00:12:39 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 00:25:47 -!- Blkt [~user@82.84.181.193] has quit [Ping timeout: 240 seconds] 01:02:04 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 01:22:18 -!- Thra11 [~thrall@87.114.93.105] has quit [Quit: kthxbai] 02:20:26 prxq_ [~mommer@mnhm-590c2053.pool.mediaWays.net] has joined #sbcl 02:23:26 -!- prxq [~mommer@mnhm-5f75e325.pool.mediaWays.net] has quit [Ping timeout: 255 seconds] 03:31:51 echo-area [~user@182.92.247.2] has joined #sbcl 03:50:45 -!- Fare [~fare@173-9-65-97-NewEngland.hfc.comcastbusiness.net] has quit [Ping timeout: 248 seconds] 04:01:04 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 04:03:38 Fare [fare@nat/google/x-ovmjbyouqbgqrdsk] has joined #sbcl 04:44:49 -!- wbooze [~wbooze@xdsl-87-79-198-110.netcologne.de] has quit [Quit: none] 06:09:40 cmm- [~cmm@bzq-79-181-191-53.red.bezeqint.net] has joined #sbcl 06:14:48 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [*.net *.split] 06:14:49 -!- cmm [~cmm@bzq-79-181-191-53.red.bezeqint.net] has quit [*.net *.split] 06:20:44 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 06:34:02 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Ping timeout: 248 seconds] 06:34:38 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 06:57:46 I've got a feature request ... it would be nice if SB-PCL::SLOT-TYPECHECK would get (and report) the slot that's currently checked. 07:04:38 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Quit: Leaving.] 07:19:00 -!- prxq_ is now known as prxq 07:26:30 -!- Fare [fare@nat/google/x-ovmjbyouqbgqrdsk] has quit [Ping timeout: 264 seconds] 07:39:59 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 265 seconds] 08:31:16 kanru`` [~kanru@217.243.168.51] has joined #sbcl 08:32:34 -!- ASau [~user@46.115.54.144] has quit [Ping timeout: 276 seconds] 09:08:47 snowylike [~sn@91-67-171-156-dynip.superkabel.de] has joined #sbcl 09:14:11 -!- Krystof [~user@81.174.155.115] has quit [Ping timeout: 255 seconds] 09:27:59 tcr [~tcr@178-83-229-138.dynamic.hispeed.ch] has joined #sbcl 10:36:12 stassats [~stassats@wikipedia/stassats] has joined #sbcl 11:23:10 hydan [~user@ip-89-102-13-27.net.upcbroadband.cz] has joined #sbcl 11:30:11 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 255 seconds] 11:40:21 -!- hydan [~user@ip-89-102-13-27.net.upcbroadband.cz] has quit [Remote host closed the connection] 11:40:54 hydan [~user@ip-89-102-13-27.net.upcbroadband.cz] has joined #sbcl 12:41:00 Krystof [~user@81.174.155.115] has joined #sbcl 12:41:00 -!- ChanServ has set mode +o Krystof 13:13:54 Fare [fare@nat/google/x-pqbvegjuoufkxzpn] has joined #sbcl 13:25:08 Thra11 [~thrall@87.114.93.105] has joined #sbcl 13:48:18 -!- Krystof [~user@81.174.155.115] has quit [Remote host closed the connection] 14:15:23 Krystof [~user@81.174.155.115] has joined #sbcl 14:15:23 -!- ChanServ has set mode +o Krystof 14:20:50 stassats [~stassats@wikipedia/stassats] has joined #sbcl 14:27:48 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 256 seconds] 14:33:52 -!- snowylike [~sn@91-67-171-156-dynip.superkabel.de] has quit [Quit: Nettalk6 - www.ntalk.de] 14:33:56 wbooze [~wbooze@xdsl-78-35-130-21.netcologne.de] has joined #sbcl 14:35:12 LiamH [~none@pdp8.nrl.navy.mil] has joined #sbcl 14:39:53 -!- huangjs [~huangjs@69.84.244.131] has quit [Quit: This computer has gone to sleep] 14:51:36 psilord [~psilord@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 15:02:34 -!- hydan [~user@ip-89-102-13-27.net.upcbroadband.cz] has quit [Remote host closed the connection] 15:07:18 huangjs [~huangjs@69.84.244.131] has joined #sbcl 15:12:41 hydan [~user@ip-89-102-13-27.net.upcbroadband.cz] has joined #sbcl 15:13:53 -!