01:00:31 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 01:58:07 -!- Bike [~Glossina@75-175-67-103.ptld.qwest.net] has quit [Ping timeout: 240 seconds] 02:00:37 echo-area [~user@182.92.247.2] has joined #sbcl 02:00:40 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 02:06:23 Bike [~Glossina@75-175-67-103.ptld.qwest.net] has joined #sbcl 02:07:03 Bike_ [~Glossina@75-175-67-103.ptld.qwest.net] has joined #sbcl 02:11:06 -!- Bike [~Glossina@75-175-67-103.ptld.qwest.net] has quit [Ping timeout: 268 seconds] 02:13:14 -!- Bike_ is now known as Bike 02:33:30 ehaliewi` [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 02:33:35 -!- ehaliewi` [~user@50-0-51-11.dsl.static.sonic.net] has quit [Remote host closed the connection] 02:39:28 -!- christoph_debian [~christoph@host-188-174-218-212.customer.m-online.net] has quit [Ping timeout: 256 seconds] 02:41:58 echo-are` [~user@182.92.247.2] has joined #sbcl 02:44:55 milosn_ [~milosn@user-5af5015d.broadband.tesco.net] has joined #sbcl 02:47:47 -!- echo-area [~user@182.92.247.2] has quit [Ping timeout: 240 seconds] 02:47:48 -!- Kromitvs [~quassel@mcqueen.rnl.ist.utl.pt] has quit [Quit: No Ping reply in 180 seconds.] 02:47:49 Kromitvs [~quassel@mcqueen.rnl.ist.utl.pt] has joined #sbcl 02:47:49 -!- |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has quit [Remote host closed the connection] 02:47:50 |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has joined #sbcl 02:47:50 -!- milosn [~milosn@user-5af5015d.broadband.tesco.net] has quit [Read error: Connection reset by peer] 02:52:43 christoph_debian [~christoph@ppp-188-174-122-198.dynamic.mnet-online.de] has joined #sbcl 02:57:38 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Ping timeout: 245 seconds] 03:01:21 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 03:14:13 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 03:27:30 -!- Bike [~Glossina@75-175-67-103.ptld.qwest.net] has quit [Ping timeout: 245 seconds] 03:29:53 Bike [~Glossina@75-175-67-103.ptld.qwest.net] has joined #sbcl 03:30:06 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 03:44:47 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Read error: Connection reset by peer] 03:45:10 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 03:47:26 -!- echo-are` is now known as echo-area 04:04:23 -!- Bike [~Glossina@75-175-67-103.ptld.qwest.net] has quit [Ping timeout: 240 seconds] 04:13:01 Bike_ [~Glossina@75-175-67-103.ptld.qwest.net] has joined #sbcl 04:20:45 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 256 seconds] 04:46:18 -!- Bike_ is now known as Bike 05:14:08 sdemarre [~serge@28.96-64-87.adsl-dyn.isp.belgacom.be] has joined #sbcl 05:22:15 -!- sdemarre [~serge@28.96-64-87.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 260 seconds] 05:36:49 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 06:03:12 dto [~user@pool-96-252-62-13.bstnma.fios.verizon.net] has joined #sbcl 06:03:54 pkhuong: some more interesting sbcl experiments: http://blocky.io/monks.ogv 06:14:03 -!- wbooze [~wbooze@xdsl-84-44-155-251.netcologne.de] has quit [Ping timeout: 260 seconds] 06:22:48 -!- jarod_ch_ [~jarod_che@115.193.190.118] has quit [Quit: Textual IRC Client: http://www.textualapp.com/] 06:29:37 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 06:49:21 -!- dto [~user@pool-96-252-62-13.bstnma.fios.verizon.net] has quit [Remote host closed the connection] 07:03:04 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 07:18:26 -!- minion [~minion@common-lisp.net] has quit [Remote host closed the connection] 07:18:33 minion [~minion@common-lisp.net] has joined #sbcl 07:26:33 benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has joined #sbcl 07:29:00 morning everyone 08:16:42 prxq [~mommer@mnhm-590c373f.pool.mediaWays.net] has joined #sbcl 09:15:17 -!- ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has quit [Ping timeout: 248 seconds] 10:04:46 -!- benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 10:09:44 -!- segv- [~mb@95-91-241-45-dynip.superkabel.de] has quit [Remote host closed the connection] 10:28:24 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 10:30:25 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 10:30:45 -!