00:04:14 -!- tcr1 [~tcr@213.55.131.17] has quit [Quit: Leaving.] 00:38:01 stassats [~stassats@wikipedia/stassats] has joined #sbcl 00:42:56 tcr1 [~tcr@213.55.131.17] has joined #sbcl 00:44:47 -!- tcr1 [~tcr@213.55.131.17] has quit [Client Quit] 01:07:48 homie` [~levgue@xdsl-78-35-152-89.netcologne.de] has joined #sbcl 01:10:58 -!- homie [~levgue@xdsl-78-35-158-247.netcologne.de] has quit [Ping timeout: 276 seconds] 01:16:47 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 01:29:57 hargettp [~hargettp@pool-71-184-177-218.bstnma.east.verizon.net] has joined #sbcl 02:21:52 redline6561 [~user@c-66-56-55-169.hsd1.ga.comcast.net] has joined #sbcl 02:37:38 -!- redline6561 [~user@c-66-56-55-169.hsd1.ga.comcast.net] has quit [Remote host closed the connection] 02:39:18 -!- slyrus_ [~chatzilla@adsl-75-36-217-249.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 260 seconds] 02:47:52 -!- hargettp [~hargettp@pool-71-184-177-218.bstnma.east.verizon.net] has quit [Quit: Leaving...] 03:32:57 redline6561 [~user@c-66-56-55-169.hsd1.ga.comcast.net] has joined #sbcl 04:03:03 -!- nyef [~nyef@pool-70-20-57-204.man.east.myfairpoint.net] has quit [Quit: G'night all.] 05:04:19 -!- Sake [~Sakeeee@e179061045.adsl.alicedsl.de] has quit [Ping timeout: 240 seconds] 05:04:33 slyrus_ [~chatzilla@adsl-75-36-217-249.dsl.pltn13.sbcglobal.net] has joined #sbcl 14:12:11 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 14:12:11 -!- names: ccl-logbot hargettp homie` stassats angavrilov mega1 jsnell derrotebaron cmm christoph_debian flip214 slyrus_ redline6561 lnostdal ASau pkhuong minion specbot lisppaste2 fe[nl]ix @Xof tsuru derrida deepfire foom luis slyrus froydnj jiacobucci1 joshe gnooth |3b|` 14:20:27 nyef [~nyef@pool-70-20-57-204.man.east.myfairpoint.net] has joined #sbcl 14:34:45 -!- ASau [~user@95-27-211-235.broadband.corbina.ru] has quit [Remote host closed the connection] 14:36:25 -!- hargettp [~hargettp@dhcp-162.mirrorimage.net] has quit [Quit: hargettp] 14:37:02 hargettp [~hargettp@dhcp-162.mirrorimage.net] has joined #sbcl 14:45:34 -!- flip214 [~marek@2001:858:107:1:baac:6fff:fe6b:9183] has quit [Remote host closed the connection] 15:19:31 -!- hargettp [~hargettp@dhcp-162.mirrorimage.net] has quit [Remote host closed the connection] 15:41:12 hargettp [~hargettp@dhcp-162.mirrorimage.net] has joined #sbcl 15:46:05 attila_lendvai [~attila_le@catv-80-98-24-21.catv.broadband.hu] has joined #sbcl 16:39:34 -!- redline6561 [~user@c-66-56-55-169.hsd1.ga.comcast.net] has quit [Ping timeout: 255 seconds] 17:19:26 Krystof [~csr21@158.223.51.76] has joined #sbcl 17:19:26 -!- ChanServ has set mode +o Krystof 17:32:44 rmarynch [~roman@88.135.194.233] has joined #sbcl 17:43:15 Good evening 17:45:14 Could somebody review (and maybe commit) this simple fix: https://bugs.launchpad.net/sbcl/+bug/520607 ? Also, will there be a release of SBCL at the end of this month? 17:51:40 I might take a look at it in a bit, since I'm planning to commit something later today anyway. 17:51:52 nyef: thank you! 17:54:06 In fact, there are near 20 tickets with the 'review' tag from different contributors, so the next release has some potential basis 18:06:44 I've got a commit series that I've been holding off on for months. :-/ 18:07:10 To the point, in fact, where I've forgotten where I put it and what it does. 18:07:25 Oh, right. TLS-index fixups. 18:10:37 -!- Krystof [~csr21@158.223.51.76] has quit [Ping timeout: 265 seconds] 18:15:48 redline6561 [~user@gate-22.spsu.edu] has joined #sbcl 18:20:56 -!- gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has quit [Quit: Leaving] 18:23:39 nyef: not wider fixnums? :) 18:26:17 I should probably look those over too, huh? 18:31:13 gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has joined #sbcl 18:32:25 do wider fixnums have any trade-offs? 18:32:54 Fewer spare lowtags? 18:33:06 besides that 18:33:32 And, while we figured out how to make them wider on 32-bit targets, it makes typechecking more expensive for some things. 18:34:06 and it's better on 64-bit? 18:34:28 64-bit already has sufficient "spare" lowtags. 18:34:50 Oh! The other cost on 32-bit was doubling the allocation granularity for non-cons objects. 