00:39:07 tcr [~tcr@196.214.183.68] has joined #sbcl 00:43:30 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 252 seconds] 01:31:35 LiamH [~none@96.231.227.13] has joined #sbcl 01:32:20 <|3b|> i'm getting 5-8 second compiles for ~100 line functions calling lots of cl-opengl functions, profiles show ~60-70% of time spent in SB-C::INSERT-BLOCK-GLOBAL-CONFLICT 01:32:51 <|3b|> same function with most of the cl-opengl calls replaced with normal functions compiles in ~0.5 sec 01:35:33 <|3b|> the gl functions are inlined, and just call another function, also inlined, which calls a foreign function and another inlined function which does another foreign call and a bit of error checking 01:36:02 <|3b|> (which now that i think about it might be inlining a bit much) 01:36:18 i think it's inlining + foreign call. 01:37:59 <|3b|> hmm, that took a while to disassemble 01:38:37 *|3b|* should possibly make the error check not inline 01:39:19 <|3b|> 27k lines of disassembly 01:39:52 tcr [~tcr@196.214.183.68] has joined #sbcl 01:40:55 <|3b|> still ~12k for the version that compiles in 0.5 sec though 01:42:22 <|3b|> yeah, 0.8 sec for full thing with error checks not inline, wonder how much effect it has on performance 01:42:23 |3b|: probably fewer global (i.e. > basic block) liveness conflicts though 01:44:22 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 252 seconds] 01:44:33 but conflict analysis doesn't really scale. 01:44:53 *|3b|* supposes not inline error checks gets rid of all the annoying warnings from failing to optimize restart-case as well 01:46:10 yena [~yena@cpe-72-177-30-155.austin.res.rr.com] has joined #sbcl 01:47:13 oh, restart-case would be a good chunk of complex code. 01:49:42 <|3b|> possibly not affecting speed much at least 02:12:26 <|3b|> yeah, looks like it was the restart-case, with just that part not inlined, compiles reasonably fast 02:13:51 *|3b|* 's benchmark is too noisy to distinguish the 3 cases though, so probably will just leave the whole error test not inlined 02:17:02 *|3b|* probably loses more performance to cffi anyway 02:40:24 tcr [~tcr@196.214.183.68] has joined #sbcl 02:44:51 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 252 seconds] 03:41:38 tcr [~tcr@196.214.183.68] has joined #sbcl 03:46:05 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 252 seconds] 04:42:21 tcr [~tcr@196.214.183.68] has joined #sbcl 04:46:54 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 264 seconds] 05:24:54 -!- LiamH [~none@96.231.227.13] has quit [Ping timeout: 240 seconds] 05:30:11 -!- Vivitron [~Vivitron@pool-98-110-213-33.bstnma.fios.verizon.net] has quit [Remote host closed the connection] 05:43:12 tcr [~tcr@196.214.183.68] has joined #sbcl 05:47:26 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 252 seconds] 06:00:12 tcr [~tcr@196.214.183.68] has joined #sbcl 06:11:02 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 265 seconds] 06:22:53 stassats [~stassats@wikipedia/stassats] has joined #sbcl 06:29:49 -!- edgar-rft [~GOD@HSI-KBW-091-089-000-047.hsi2.kabelbw.de] has quit [Quit: mental blackout] 06:36:33 tcr [~tcr@196.214.183.68] has joined #sbcl 06:44:16 -!- tcr [~tcr@196.214.183.68] has quit [Ping timeout: 252 seconds] 06:50:28 tcr [~tcr@41.115.173.189] has joined #sbcl 07:20:05 sdemarre [~serge@91.176.85.107] has joined #sbcl 07:47:42 -!- ASau [~user@46.115.73.240] has quit [Ping timeout: 265 seconds] 07:54:29 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Quit: Leaving.] 08:14:23 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 08:23:01 attila_lendvai [~attila_le@92.47.248.232] has joined #sbcl 08:23:01 -!- attila_lendvai [~attila_le@92.47.248.232] has quit [Changing host] 08:23:01 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 08:28:04 prxq [~mommer@mnhm-5f75e3f4.pool.mediaWays.net] has joined #sbcl 09:17:44 -!- tcr [~tcr@41.115.173.189] has quit [Quit: Leaving.] 