- danlentz [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has quit [Ping timeout: 264 seconds] 10:57:12 danlentz [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has joined #sbcl 11:18:57 benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has joined #sbcl 11:42:01 ASau` [~user@p4FF96FA0.dip0.t-ipconnect.de] has joined #sbcl 11:45:25 -!- ASau [~user@p4FF97D53.dip0.t-ipconnect.de] has quit [Read error: Operation timed out] 11:47:56 segv- [~mb@95-91-241-45-dynip.superkabel.de] has joined #sbcl 11:55:40 teggi [~teggi@123.25.129.139] has joined #sbcl 12:07:19 -!- Subfusc_ is now known as Subfusc 12:08:59 stassats [~stassats@wikipedia/stassats] has joined #sbcl 12:30:42 *stassats* got a gcc-compile-farm account 12:30:59 of useful system only a POWER7 machine seems to be working 12:31:19 good enough for me, maybe i can update the current binary 12:40:25 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 12:43:29 wasn't sbcl support only for ppc32 ? 12:43:46 power7 can run both 12:43:52 but yes 12:44:21 not only power7, but i think any ppc64 can run ppc32 too 12:45:25 -!- joshe [~joshe@2001:470:e862::1:1] has quit [Ping timeout: 245 seconds] 12:46:12 joshe [~joshe@2001:470:e862::1:1] has joined #sbcl 12:47:15 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 12:47:27 and the current HEAD is built successfully 12:47:43 just under 15 minutes, not bad 12:48:17 stassats: I was trying to build sbcl with the modified length function without the define-static-fun part 12:48:30 it builds, but blows up as I start it 12:48:43 note that I removed only the define-static-fun only for %list-length 12:48:55 did you remove it from the static functions list as well? 12:49:00 should I remove all define-satatic-fun for length aswell? 12:49:06 yes I did 12:49:40 does it die in cold init? 12:49:44 with a primitive halt? 12:50:19 yes if I remember correctly 12:50:31 and what is the backtrace? 12:50:37 let me compile it again 12:50:52 I have cleaned the tree somehow 12:51:14 -!- joshe [~joshe@2001:470:e862::1:1] has quit [Ping timeout: 264 seconds] 12:51:29 porting SBCL to ppc64 could be fun 12:51:35 imagine the enterprise possibilities! 12:51:53 true 12:52:18 if only the said enterprise could sponsor it 12:57:29 are you talking about IBM? 12:57:52 Blkt: enterprise as a whole 12:58:15 including IBM, but that seems unlikely 12:58:27 which enterprise machines are using ppc64? 12:59:01 well, ibm mainframes, but i rather meant people who use such machines 12:59:34 s/mainframes/servers/ 13:03:00 darn, the problem was my init configuration I forgot about 13:03:03 got a couple of test failures, 3 unhandled errors and 3 unexpected successes 13:03:28 (SLEEP NON-CONSING) again 13:04:17 damn thing, i think the best approach would be to have a C function which is called with integers or floats which does calculation into seconds/microsecnds on the C side 13:07:41 how much would it cost to port SBCL to ppc64? 13:08:19 i don't know anything about PPC right now, so, no idea 13:23:05 -!- psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has quit [Quit: Leaving.] 13:29:07 -!- segv- [~mb@95-91-241-45-dynip.superkabel.de] has quit [Remote host closed the connection] 13:31:51 how hard can it be to sleep without consing really? :p 13:32:21 foom: sleeping is easy, computing the amount by which it should sleep isn't 13:32:56 with floats, in particular 13:34:56 pnpuff [~Cijkl@unaffiliated/pnpuff] has joined #sbcl 13:35:21 -!- pnpuff [~Cijkl@unaffiliated/pnpuff] has left #sbcl 13:39:58 segv- [~mb@95-91-241-45-dynip.superkabel.de] has joined #sbcl 13:52:25 is sbcl's ppc assembly messed up or what 13:53:01 i mean registry names etc. 13:53:50 -!- Bike [~Glossina@75-175-67-103.ptld.qwest.net] has quit [Ping timeout: 240 seconds] 13:55:49 Bike [~Glossina@75-175-67-103.ptld.qwest.net] has joined #sbcl 13:58:24 drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has joined #sbcl 13:59:03 I don't think so. 13:59:27 it's really hard to see what's going on 14:01:54 ppc mtlr 14:01:54 Sorry, I couldn't find anything for mtlr. 14:02:04 ppc addi 14:02:05 Add Immediate: http://www.nersc.gov/vendor_docs/ibm/asm/addi.htm#idx262 14:02:19 "Page not found" 14:05:44 -!