18:35:25 i should really go back to using a 64-bit system, but i'm too lazy 18:36:24 Yeah, well, I wouldn't mind 64-bithood here, but it'd involve retargetting the compiler. 18:37:25 Isn't 29 bit the most common fixnum value length for 32-bit implementations? Why should it be longer than that? 18:38:06 Umm. On SBCL it's 30 bits, and we could make it 31 with a bit of work. 18:38:21 rmarynch: because it's never enough 18:38:41 rmarynch: because ocaml can. 18:39:01 fair enough :) 18:43:03 -!- redline6561 [~user@gate-22.spsu.edu] has quit [Ping timeout: 246 seconds] 18:54:41 My IMHO is that speed is more important. Also, my colleague (he has LW license) has compared LW Ent with SBCL on cl-bench, and he said that LW x64 is faster in more than half of the tests. I was really surprised to hear that :( 18:55:24 fixnums give you se 18:55:26 speed 18:58:22 Better closure representations don't suck either. 18:59:18 And, and... Umm... Last night I took a look at how template VOP selection works, and was somewhat underwhelmed. I think we should be able to do better, which could help numeric code. 18:59:33 Unfortunately, I'm not sure /how/ we can do better yet, just that we should be able to. 18:59:38 BTW, can SBCL profile itself? 19:00:14 Statistical profiler, maybe. The other profiler would end up recursing, which would be bad. 19:01:03 not everything in cl-bench is meaningful for everyone, and what benchmarks actually measure can be surprising 19:01:26 Can surprise even the people writing them. 19:01:40 :) 19:03:57 I can see the floats microbenchmarks putting more pressure on the GC than anything else. 19:06:01 Because the floats get closed over? 19:06:24 Or because they're undeclared? 19:06:28 passed/returned over full calls. 19:06:34 Ah. 19:06:39 *nyef* shudders. 19:09:16 As far as I remember from that report, SBCL won numbers crunching (FFT etc), but it is slower in generic tests (calls/strings/MOP etc). Sometimes the difference was up to 3x, but usually 10-50%. Not so dramatically, but anyway it is a kind of an alarm :) 19:09:47 rmarynch: some of that may be due to a different type lattice, or incomplete standard compliance. 19:10:12 Mmm... NIL arrays, anyone? 19:10:14 (especially wrt CLOS and the MOP) 19:13:45 -!- nyef [~nyef@pool-70-20-57-204.man.east.myfairpoint.net] has quit [*.net *.split] 19:13:45 -!- christoph_debian [~user@oteiza.siccegge.de] has quit [*.net *.split] 19:13:45 -!- hargettp [~hargettp@dhcp-162.mirrorimage.net] has quit [*.net *.split] 19:13:45 -!- gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has quit [*.net *.split] 19:13:45 -!- homie` [~levgue@xdsl-84-44-209-222.netcologne.de] has quit [*.net *.split] 19:13:45 -!- jsnell [~jsnell@ash.snellman.net] has quit [*.net *.split] 19:13:45 -!- stassats [~stassats@wikipedia/stassats] has quit [*.net *.split] 19:13:45 -!- mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has quit [*.net *.split] 19:13:46 -!- cmm [~cmm@bzq-79-178-206-31.red.bezeqint.net] has quit [*.net *.split] 19:13:46 -!- derrotebaron [~derroteba@static.7.69.47.78.clients.your-server.de] has quit [*.net *.split] 19:13:49 -!- luis [~luis@nhop.r42.eu] has quit [*.net *.split] 19:13:49 -!- foom [~jknight@ita4fw1.itasoftware.com] has quit [*.net *.split] 19:18:25 gnooth [~gnooth@ip98-176-79-151.sd.sd.cox.net] has joined #sbcl 19:18:25 hargettp [~hargettp@dhcp-162.mirrorimage.net] has joined #sbcl 19:18:25 nyef [~nyef@pool-70-20-57-204.man.east.myfairpoint.net] has joined #sbcl 19:18:25 homie` [~levgue@xdsl-84-44-209-222.netcologne.de] has joined #sbcl 19:18:25 stassats [~stassats@wikipedia/stassats] has joined #sbcl 19:18:25 mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has joined #sbcl 19:18:25 jsnell [~jsnell@ash.snellman.net] has joined #sbcl 19:18:25 derrotebaron [~derroteba@static.7.69.47.78.clients.your-server.de] has joined #sbcl 19:18:25 cmm [~cmm@bzq-79-178-206-31.red.bezeqint.net] has joined #sbcl 19:18:25 christoph_debian [~user@oteiza.siccegge.de] has joined #sbcl 19:18:25 foom [~jknight@ita4fw1.itasoftware.com] has joined #sbcl 19:18:25 luis [~luis@nhop.r42.eu] has joined #sbcl 19:28:08 -!- rmarynch [~roman@88.135.194.233] has quit [Quit: Leaving] 19:35:34 -!