10:31:17 -!- yena [~yena@cpe-72-177-30-155.austin.res.rr.com] has quit [Quit: yena] 10:47:22 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 268 seconds] 11:01:45 echo-area [~user@182.92.247.2] has joined #sbcl 11:08:58 -!- sdemarre [~serge@91.176.85.107] has quit [Ping timeout: 246 seconds] 11:24:49 sdemarre [~serge@91.176.68.235] has joined #sbcl 11:29:49 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 11:41:26 -!- cmm- [~cmm@109.65.202.199] has quit [Remote host closed the connection] 12:15:52 -!- minion [~minion@tiger.common-lisp.net] has quit [Remote host closed the connection] 12:15:52 minion [~minion@tiger.common-lisp.net] has joined #sbcl 12:17:13 hlavaty [~user@91-65-217-229-dynip.superkabel.de] has joined #sbcl 12:17:55 -!- sdemarre [~serge@91.176.68.235] has quit [Ping timeout: 246 seconds] 12:34:02 stassats [~stassats@wikipedia/stassats] has joined #sbcl 12:50:17 sdemarre [~serge@91.176.68.235] has joined #sbcl 12:55:10 yena [~yena@ip-64-134-27-116.public.wayport.net] has joined #sbcl 13:09:54 -!- sdemarre [~serge@91.176.68.235] has quit [Ping timeout: 240 seconds] 13:25:01 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 13:32:35 Blinda [~antani@host45-237-dynamic.6-87-r.retail.telecomitalia.it] has joined #sbcl 13:32:44 -!- Blinda [~antani@host45-237-dynamic.6-87-r.retail.telecomitalia.it] has quit [Client Quit] 13:54:59 -!- yena [~yena@ip-64-134-27-116.public.wayport.net] has quit [Quit: yena] 14:14:10 LiamH [~none@pdp8.nrl.navy.mil] has joined #sbcl 14:44:30 psilord [~psilord@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 15:12:46 milanj [~milanj_@79-101-250-98.dynamic.isp.telekom.rs] has joined #sbcl 15:18:38 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Quit: No Ping reply in 180 seconds.] 15:21:00 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 15:32:23 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Quit: No Ping reply in 180 seconds.] 15:32:44 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 15:35:12 sdemarre [~serge@91.176.68.235] has joined #sbcl 15:36:07 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Client Quit] 15:36:30 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 15:44:23 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Quit: No Ping reply in 180 seconds.] 15:44:44 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 15:47:20 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Remote host closed the connection] 15:47:32 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 15:50:55 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Client Quit] 15:51:17 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 15:54:41 -!- fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has quit [Client Quit] 16:10:41 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 16:19:54 fe[nl]ix [~quassel@pdpc/supporter/professional/fenlix] has joined #sbcl 16:35:08 -!- specbot [~specbot@tiger.common-lisp.net] has quit [Disconnected by services] 16:35:12 specbot [~specbot@tiger.common-lisp.net] has joined #sbcl 16:47:04 -!- sdemarre [~serge@91.176.68.235] has quit [Ping timeout: 246 seconds] 17:06:59 -!- hroptatyr [~asathor@sxemacs/devel/hroptatyr] has quit [Remote host closed the connection] 17:25:55 -!- scymtym [~user@2001:638:504:2093:226:b9ff:fe7d:3e1f] has quit [Remote host closed the connection] 17:25:59 pnpuff [~aeiou@unaffiliated/pnpuff] has joined #sbcl 17:32:38 scymtym [~user@2001:638:504:2093:226:b9ff:fe7d:3e1f] has joined #sbcl 17:45:58 Krystof: well, I had some stuff that only works on the standard combination, for instance. 