Show enters and exits. Hide enters and exits.
| 00:09:49 | boyscout | Add ability to load options from a file - 878ca1b - Evan Phoenix |
| 00:09:50 | boyscout | Config files can load other config files - 97de677 - Evan Phoenix |
| 00:12:08 | evan | ok, enough config file stuff :D |
| 00:23:55 | rue | Tee hee, I have a new favourite term: Java is an enormity. |
| 00:24:32 | evan | rad |
| 00:24:39 | evan | cscope -R vm |
| 00:24:41 | evan | segfaults. |
| 00:25:25 | rue | Trying to prune stuff? |
| 00:25:39 | evan | a little |
| 00:25:49 | evan | trying to clean up CallFrame a little |
| 00:26:20 | evan | so the JIT doesn't have to so much work to build one that the VM understands |
| 00:31:45 | rue | That should help a bit. Have you looked into how well LLVM groks ByteArray and Tuple? |
| 00:32:59 | evan | it groks them fine |
| 00:33:08 | evan | it's easy to explain to them how to fetch raw parts |
| 00:33:13 | evan | be it pointers or bytes |
| 00:34:02 | rue | Being able to "inline" that access should be helpful |
| 00:34:12 | evan | already doing that for Tuple#at |
| 00:34:13 | evan | yeah |
| 00:34:28 | evan | i'll hand inline the other 3 too |
| 00:34:30 | evan | they're easy |
| 01:14:39 | slava | hi evan |
| 01:14:50 | evan | allo slava |
| 01:16:19 | headius | afternoon |
| 01:16:24 | evan | headius: hey there |
| 01:16:27 | evan | we're off by one week |
| 01:16:30 | evan | i'm in SF next week |
| 01:16:47 | headius | ahh I was wondering about that |
| 01:16:47 | headius | oh well |
| 01:24:33 | evan | my parents are in town this week |
| 01:24:35 | evan | thats why i'm not up there |
| 02:03:03 | boyscout | Refactor CallFrame::name - 84d2003 - Evan Phoenix |
| 02:03:03 | boyscout | Move CallFrame::name to CallFrame::msg - 2f7908a - Evan Phoenix |
| 02:27:59 | headius | evan: ahhh I see |
| 02:28:07 | evan | glances about |
| 02:28:07 | headius | well, you can hang out with laurent next week :D |
| 02:28:15 | evan | ah! |
| 02:28:16 | evan | hehe |
| 02:28:25 | evan | it was partly because of WWDC |
| 02:28:30 | evan | Marcus is in town again |
| 02:28:41 | headius | ahhh |
| 02:30:13 | evan | I forgot JavaOne and friends was this week |
| 02:30:27 | evan | for I am a bad nerd. |
| 02:31:29 | headius | yeah, crazytown |
| 02:32:50 | evan | i'm just happy I can stay in the EY apt |
| 02:32:57 | evan | i'm sure there are no hotels |
| 02:44:58 | headius | evan: yeah, probably would be tougher to get a hotel at least |
| 02:45:36 | evan | headius: yeah |
| 02:45:48 | evan | so, because our paths are intertwinned |
| 02:45:54 | evan | i'm trying to slim down my call frame |
| 02:45:55 | evan | :D |
| 02:47:48 | headius | you probably have somewhat less cost than we do because of alloca, but it can't hurt |
| 02:48:02 | evan | yeah |
| 02:48:10 | evan | it's not so much the allocation i'm working on |
| 02:48:15 | evan | but the ammount of data in it. |
| 02:48:46 | headius | yeah, in 1.4 I'm hoping that --fast will actually only initialize what's needed |
| 02:48:53 | headius | rather than an all-or-nothing thing |
| 02:49:04 | evan | yeah, for instance |
| 02:49:16 | evan | i'm trying to get my StaticScope object in the call frame lazily used |
| 02:49:22 | evan | because it's only there so that instance_eval works right. |
| 02:49:31 | headius | ah, yes |
| 02:52:44 | headius | evan: yeah, we keep trying to reduce frame size, and probably will also fix stack size at 1000 or something soon |