- hargettp [~hargettp@dhcp-162.mirrorimage.net] has quit [Read error: Connection reset by peer] 19:55:15 rmarynch [~roman@88.135.194.233] has joined #sbcl 19:57:25 http://sbcl.boinkor.net/boinkmarks/index site shows that some tests have 2x degradation in 2010 year releases. Interesting... 20:05:53 CLOS/methodcalls, maybe? 20:07:10 Heh. And that jump in particular was caused by a commit titled "much faster make-initial-dfun". 20:08:12 nyef: scroll more:) Well, some have no 2x, but close to that. 20:08:40 Yeah, the site is just really slow to respond in my browser, for whatever reason. :-/ 20:09:32 nyef: In my too. 20:09:49 Hunh. What happened to CTAK? 20:10:57 STAK has the degradation too, along with PUZZLE 20:11:10 some of might just be threads being enabled. 20:11:18 (1.0.33.26) 20:11:43 Spot on, 1.0.33.26. 20:13:14 that's almost the only change that's not realted to bootstrap in the release. 20:15:42 wonder what string-concat is 20:16:44 "Almost"? Feels like a good chunk of the 1.0.33.x series is windows and ldb fixes. 20:17:08 ok, that's out of the way for these tests as well ;) 20:18:56 <|3b|`> threading related ones are pretty obvious when you turn on the display of (:sb-thread) implementation 20:19:54 cl-bench.arrays:bench-string-concat code: http://paste.lisp.org/display/118808 20:20:30 rmarynch: I know what the test is; I'm wondering about the jump 20:21:17 pkhuong: clear 20:22:15 mm... didn't we change the way strings were read at some point? (base-string versus full string) 20:23:34 I feel that we should include cl-bench to the regression suite in some way 20:24:05 it's a post-commit hook already. 20:24:55 pkhuong: does it report the degradation if any? 20:27:49 I don't think so. If there's a simple way to detect meaningful degradation, antifuchs might be open to a patch. 20:28:05 <|3b|`> didn't that sort of thing go to planet sbcl at one point? 20:29:05 no, there is only a link to the site there 20:29:56 (the last 'there' was redundant in the above sentence) 20:34:36 yes it used to but then a new version of boinkmarks was made without an RSS feed Xach could 'plug' into.. IIRC 20:39:48 -!- rmarynch [~roman@88.135.194.233] has quit [Quit: Leaving] 20:44:27 We could do with a new version of boinkmarks that isn't so abominably slow to browse. 20:47:24 nyef: I think it's the fancy graphing that slows things down. the json data appears to come back pretty fast 20:51:18 rpg [~rpg@mpls.sift.info] has joined #sbcl 20:51:30 Krystof [~csr21@csrhodes.plus.com] has joined #sbcl 20:51:30 -!- ChanServ has set mode +o Krystof 21:14:29 Hrm... fasloading seems to slowly get better over time. 21:26:07 Okay, PUZZLE, RICHARDS, and probably STAK are all SYMBOL-VALUE and SET-SYMBOL-VALUE-bound. 21:27:30 -!- mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has quit [Read error: Connection reset by peer] 21:27:43 mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has joined #sbcl 21:28:48 ... And the STRING-CONCAT damage is due to the recent numeric type union approximation stuff... 21:29:00 And the more I think about that, the less I like it. 21:41:31 Fare [~Fare@ita4fw1.itasoftware.com] has joined #sbcl 21:49:51 -!- rpg [~rpg@mpls.sift.info] has quit [Ping timeout: 240 seconds] 22:29:26 redline6561 [~user@c-66-56-55-169.hsd1.ga.comcast.net] has joined #sbcl 22:30:22 -!- luis [~luis@nhop.r42.eu] has quit [*.net *.split] 22:30:22 -!- foom [~jknight@ita4fw1.itasoftware.com] has quit [*.net *.split] 22:32:57 luis [~luis@nhop.r42.eu] has joined #sbcl 22:32:58 foom [~jknight@ita4fw1.itasoftware.com] has joined #sbcl 22:36:41 -!- mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has quit [Ping timeout: 276 seconds] 23:00:08 -!- slyrus_ [~chatzilla@adsl-75-36-217-249.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 265 seconds] 23:10:59 rpg [~rpg@216.243.156.16] has joined #sbcl 23:17:38 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 276 seconds] 23:23:20 redline6` [~user@c-66-56-55-169.hsd1.ga.comcast.net] has joined #sbcl 23:24:01 -!- redline6` [~user@c-66-56-55-169.hsd1.ga.comcast.net] has quit [Client Quit] 23:42:45 -!- rpg [~rpg@216.243.156.16] has quit [Quit: rpg] 23:51:39 -!- Fare [~Fare@ita4fw1.itasoftware.com] has quit [Quit: Leaving] 23:57:43 Blkt [~Blkt@dynamic-adsl-94-37-227-177.clienti.tiscali.it] has joined #sbcl