17:46:22 not that I know what cl-protobuf they're referring to in any case. 17:49:12 milosn [~milosn@5ac0f955.bb.sky.com] has joined #sbcl 17:50:23 your stuff is probably more clever than my stuff 17:52:04 my 0th take was to make a final dfun for all the method specializers; the 1st take was to make one for all the specializers and their subclasses (cross-producted) 17:52:19 the 2nd one was "you might have a class hierarchy that makes this prohibitive -- let me know" 17:54:37 I think everything converges to something like your 2nd. 17:55:38 yeah 17:55:55 but then again, we already do something like that, iirc. 17:56:17 ASau [~user@46.115.56.237] has joined #sbcl 17:56:24 well, we incrementally get there, but I can see why one would want to shortcut that incrementalism 17:56:54 right 17:57:04 anyway, must brave public transport 17:57:11 keep talking, I'll reply post bedtime 17:59:34 I don't know... Maybe we can have a mixed perfect hash/graph walk that works well on huge GFs, and something like the final dfun otherwise (which might even be inlined). 18:00:05 this situation with hundreds of methods reminds me of kpreid's stuff. 18:00:27 -!- pnpuff [~aeiou@unaffiliated/pnpuff] has left #sbcl 18:01:26 The problem is that, in theory, given hundreds of methods that specialise on a set of classes C that are all disjoint, future classes could inherit from any subset of C, in any order. 18:02:19 we will always need to have an adaptive method in there 18:02:48 right, but it seems like people who worry about the cache tend to have this really bad corner case 18:02:55 and I don't currently see anyy "good" way to do it. 18:03:53 but I declare it to be Their Problem if they simultaneously want instantly good (i.e. precached) performance and also to extend their class hieararchy 18:03:59 oh well, specialise on what exists, and punt to the adaptive way. 18:04:04 good. 18:04:12 I think we can provide good precached and correct adaptive 18:04:18 Same. 18:04:21 fine :) 18:04:22 CL-USER> (colour/ (colour+ "#123456" "#789abc") 2) 18:04:22 # R: 0.34 G: 0.46 B: 0.58 (#577594)> 18:04:41 meanwhile, 21-bit floats in "actually used in a calculation" shock 18:04:57 scary (: 18:05:58 I'm not going to admit how many tries it took me to get the logical block stuff right for the indentation 18:06:13 well, if we do that, it looks like a plain precompilation of effective methods and a hash table. 18:09:50 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 260 seconds] 18:23:17 attila_lendvai [~attila_le@92.47.248.255] has joined #sbcl 18:23:17 -!- attila_lendvai [~attila_le@92.47.248.255] has quit [Changing host] 18:23:17 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 18:58:23 homie [~homie@94.123.220.37] has joined #sbcl 19:00:29 Blkt [~user@62.10.10.99] has joined #sbcl 19:29:27 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 260 seconds] 19:32:14 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Read error: Operation timed out] 19:32:34 -!- slyrus [~chatzilla@adsl-99-190-98-53.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 240 seconds] 19:50:34 cmm [~cmm@bzq-79-181-240-254.red.bezeqint.net] has joined #sbcl 20:00:13 pnpuff [~aeiou@unaffiliated/pnpuff] has joined #sbcl 20:16:02 -!- pnpuff [~aeiou@unaffiliated/pnpuff] has quit [Quit: Bye.] 20:47:44 slyrus [~chatzilla@173.228.44.92] has joined #sbcl 21:10:30 tsuru` [~charlie@adsl-74-179-25-48.bna.bellsouth.net] has joined #sbcl 21:16:01 sdemarre [~serge@91.176.68.235] has joined #sbcl 22:00:54 -!- sdemarre [~serge@91.176.68.235] has quit [Ping timeout: 240 seconds] 22:12:33 -!- milosn [~milosn@5ac0f955.bb.sky.com] has quit [Read error: Connection reset by peer] 22:13:20 milosn [~milosn@90.192.249.85] has joined #sbcl 22:16:41 -!- psilord [~psilord@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Quit: Leaving.] 22:17:19 Thra11 [~thrall@87.113.150.33] has joined #sbcl 22:17:27 Bike [~Glossina@67-5-240-27.ptld.qwest.net] has joined #sbcl 22:34:09 xxhx [~user@host115.190-226-93.telecom.net.ar] has joined #sbcl 22:35:31 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 23:05:46 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 23:16:05 psilord [~psilord@69.180.173.249] has joined #sbcl