| 02:53:08 | evan | really? |
| 02:53:09 | headius | I keep feeling like there's some magic way we could pass this data across calls that doesn't involve heap frames |
| 02:53:11 | evan | python style? |
| 02:53:38 | headius | well, it would be fixed but still possibly to go over it with a perf penalty |
| 02:53:39 | evan | going to just preallocate all the frames? |
| 02:53:43 | headius | yes |
| 02:53:45 | headius | we already do that |
| 02:55:00 | evan | interesting |
| 04:36:04 | slava | evan: how's the JIT going? |
| 21:16:28 | evan | slava: poke |
| 22:55:18 | lypanov | wibbles |
| 22:55:41 | evan | yo yo |
| 22:55:51 | lypanov | yoyo evan. how goes? |
| 22:55:57 | evan | pretty good |
| 22:55:58 | evan | you? |
| 22:56:38 | lypanov | completely awesome |
| 22:56:45 | evan | good to hear! |
| 22:56:52 | lypanov | how goes to llvm? |
| 22:56:56 | lypanov | teh* |
| 22:57:05 | evan | good good |
| 22:57:13 | evan | working great now that i've got the architecture setup properly |
| 22:57:20 | evan | and LLVM complication is shunted to a background thread |
| 22:57:27 | evan | it's never decreasing performance |
| 22:57:33 | evan | only increas. |
| 22:58:06 | lypanov | awesome! |
| 22:58:38 | lypanov | getting any good speed ups? |
| 22:59:11 | evan | yeah |
| 22:59:15 | evan | small wins right now |
| 22:59:22 | evan | about 15% improvement on the spec run |
| 22:59:27 | evan | Array#each is 100% faster |
| 22:59:41 | evan | because i've taught the JIT a bit about how to make it fast |
| 22:59:49 | lypanov | maybe time for me to fuck around with it for a bit |
| 22:59:52 | evan | ie, it knows how to perform Tuple#at directly |
| 23:00:03 | evan | no method or function call |
| 23:00:11 | lypanov | is getting married in 2 weeks, will have time free to code |
| 23:00:12 | evan | a guard and then a memory move |
| 23:00:19 | evan | hey! congrats! |
| 23:00:26 | lypanov | :D thx |
| 23:00:54 | evan | big wedding? small? |
| 23:00:58 | lypanov | new iphone, time to hack, and a wife. couldn't get better |
| 23:01:19 | lypanov | (i'm assuming the iphone will be released anyway :P ) |
| 23:03:14 | lypanov | evan: not sure really. 70. no idea if thats big or small |
| 23:03:23 | evan | small. |
| 23:03:24 | lypanov | guess small :) |
| 23:03:36 | lypanov | would prefer 20 :P |
| 23:03:49 | evan | hehe |
| 23:03:49 | boyscout | Lazily use CallFrame::static_scope - 09beb70 - Evan Phoenix |
| 23:03:49 | boyscout | Make usage of CallFrame::top_scope conditional - 931fb04 - Evan Phoenix |
| 23:03:49 | boyscout | Remove CallFrame::stack_size - 584c8cf - Evan Phoenix |
| 23:03:49 | boyscout | Remove CallFrame::args - 5f346f2 - Evan Phoenix |
| 23:03:56 | evan | wtf. |
| 23:04:03 | evan | i thought that push failed. |
| 23:04:06 | evan | bizarre. |
| 23:27:00 | slava | hie van |
| 23:27:02 | slava | evan* |
| 23:29:06 | evan | allo |
| 23:29:16 | evan | i was curious how you profile things |
| 23:29:28 | slava | I have a counting profiler |
| 23:31:44 | lypanov | not the special profiling instructions? |
| 23:31:53 | lypanov | not worth the effort? |
| 23:46:28 | rue | evan: Apropos "compilation" misspelling :P |
| 23:49:53 | evan | eh? |
| 23:50:37 | rue | "LLVM complication" |
| 23:51:05 | evan | hah |
| 23:51:49 | jackdempsey | hahaha |
| 23:51:55 | jackdempsey | great typo |
| 23:54:42 | rue | Keep thinking there should be a Nil::to<>() or something |