- hydan [~user@ip-89-102-13-27.net.upcbroadband.cz] has quit [Remote host closed the connection] 15:14:52 Is it possible somehow to create a "true" (byte-array 8) in SBCL? Some kind of combination of declarations that would cause an array to only allocate 8 bits per byte, and not use any tagging for the members? 15:15:56 a good question.... 15:17:23 <|3b|> (make-array size :element-type '(unsigned-byte 8))? 15:18:03 |3b|: Just tried. The entries still have type tags 15:18:32 <|3b|> then you did something wrong 15:20:08 does he have to set optimizing off ? 15:20:38 <|3b|> not to create an array 15:20:54 ok 15:21:07 http://paste.lisp.org/display/134501 15:22:12 *|3b|* has no idea what that is supposed to show, but notes that the array indices are probably fixnums, the values extracted from the array might be being converted to fixnums, etc 15:22:32 I pasted the wrong version of the function. It has (declare (optimize (speed 3) (safety 0) (debug 0))) in there 15:22:33 um, so that assembly is clearly showing the elements being read from the array and then tagged 15:22:59 left... 15:23:00 argh 15:23:13 spank me, please. 15:25:27 But it does beg another question: It seems a bit inefficient to shift in the tag before the addition, yes? Is this a limit of the optimiser? 15:28:37 wrap the + inside a (ldb (byte 0 32) ...) 15:29:05 or is that 32 0? 15:30:47 but yes, in an ideal world the tagging would only happen once 15:32:04 You're right: http://paste.lisp.org/display/134506 15:32:59 that doesn't look like it has an addition, so the args should probably have been the other way around 15:33:05 I am not entirely sure how it works though... It seems to overwrite EAX with each cell, but there is no addition happning? 15:33:11 ah 15:33:49 Yes, with the arguments the other way I'm back to the original version. SHL before each add 15:40:01 ok, not entirely surprising 15:42:07 jsnell: I wonder if there would be any point in adding such optimisation to SBCL, or if I've hit an edge case 15:42:38 Basically, is my use-case somewhat unique? (I've been trying to create the fastest possible Brainfuck compiler) 15:48:43 sbcl is very bad at optimizing the number of shift operations. 15:48:57 luckily, in my experience, x86 cpus are very good at it. 15:49:20 the extra shifts seem to have pretty negligible performance impact. 15:49:48 fair enough 15:50:46 this particular case should also be very low-hanging fruit 15:51:53 since it'd likely be due to the way the unsigned + operations are set up to have a higher cost than fixnum + 15:51:53 In brainfuck in particular, there is really nothing but loops and add/subtract of elements in a byte array 15:52:10 ooh 15:52:26 so you're saying is that if the cells were fixnum, it might not do the shifting? 15:52:52 aaaaand... 15:52:54 you're right 15:53:16 changing the array elements to fixnum eliminated the SHL's 15:53:47 that said... hmm... perhaps the array elements are already tagged? 15:54:23 yes, that's what a fixnum :element-type would mean. and it's going to be 8 bytes per element 15:54:39 yeah 16:01:37 OK, I might have improved things 16:03:06 another way would be to influence representation selection with appropriate type declarations and assignments to bamboozle the optimiser. 