- rpg [~rpg@216.243.156.16.real-time.com] has quit [Quit: rpg] 14:08:00 -!- benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 14:08:23 -!- edgar-rft [~GOD@HSI-KBW-149-172-63-75.hsi13.kabel-badenwuerttemberg.de] has quit [Quit: eternal darkness] 14:11:55 attila_lendvai [~attila_le@84-236-118-220.pool.digikabel.hu] has joined #sbcl 14:11:55 -!- attila_lendvai [~attila_le@84-236-118-220.pool.digikabel.hu] has quit [Changing host] 14:11:55 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 14:19:46 benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has joined #sbcl 14:20:45 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 14:21:17 LiamH [~none@pdp8.nrl.navy.mil] has joined #sbcl 14:24:18 -!- benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has quit [Ping timeout: 245 seconds] 14:25:05 -!- drmeister [~drmeister@pool-71-185-168-200.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 14:26:12 benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has joined #sbcl 14:31:17 psilord [~pkeller@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 14:31:50 rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 14:48:09 -!- benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has quit [Ping timeout: 264 seconds] 14:53:05 drmeister [~drmeister@166.137.87.193] has joined #sbcl 14:53:09 -!- teggi [~teggi@123.25.129.139] has quit [Remote host closed the connection] 14:53:57 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 248 seconds] 14:58:43 -!- rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Quit: rpg] 15:00:24 benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has joined #sbcl 15:01:15 pranavrc [~pranavrc@122.164.110.200] has joined #sbcl 15:01:15 -!- pranavrc [~pranavrc@122.164.110.200] has quit [Changing host] 15:01:15 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 15:07:18 -!- benkard [~benkard@dhcp-138-246-84-221.dynamic.eduroam.mwn.de] has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz] 15:16:49 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 15:16:49 -!- drmeister [~drmeister@166.137.87.193] has quit [Remote host closed the connection] 15:23:15 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 15:27:46 apparently, sleep consing in this case is actually when calling nanosleep, not time calculations 15:45:10 -!- foom [jknight@nat/google/x-nfwyiltaivbteiti] has quit [Quit: Leaving] 16:21:33 *stassats* is surprised that undefined_tramp on ppc is broken 16:21:53 the register with FDEFN is just encoded wrongly 16:22:04 how could that be not caught before? 16:24:12 iiuc, nyef is surprised that any of the ~x86 ports work at all. 16:24:52 i expect ppc to be the next best 16:25:40 at least now i can test on ppc regularly 16:32:42 .byte 254, sc_DescriptorReg+0x40, 1 could use something better than hard-coding register offsets 16:33:04 that's the current definition, the working one is +0x80, 2 16:33:34 stassats: http://paste.lisp.org/+2YRF 16:33:51 those are the hosts I was using to test libfixposix 16:34:14 i'm on gcc110 right now, all other appear to be down 16:34:26 haven't tried ia64 16:34:31 but they might come back up 16:34:40 anyway, those aliases are handy 16:34:43 yeah, but i got bugs on ppc to fix so far 16:34:47 especially with shell completion 16:34:54 so, i'm occupied 16:34:57 :) 16:35:27 and it's big-endian, so, good for testing bit-fiddling code 16:42:10 -!- ASau` is now known as ASau 16:47:22 oh, i see whom to blame for ppc breakage 16:48:03 pkhuong changed the format from 5 bits for SCN to 6 bits, hence it needs to go from 140 to 280 16:48:17 at least that explain why nobody noticed that 16:52:30 same thing on mips 16:53:50 and probably on sparc 16:55:53 stassats: raw numbers :\ 16:55:58 yep 16:56:15 the proper fix would be to export register offsets, and then calculate from them 16:57:05 mips-assem already tries to calculate, .byte ((14 << 5) + sc_DescriptorReg) % 0x100 16:58:04 registers offsets aren't currently exported from genesis 16:58:27 i'll add a quick fix now, and will note do a proper one later 16:58:51 incidentally, fixing makes only one failure on ppc, (SLEEP NON-CONSING) 16:58:57 and 3 unexpected successes 16:59:03 with threads? 16:59:11 don't think so 16:59:34 no threads 17:02:22 foom [jknight@nat/google/x-zxkmfprfmvinmljn] has joined #sbcl 17:12:30 -!