00:08:32 -!- whoops [u549@gateway/web/irccloud.com/x-dxvkgbtwayparack] has quit [Remote host closed the connection] 00:34:55 Volundr [~whoops@li257-204.members.linode.com] has joined #sbcl 00:37:08 -!- Volundr is now known as whoops 00:40:51 -!- homie` [~levgue@xdsl-78-35-162-208.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 00:43:42 homie [~levgue@xdsl-78-35-162-208.netcologne.de] has joined #sbcl 03:05:58 attila_lendvai [~attila_le@87.247.48.240] has joined #sbcl 03:05:58 -!- attila_lendvai [~attila_le@87.247.48.240] has quit [Changing host] 03:05:58 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 03:37:25 echo-area [~user@182.92.247.2] has joined #sbcl 03:57:09 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 04:00:56 -!- pkhuong [~pkhuong@gravelga.xen.prgmr.com] has quit [Ping timeout: 252 seconds] 04:09:39 pkhuong [~pkhuong@gravelga.xen.prgmr.com] has joined #sbcl 04:13:54 -!- pkhuong [~pkhuong@gravelga.xen.prgmr.com] has quit [Ping timeout: 245 seconds] 04:16:35 pkhuong [~pkhuong@gravelga.xen.prgmr.com] has joined #sbcl 06:30:05 sdemarre [~serge@91.176.94.14] has joined #sbcl 06:50:22 Cryotank2011 [~Cryotank2@c-24-17-62-152.hsd1.wa.comcast.net] has joined #sbcl 06:54:28 angavrilov [~angavrilo@217.71.235.212] has joined #sbcl 07:05:45 echo-area [~user@182.92.247.2] has joined #sbcl 07:16:53 -!- angavrilov [~angavrilo@217.71.235.212] has quit [*.net *.split] 07:16:53 -!- christoph_debian [~user@oteiza.siccegge.de] has quit [*.net *.split] 07:16:53 -!- drl [~lat@110.139.229.172] has quit [*.net *.split] 07:17:58 angavrilov [~angavrilo@217.71.235.212] has joined #sbcl 07:17:58 christoph_debian [~user@oteiza.siccegge.de] has joined #sbcl 07:17:58 drl [~lat@110.139.229.172] has joined #sbcl 07:44:25 -!- Cryotank2011 [~Cryotank2@c-24-17-62-152.hsd1.wa.comcast.net] has quit [Quit: Cryotank2011] 07:57:52 -!- ddp [~ddp@anon-158-35.relakks.com] has quit [Remote host closed the connection] 07:58:02 ddp [~ddp@216.243.111.165] has joined #sbcl 08:00:16 -!- ddp [~ddp@216.243.111.165] has quit [Client Quit] 08:30:11 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 08:35:59 -!- Kryztof [~user@77-58-246-74.dclient.hispeed.ch] has quit [Ping timeout: 245 seconds] 08:37:38 Blkt [~user@89-96-199-46.ip13.fastwebnet.it] has joined #sbcl 08:40:01 -!- scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has quit [Ping timeout: 252 seconds] 09:08:44 homie` [~levgue@xdsl-84-44-152-107.netcologne.de] has joined #sbcl 09:11:49 -!- homie [~levgue@xdsl-78-35-162-208.netcologne.de] has quit [Ping timeout: 276 seconds] 09:26:26 Kryztof [~user@178-83-229-138.dynamic.hispeed.ch] has joined #sbcl 09:48:49 -!- drl [~lat@110.139.229.172] has quit [Quit: Leaving] 12:33:42 scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has joined #sbcl 12:35:35 -!- scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has quit [Remote host closed the connection] 12:38:42 scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has joined #sbcl 13:05:16 -!- homie` [~levgue@xdsl-84-44-152-107.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 13:32:34 homie [~levgue@xdsl-84-44-152-107.netcologne.de] has joined #sbcl 14:48:02 LiamH [~none@pdp8.nrl.navy.mil] has joined #sbcl 15:46:12 loke [~elias@bb115-66-85-121.singnet.com.sg] has joined #sbcl 15:47:38 Is there a reason why SBCL does not optimise (+ 1 (+ 1 x)) to one ADD instruction? This happens even if X is declared (type (integer 0 1000) x) 15:48:05 Shouldn't it be able to fold it into a single ADD instruction? 15:50:51 dtw [dtw@pdpc/supporter/active/dtw] has joined #sbcl 15:51:11 loke: it doesn't have anything for that pattern. 