <&McMartin> | Half second the second time. |
| 22:50 | <@ToxicFrog> | IME it loads stuff on-demand, so the first time you enter a command that requires a new library there'll be a delay |
| 22:51 | <@ToxicFrog> | That said, I'm not sure what would take 15 seconds to load for echoing a single string |
| 22:51 | | maoranma is now known as Noah |
| 22:51 | | * RichyB prods. |
| 22:51 | <&McMartin> | [mcmartin@iodine tests]$ time scala -e "java.lang.System.out.format(\"Hello, world\\n\")" |
| 22:52 | <&McMartin> | Hello, world |
| 22:52 | <&McMartin> | real 0m20.995s |
| 22:52 | <&McMartin> | user 0m1.028s |
| 22:52 | <&McMartin> | sys 0m0.166s |
| 22:52 | < RichyB> | Takes about the same time for me (once the JVM itself is in disk cache) to run hello world with OpenJDK 1.6 instead of Sun's. |
| 22:52 | <&McMartin> | Yeah |
| 22:52 | <@ToxicFrog> | McMartin: that command takes 4s for me with a cold cache and 1.2 with a wamr one. |
| 22:53 | <&McMartin> | What makes it "warm"? |
| 22:53 | <&McMartin> | Just having recently run Scala? |
| 22:53 | <@ToxicFrog> | Using openJDK 1.6 64-bit. |
| 22:53 | <@ToxicFrog> | Yeah. A warm OS block cache, I mean. |
| 22:53 | <&McMartin> | OK, we're using the same JVM here |
| 22:53 | <&McMartin> | I have no idea what the leeg |
| 22:53 | <&McMartin> | 7.5 seconds with a warm cache, it seems |
| 22:53 | <@ToxicFrog> | Actually |
| 22:54 | <@ToxicFrog> | What's java for "get version of host VM"? |
| 22:54 | <&McMartin> | java -version |
| 22:54 | <&McMartin> | one hyphen |
| 22:54 | <@ToxicFrog> | No, from Java code |
| 22:54 | <@ToxicFrog> | I have both openJRE and sunJRE installed and I want to be sure which one Scala is using |
| 22:54 | <&McMartin> | I don't have that swapped in offhand |
| 22:54 | <&McMartin> | I bet it's in java.lang.Runtime |
| 22:55 | <&McMartin> | Ah, no |
| 22:55 | <&McMartin> | It's in java.lang.System |
| 22:56 | <@ToxicFrog> | System.getProperties("java.version") looks like |
| 22:56 | <&McMartin> | java.vendor may also be useful |
| 22:56 | <@ToxicFrog> | Yep, it's using openjdk |
| 22:56 | <@ToxicFrog> | Conveniently, Properties.toString dumps everything in it |
| 22:56 | <&McMartin> | Could be some hangup inside Iodine :/ |
| 22:57 | <@ToxicFrog> | I'm using Scala 2.9 here, but I haven't noticed any performance differences relative to 2.8 |
| 22:59 | <&McMartin> | I guess for completeness |
| 23:00 | <&McMartin> | [mcmartin@iodine tests]$ java -version |
| 23:00 | <&McMartin> | java version "1.6.0_24" |
| 23:00 | <&McMartin> | OpenJDK Runtime Environment (IcedTea6 1.11.1) (fedora-65.1.11.1.fc16-x86_64) |
| 23:00 | <&McMartin> | OpenJDK 64-Bit Server VM (build 20.0-b12, mixed mode) |
| 23:01 | <@ToxicFrog> | ben@thoth ~ $ java -version |
| 23:01 | <@ToxicFrog> | java version "1.6.0_22" |
| 23:01 | <@ToxicFrog> | OpenJDK Runtime Environment (IcedTea6 1.10.6) (6b22-1.10.6-0ubuntu1) |
| 23:01 | <@ToxicFrog> | OpenJDK 64-Bit Server VM (build 20.0-b11, mixed mode) |
| 23:01 | <&McMartin> | Yeah, that really shouldn't matter unless Fedora's patches somehow kill the shit out of Scala or Ubuntu's somehow save it, both of which seem unlikely |
| 23:01 | <@ToxicFrog> | I've got nothing. |
| 23:02 | <&McMartin> | Possible culprit: Iodine only has 1GB of RAM. |
| 23:02 | <@ToxicFrog> | Especially since Clojure works, so it's clearly not that your java install is intrinsically broken or something |
| 23:02 | < RichyB> | I have 4GB of RAM here. |
| 23:02 | <@ToxicFrog> | Oh. Yeah, that would definitely do it. |
| 23:03 | <@ToxicFrog> | I fire up the REPL here and it goes to 700 virt/100 res instantly. |
| 23:04 | <&McMartin> | Mm. That's not the whole story, clearly, as clojure's java process is also grabbing more virt than I have RAM immediately |
| 23:04 | <@ToxicFrog> | Oh, huh. |
| 23:04 | <&McMartin> | But the res is only "62m" |