16:03:12 At least things changed a bit... My intel assembly is too bad to understand fully what's going on 16:03:27 http://paste.lisp.org/display/134509 16:04:44 Frankly, I don't understand what happens here 16:05:47 why do you care? 16:05:55 (bug in my f+, but nothing that has any impact here) 16:06:19 pkhuong: pathological interest I suppose 16:06:57 pkhuong: I found it to be somewhat interesting to see how my code is represented by the sbcl compiler 16:07:20 6 times, it loads a byte and zero-extends it to 32 bit (which are then sign-extended to 64), then tags the result as a fixnum 16:07:22 that's not really better, just a different way of shifting 16:08:48 jsnell: Ah yes. I see that now. LEA ... [RAX+RAX] is of course loading the effective address, which has been doubled (i.e. equivalent to SHL) 16:09:01 so obvious now 16:09:05 thanks for your help 16:32:26 -!- tcr [~tcr@178-83-229-138.dynamic.hispeed.ch] has quit [Quit: Leaving.] 16:36:39 Thra11_ [~thrall@87.115.149.255] has joined #sbcl 16:38:47 -!- Thra11 [~thrall@87.114.93.105] has quit [Ping timeout: 260 seconds] 17:10:03 MrMc [~user@91-64-125-247-dynip.superkabel.de] has joined #sbcl 17:14:07 -!- Thra11_ [~thrall@87.115.149.255] has quit [Ping timeout: 246 seconds] 17:15:27 hello all 17:41:43 Thra11 [~thrall@87.113.136.86] has joined #sbcl 17:42:50 Urfin [~user@213.57.184.94] has joined #sbcl 17:43:03 -!- Urfin [~user@213.57.184.94] has left #sbcl 18:02:44 do threads work in OpenBSD 18:02:57 ? 18:03:01 MrMc: no. 18:03:33 where in the code shall i start 18:03:53 looking? 18:04:24 MrMc: IIUC, OpenBSD recently moved to a 1:1 model, so it might be doable. I doubt there's anything like Free's linux compatibility layer, so emulating futexes doesn't sound like a good idea. 18:04:46 Instead, you'd want to go with something like Darwin's approach. 18:05:34 -!- kanru`` [~kanru@217.243.168.51] has quit [Ping timeout: 260 seconds] 18:05:50 pkhuong:what is the approach in Darwin? 18:06:29 note that the darwin port has extra cruft for mach, so i'd start with a hack job with that and the Free code 18:06:38 see src/runtime/*-os.c 18:10:14 stassats [~stassats@wikipedia/stassats] has joined #sbcl 18:14:02 ASau [~user@92.116.69.196] has joined #sbcl 18:20:23 snowylike [~sn@91-67-171-156-dynip.superkabel.de] has joined #sbcl 18:28:37 -!- wbooze [~wbooze@xdsl-78-35-130-21.netcologne.de] has quit [Ping timeout: 265 seconds] 18:30:26 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 18:39:17 pkhuong:thanks I am starting to work 18:39:39 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 260 seconds] 18:40:00 is the bsd-os.c for freebsd? 18:41:01 netbsd and openbsd too 18:42:02 and darwin. 18:52:58 edgar-rft [~GOD@HSI-KBW-091-089-005-153.hsi2.kabelbw.de] has joined #sbcl 18:59:10 -!- Fare [fare@nat/google/x-pqbvegjuoufkxzpn] has quit [Ping timeout: 276 seconds] 19:12:09 pnpuff [~Dioxirane@unaffiliated/pnpuff] has joined #sbcl 19:16:53 Fare [fare@nat/google/x-wikpnymfljcamyke] has joined #sbcl 19:59:16 -!- pnpuff [~Dioxirane@unaffiliated/pnpuff] has quit [Quit: .] 20:37:29 -!- huangjs [~huangjs@69.84.244.131] has quit [Quit: Leaving] 20:40:34 -!- Fare [fare@nat/google/x-wikpnymfljcamyke] has quit [Ping timeout: 265 seconds] 20:49:06 pnpuff [~Dioxirane@unaffiliated/pnpuff] has joined #sbcl 20:53:32 rbarraud [~rbarraud@125-239-32-133.jetstream.xtra.co.nz] has joined #sbcl 20:58:22 I am trying to compile SBCL from git on OpenBSD current amd64 and the first sb-thread test fails 21:00:26 compilation comes to a grinding halt when doing an sb-sprof test there is mention of SB-SAFEPOINT-STRICTLY 21:01:50 in the test's comments can this be ignored or this has to be implemented for OpenBSD? 21:02:14 is that with or without threads? 21:02:26 this is with threads 21:03:36 I'd try a couple trivial smoke tests first. Spawn a thread, get its return value, print from a thread, trigger GC from a thread, ... 