15:51:23 pkhuong: would it be easy to implement? 15:51:31 It Depends? 15:51:38 (and its infrastructure isn't designed to implicitly simplify those) 15:51:44 I see 15:52:21 probably easier at the VOP level than on the expression-oriented IR 15:53:25 I mean, I don't worry that much about the somewhat synthetic case of (+ (+ ...)) 15:53:51 But I wondered if it doesn't do that, are there some major set of optimisations like that that it does not do? 15:54:29 no CSE, no code hoisting, no strength reduction of loop variables 15:54:45 regalloc on variables, not values, no live range splitting 15:55:06 only greedy instruction selection 15:55:09 no scheduling 15:56:44 I just checked CCL, it does much, much worse :-) 15:56:55 OPTIMIZATION has more stuff 15:58:05 ECL does that kind of reduction though... :-) 15:59:37 GCC. 16:00:16 yeah, but I checked the disassembly from ECL, which shows some kind of intermediate code. It already reduced it to a call to (+ 2 ...) there 16:01:55 a peephole pass would get that right. 16:01:58 But in general terms. Do you feel that the SBCL compilers leaves a lot to be desired? 16:02:29 yes. It also doesn't matter that much. 16:02:41 c.f. Proebsting's law. 16:02:53 which law is that? 16:03:07 google (: 16:08:48 argh 16:08:57 So sad :-) 16:09:20 Then again, 18 years already passed. almost 3 times since Lisp came to be 16:17:48 -!- ASau [~user@95-24-150-243.broadband.corbina.ru] has quit [Remote host closed the connection] 16:18:07 ASau [~user@93-80-123-238.broadband.corbina.ru] has joined #sbcl 16:22:09 SBCL probably hits that ~8x figure easily. 16:33:32 milanj [~milanj_@109-93-200-141.dynamic.isp.telekom.rs] has joined #sbcl 16:34:27 quite: see e.g. http://sbcl.boinkor.net/boinkmarks/index#BENCH-STRINGS for a 4x in 5 years! So far ahead of the curve 16:36:07 silver [~kingrat@178.121.134.139] has joined #sbcl 16:36:47 -!- redline6561 is now known as redline6561_brb 16:38:02 fun to see the effect of turning on threads by default: http://sbcl.boinkor.net/boinkmarks/index#BENCH-STRINGS;SBCL,(:ARCH :X86_64 :FEATURES NIL),baker/1.0.33 16:44:44 -!- Kryztof [~user@178-83-229-138.dynamic.hispeed.ch] has quit [Ping timeout: 240 seconds] 17:01:32 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 17:06:02 hroptatyr [~asathor@sxemacs/devel/hroptatyr] has joined #sbcl 17:17:37 -!- Blkt [~user@89-96-199-46.ip13.fastwebnet.it] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 17:19:53 has anyone ever seen the semaphore code hanging in run-tests.sh? 17:20:15 here's a live log of an sbcl build on the opensuse build server 17:20:29 its waiting for a semaphore for 8 hours now 17:20:32 https://build.opensuse.org/package/live_build_log?arch=x86_64&package=sbcl&project=home%3Arudi_m%3Alisp&repository=openSUSE_Factory 17:25:56 -!- cmm [~cmm@109.65.208.18] has quit [Ping timeout: 240 seconds] 17:27:52 cmm [~cmm@bzq-79-177-193-169.red.bezeqint.net] has joined #sbcl 17:34:51 antgreen [~user@bas3-toronto06-1177890396.dsl.bell.ca] has joined #sbcl 17:36:07 -!- redline6561_brb is now known as redline6561 17:44:15 hroptatyr: on linux? That's surprising. 17:45:45 yes on linux 17:46:05 x86_64 17:46:38 qemu (?!) 