- nicdev [user@2600:3c03::f03c:91ff:fedf:4986] has quit [Ping timeout: 245 seconds] 17:21:34 benkard [~benkard@tmo-106-38.customers.d1-online.com] has joined #sbcl 17:27:25 -!- benkard [~benkard@tmo-106-38.customers.d1-online.com] has quit [Ping timeout: 276 seconds] 17:35:11 benkard [~benkard@tmo-107-133.customers.d1-online.com] has joined #sbcl 18:17:00 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 18:24:07 drmeister [~drmeister@farnsworth.chem.temple.edu] has joined #sbcl 18:27:58 Can I easily recompile parts of SBCL from SLIME? Naively trying to use C-c C-k makes it complain about missing packages (e.g. SB!C). 18:28:25 use C-c C-c instead 18:28:32 benkard: C-c C-c will just work, with slime magic to interpolate sb!foo 18:28:54 if you have literal sb!foo:bar in the source, you'll have to change that, though. 18:29:23 pkhuong: that'll work too 18:29:56 stassats: not always, on my end. 18:30:13 It might only have been. #.sb!foo:bar 18:30:38 Hmm. OK. So I need to compile each form individually, then. 18:30:57 Thanks. 18:31:05 #.sb!foo:bar should work too 18:32:05 benkard: you can also copy/paste into a buffer with (in-package "SB-C"). 18:32:24 stassats: we'll see if that works better now that I'm on a post 2012 SLIME. 18:32:27 benkard: C-c C-c can compile a region 18:32:52 stassats: Oh, it can? I didn't know that. :) That's useful. 18:33:11 pkhuong: for #.sb!foo:bar to work it should be inside an sbcl package 18:33:33 sbcl package defined as (let ((name (package-name package))) (eql (mismatch "SB-" name) 3)) 18:34:00 and that's fairly old 18:37:30 -!- Bike [~Glossina@75-175-67-103.ptld.qwest.net] has quit [Ping timeout: 245 seconds] 18:39:39 Bike [~Glossina@67-5-200-199.ptld.qwest.net] has joined #sbcl 18:39:40 I've made a fork of SBCL that uses my interpreter when (eq sb-ext:*evaluator-mode* :interpret). (Which means I break basic REPL functionality all the time and need to restart it frequently.) 18:40:24 It's at: https://github.com/benkard/sbcl 18:41:12 davazp [~user@31.200.185.16] has joined #sbcl 18:41:52 sdemarre [~serge@91.180.103.212] has joined #sbcl 18:43:04 benkard: sticking to :compile should avoid the evaluator for everything except some foreign function stuff. 18:43:25 and maybe some PCL. 18:44:55 pkhuong: Yes, but I deliberately set it to be able to use C-c C-l to load stuff using the interpreter. 18:45:16 Hmm... Maybe better to use LOAD and bind *evaluator-mode* temporarily. 18:45:31 -!- Bike [~Glossina@67-5-200-199.ptld.qwest.net] has quit [Ping timeout: 264 seconds] 18:46:11 yes, that's more robust. You can also try stuff like (set '*evaluator-mode* :compile), which works as long as function calls work. 18:47:27 Bike [~Glossina@67-5-214-243.ptld.qwest.net] has joined #sbcl 18:50:38 -!- bege [~bege@S0106001d7e5132b0.ed.shawcable.net] has quit [Quit: leaving] 19:00:04 yeah, ppc with threads is no good 19:02:53 Bike_ [~Glossina@67-5-226-71.ptld.qwest.net] has joined #sbcl 19:03:50 -!- Bike [~Glossina@67-5-214-243.ptld.qwest.net] has quit [Ping timeout: 240 seconds] 19:06:59 Anyway, debugging now works a bit. At least it displays local variables. 19:07:30 -!- Bike_ [~Glossina@67-5-226-71.ptld.qwest.net] has quit [Ping timeout: 245 seconds] 19:09:13 Bike [~Glossina@67-5-226-71.ptld.qwest.net] has joined #sbcl 19:09:13 -!- ASau [~user@p4FF96FA0.dip0.t-ipconnect.de] has quit [Ping timeout: 268 seconds] 19:18:47 ASau [~user@p4FF96FA0.dip0.t-ipconnect.de] has joined #sbcl 19:20:25 how about source locations? 19:23:57 in particular, defun-cached and multiple threads are broken 19:24:14 ppc probably breaks some assumptions about atomicity 19:24:15 Sorry, I couldn't find anything for probably breaks some assumptions about atomicity. 19:24:23 I think I'm computing source locations the wrong way. sb!c::make-definition-source-location throws errors. 19:25:50 But I'm not really manipulating them directly. Shouldn't ensure-source-path create valid source paths that make-definition-source-location can take as input? Is there something else that needs doing to get a valid source path? 19:26:22 yea, x86(-64) makes life too easy on developers, so they forget to put in proper barriers. 