| 23:05 | <@ToxicFrog> | (also there is something seriously wrong with nm-applet on my computer) |
| 23:07 | < RichyB> | Virt is address space allocated. |
| 23:07 | < RichyB> | Virt is often much larger even than swap usage with processes that mmap() scadloads of things that they never touch. |
| 23:08 | < RichyB> | AIUI, someone please correct me strongly if I'm wrong. |
| 23:08 | <&McMartin> | Mmm. My Clojure times look like your scala times. |
| 23:08 | <&McMartin> | [mcmartin@iodine tests]$ time clojure -e '(.println System/out "Hello, world")' |
| 23:08 | <&McMartin> | Hello, world |
| 23:08 | <&McMartin> | real 0m1.903s |
| 23:08 | <&McMartin> | user 0m1.677s |
| 23:08 | <&McMartin> | sys 0m0.217s |
| 23:12 | < RichyB> | Oh, here's one idea. |
| 23:12 | < RichyB> | strace -f -o calls -- /usr/bin/time scala -e 'java.lang.System.out.format("hello world\n")' |
| 23:12 | < RichyB> | /usr/bin/time should give you a pagefault count in addition to the raw times. |
| 23:13 | < RichyB> | and maybe you can get a rough idea of what's happening by counting syscalls? |
| 23:13 | < RichyB> | awk '{print $2}' <jvm | sed -r 's/^([^(]+).*/\1/' | sort | uniq -c | sort -n |
| 23:14 | < RichyB> | Time tells me that scala makes 1major, 226minor pagefaults, 0 swaps. |
| 23:14 | < RichyB> | About 2800 read() calls, about 2700 lseek() calls, 500 rt_sigprocmask(), 500 stat(), 400 open()... |
| 23:14 | <&McMartin> | Do I pipe the result of strace into that awk? |
| 23:14 | < RichyB> | Yes. |
| 23:15 | < RichyB> | Er, I meant "<calls" when I wrote "<jvm" there. |
| 23:15 | < RichyB> | Sorry. I also don't know a cleaner way to carve up strace's output than that. |
| 23:15 | <&McMartin> | Nod |
| 23:15 | <&McMartin> | Also, strace missing on this system |
| 23:15 | | * McMartin will revisit later, he guesses |
| 23:15 | < RichyB> | Going through both awk and sed because sed's syntax makes me twitch for simple things like cutting a field out. |
| 23:16 | < RichyB> | At this point it's just bloody-minded determination to know everything. ;) |
| 23:21 | <&McMartin> | 0inputs+72outputs (0major+8150minor)pagefaults 0swaps |
| 23:21 | <&McMartin> | ...... |
| 23:21 | <&McMartin> | Here are my top three system calls |
| 23:21 | <&McMartin> | 2607 lseek |
| 23:21 | <&McMartin> | 2673 read |
| 23:21 | <&McMartin> | 42047 gettimeofday |
| 23:22 | <@TheWatcher> | ... that's some serious time checking going on there |
| 23:22 | < RichyB> | I wonder what's up with the gettimeofday calls. |
| 23:23 | <&McMartin> | 42047 gettimeofday, 2673 read, 2607 lseek, 1392 stat, 662 <..., 585 futex, 348 mprotect, 341 lstat, 288 rt_sigprocmask, 284 mmap, 265 open, 173 close, 150 fstat, 126 write |
| 23:24 | <&McMartin> | ... oh |
| 23:24 | <&McMartin> | The gettimeofday calls are no doubt because we're running "time" here. |
| 23:25 | < RichyB> | gettimeofday() is really fast anyway. |
| 23:25 | < RichyB> | Takes ~0.03s to do 42k of them on my CPU. |
| 23:26 | < RichyB> | http://pastebin.starforge.co.uk/518 |
| 23:27 | < RichyB> | wait, that doesn't actually make any syscalls on my box. What? |
| 23:30 | < RichyB> | Huh, gettimeofday() apparently isn't a syscall for me. |
| 23:40 | | himi [fow035@D741F1.243F35.CADC30.81D435] has joined #code |
| 23:40 | | mode/#code [+o himi] by ChanServ |
| 23:42 | | You're now known as TheWatcher[T-2] |
| 23:44 | | You're now known as TheWatcher[zZzZ] |
| 23:44 | | maoranma [maoranma@Nightstar-fc470d2b.pools.spcsdns.net] has joined #code |
| 23:46 | | Noah [maoranma@490720.C448F4.E7801D.4B8709] has quit [Ping timeout: 121 seconds] |
| 23:58 | < RichyB> | This slightly amuses me: https://consultations.external.bbc.co.uk/ZopeTime |
| 23:59 | < RichyB> | As a semi-accidental feature of one of the web applications that the company I work for runs, we have produced a remarkably inaccurate clock. |
| --- Log closed Tue May 15 00:00:12 2012 |