17:47:23 in this case yes, kvm even 17:47:36 can you try and replicate on a straight linux/baremetal box? 17:47:43 but the guy who reported it to me claimed he saw it locally too 17:48:06 It's reallly surprising, because most of the code is either straightforward or directly from the kernel docs 17:48:31 well i dont know, i was just gonna bring it to your attention 17:57:40 KVM shouldn't make a difference 17:58:49 Qworkescence [~quad@unaffiliated/quadrescence] has joined #sbcl 18:05:42 i.e., I'm not saying that it _doesn't_ make a difference technically, of course :-), but KVM is pretty good in the Linux virtualization department these days, so if SBCL fails there, it probably fails on most other Linux servers out there, too 18:09:15 i'm totally sure my mate reproduced it on a bare metal box as well 18:09:44 its just very hard to debug i imagine, in case it's a race condition 18:13:38 leuler [~user@p54905444.dip.t-dialin.net] has joined #sbcl 18:19:26 I have seen this blocked semaphore test once recently. It was not reproducible. I am on Linux x86-64, non-virtualized. It may have been with an x86 build; I don't remember. 18:22:03 sigh. 18:22:19 Our semaphore code is a bit stupid though. 18:22:50 (they're implemented with lisp-side mutex/condvar) 18:27:03 -!- scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has quit [Ping timeout: 252 seconds] 18:30:04 scymtym [~user@2001:638:504:2093:21a:a0ff:fe34:2d7d] has joined #sbcl 18:31:10 ddp [~ddp@216.243.111.165] has joined #sbcl 18:48:15 pkhuong: I seem to be able to reproduce the hang currently relatively easily. As soon as I attach strace to the seemingly hung SBCL (which has 50 threads at that moment), it starts to stop them and shortly after that the test finishes successfully. 18:51:09 -!- dtw [dtw@pdpc/supporter/active/dtw] has quit [Quit: +++ATH0] 18:52:32 This is with SBCL 1.0.55, x86-64, on Linux 3.2.0, Preemptible Kernel. 18:58:04 tsuru` [~charlie@adsl-74-179-28-117.bna.bellsouth.net] has joined #sbcl 19:07:27 See http://http://paste.lisp.org/display/127148 for the (shortened) output of strace. 19:58:31 More info: The test hangs at the point after joining the "interruptor" thread when trying to join the first thread on the "safe" list, "T99". I have annotated the paste with the stack trace from lisp, showing that the test is stuck waiting for the "thread result lock" mutex. 20:01:54 -!- homie [~levgue@xdsl-84-44-152-107.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 20:10:09 -!- antgreen [~user@bas3-toronto06-1177890396.dsl.bell.ca] has quit [Read error: Connection reset by peer] 20:11:14 -!- ddp [~ddp@216.243.111.165] has quit [Remote host closed the connection] 20:11:30 ddp [~ddp@anon-159-72.relakks.com] has joined #sbcl 20:44:53 -!- milanj [~milanj_@109-93-200-141.dynamic.isp.telekom.rs] has quit [Ping timeout: 260 seconds] 21:01:05 homie [~levgue@xdsl-84-44-152-107.netcologne.de] has joined #sbcl 21:02:25 Cryotank2011 [~Cryotank2@c-24-17-62-152.hsd1.wa.comcast.net] has joined #sbcl 21:19:14 -!- homie [~levgue@xdsl-84-44-152-107.netcologne.de] has quit [Read error: Connection reset by peer] 21:20:12 homie [~levgue@xdsl-78-35-141-71.netcologne.de] has joined #sbcl 21:36:16 -!- sdemarre [~serge@91.176.94.14] has quit [Ping timeout: 240 seconds] 21:42:49 -!- Posterdati [~tapioca@host188-225-dynamic.10-87-r.retail.telecomitalia.it] has quit [Read error: Connection reset by peer] 21:44:03 Blkt [~user@82.84.176.4] has joined #sbcl 21:52:55 Posterdati [~tapioca@host188-225-dynamic.10-87-r.retail.telecomitalia.it] has joined #sbcl 21:58:14 good evening everyone 21:59:18 hello Blkt 22:01:33 :) 22:36:03 -!- silver [~kingrat@178.121.134.139] has quit [Read error: Connection reset by peer] 22:53:30 -!- homie [~levgue@xdsl-78-35-141-71.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 23:17:50 ddp_ [~ddp@216.243.111.165] has joined #sbcl 23:21:41 -!- ddp [~ddp@anon-159-72.relakks.com] has quit [Ping timeout: 252 seconds] 23:21:41 -!- ddp_ is now known as ddp 23:24:25 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 23:32:39 homie [~levgue@xdsl-78-35-141-71.netcologne.de] has joined #sbcl