21:03:36 I'm losing track of what's going on here. You've ported sb-thread to OpenBSD within the last 3 hours? 21:03:45 -!- snowylike [~sn@91-67-171-156-dynip.superkabel.de] has quit [Quit: Nettalk6 - www.ntalk.de] 21:05:16 If so, that's exciting. And -- what pkhuong says. 21:05:28 (But after that, the main threading tests are not in the contribs at all, but rather in threads.impure.lisp. That's assuming you don't have regressions in the rest of tests/.) 21:08:32 Sorry to disappoint you but I just looked around for clues for this to compile 21:10:16 # 21:11:25 erm. If you're failing contrib tests, you have already compiled SBCL succesfully in my book. 21:11:54 -!- pnpuff [~Dioxirane@unaffiliated/pnpuff] has quit [Quit: .] 21:20:25 Fare [fare@nat/google/x-vmwhdudwmlxanvww] has joined #sbcl 21:20:59 lichtblau: i am susceptible to flattery. how do i turn off this sb-sprof test so that I can start what pkhuong suggested? 21:21:33 MrMc: just run sbcl 21:21:43 sh run-sbcl.sh --no-userinit 21:25:36 Ok. the compilation-test process is hanging and I will have to send signal -9 to the sb-sprof test so I can still do the above? 21:27:09 yes. 21:27:43 tcr [~tcr@84-72-21-32.dclient.hispeed.ch] has joined #sbcl 21:32:46 kanru`` [~kanru@62-50-217-185.client.stsn.net] has joined #sbcl 21:53:53 -!- tcr [~tcr@84-72-21-32.dclient.hispeed.ch] has quit [Quit: Leaving.] 22:01:20 well the trivial tests of creating threads work 22:05:53 (dotimes (x 2) (make-thread (lambda () (loop (sb-sprof::consalot))))) 22:06:29 should tell you whether the earlier sprof failure was due to profiling or just threading as such, since consalot itself does not involve profiling 22:07:55 -!- Fare [fare@nat/google/x-vmwhdudwmlxanvww] has quit [Quit: Leaving] 22:08:17 sb-sprof did not build :-( 22:08:22 copy the code. 22:10:09 i did and where can i paste the error 22:10:47 its essentially telling me sb-sprof does not exist 22:12:43 MrMc: the code for sb-sprof::consalot. 22:17:09 http://paste.lisp.org/+2VSG 22:18:11 Yes let me try that 22:26:55 I got nil after trying the supplied snippet 22:31:01 tcr [~tcr@84-72-21-32.dclient.hispeed.ch] has joined #sbcl 22:49:09 Fare [fare@nat/google/x-cnygogqvmzltiyqz] has joined #sbcl 22:49:19 pkhuong + lichtblau thank you for your support. I am going to install hunchentoot and do a bit if testing 22:49:39 do you understand what the snippet is testing? 22:49:45 or how failures would manifest? 22:50:07 not quite 22:51:22 could you explain 22:51:50 its making two threads 22:51:57 that consalot? 22:54:37 you're testing allocation and garbage collection, from two threads at the same time. that could fail by hanging, or by visibly crashing. in any case, it could be a while before the symptoms are showing -- long after the main thread printed that nil. If it truly runs smoothly, that's a good sign. 22:57:06 thanks for the explanation I am still a newcomer 22:59:44 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 22:59:48 things could also work v. well if threads never become active. 23:05:38 so generally, as much as you're hoping for the shiny new threads to run perfectly, your job is rather to find any way to break them 23:06:04 -!- psilord [~psilord@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Quit: Leaving.] 23:25:28 well i will retire for now and will not miss to post my findings in the mailing list 23:26:50 thank you for your support. 23:32:51 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 23:44:43 Thra11_ [~thrall@146.90.177.145] has joined #sbcl 23:47:15 -!- Thra11 [~thrall@87.113.136.86] has quit [Ping timeout: 248 seconds]