19:27:44 foom: alternatively, it's the proof that it's possible to make life easy for software devs ;) 19:28:42 the error The value #(# # NIL T) is not of type (SIMPLE-ARRAY T (*)) is strange 19:31:25 benkard: what kind of errors? 19:33:10 jsnell: ensure-source-path creates lists whose (car (last _)) is not an (or fixnum null). 19:33:55 which makes make-definition-source-location fail with a type error because toplevel-form-number is set to a form (I think) rather than a fixnum. 19:36:22 Which I don't understand, because ensure-source-path just uses cons, which shouldn't break that invariant if it's true for *current-path*. 19:42:18 hm. I don't quite understand where you're initiazing the source paths 19:43:41 where is find-source-paths getting called? 19:43:53 I don't... I rely on LOAD to do that. 19:44:37 LOAD-AS-SOURCE calls FIND-SOURCE-PATHS. 19:52:37 indeed, that looks a bit mysterious. not sure what's going on 19:53:25 -!- davazp [~user@31.200.185.16] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 20:03:10 -!- Bike [~Glossina@67-5-226-71.ptld.qwest.net] has quit [Ping timeout: 246 seconds] 20:04:57 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 20:07:07 Bike_ [~Glossina@67-5-226-71.ptld.qwest.net] has joined #sbcl 20:07:21 Hmm. OK. I'll see whether I can figure out what happens. That's a plan for tomorrow, though. 20:08:16 -!- sdemarre [~serge@91.180.103.212] has quit [Ping timeout: 264 seconds] 20:09:38 -!- Bike_ is now known as Bike 20:15:33 francogrex [~user@109.128.109.112] has joined #sbcl 20:17:03 ehaliewicz [~user@50-0-51-11.dsl.static.sonic.net] has joined #sbcl 20:17:15 Hi, how can we represent hex in sbcl VOP? for example (inst move EAX-TN 0013h) ? 20:17:33 the same way you represent it in common lisp 20:18:48 there's no such thing as a hexadecimal number, only the printed representation is different 20:20:52 yes so as a string? (inst mov EAX-TN "6D") ? 20:21:03 clhs #x 20:21:04 http://www.lispworks.com/reference/HyperSpec/Body/02_dhi.htm 20:21:16 a string is not a number, in reality 20:23:07 ok the sharp sign. e.g. #xF00 cool. thanks 20:23:23 *francogrex* goes HEY! 20:29:34 rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has joined #sbcl 20:49:02 -!- |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has quit [Ping timeout: 240 seconds] 20:51:55 |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has joined #sbcl 20:52:40 -!- benkard [~benkard@tmo-107-133.customers.d1-online.com] has quit [Ping timeout: 264 seconds] 21:08:55 -!- Bike [~Glossina@67-5-226-71.ptld.qwest.net] has quit [Ping timeout: 264 seconds] 21:11:09 Bike [~Glossina@67-5-226-71.ptld.qwest.net] has joined #sbcl 21:14:36 -!- francogrex [~user@109.128.109.112] has left #sbcl 21:17:30 -!- Bike [~Glossina@67-5-226-71.ptld.qwest.net] has quit [Ping timeout: 245 seconds] 21:18:41 Bike [~Glossina@67-5-226-71.ptld.qwest.net] has joined #sbcl 21:22:09 shafire [2e05dbbe@gateway/web/freenode/ip.46.5.219.190] has joined #sbcl 21:22:11 hi 21:22:21 will sbcl be ported to arm(64)? 21:22:51 yes 21:23:43 shafire: can you pay for the porting ? 21:24:31 I am only a student at moment. :( 21:25:03 can you help in porting it? 21:26:34 yes 21:28:37 -!- psilord [~pkeller@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Quit: Leaving.] 21:28:42 so, here's the problem with subtypep (and other cached functions) on PPC: http://paste.lisp.org/display/138367 21:29:00 (test 100000 *standard-output*) prints several hits 21:31:01 or even just changing the condition to unless (check x) do (incf hit-number) 21:31:09 (test 100000 *standard-output*) => 36 21:31:50 stassats: can you try a barrier after allocation sequences? 21:31:53 and having just two threads has the same outcome 21:32:47 we're probably vunlerable to really strange memory safetfy bugs without such barriers 21:33:50 *stassats* is really annoyed by labels like L24: L25: 21:34:10 when more than one instruction jumps to the same address, it becomes like this 21:34:14 -!- Bike [~Glossina@67-5-226-71.ptld.qwest.net] has quit [Ping timeout: 240 seconds] 21:34:27 even though L25 is used in the jump instruction itself 21:34:43 or just at the end of every pseudo-atomic. 21:35:49 would that be mbar? 21:35:57 yes. 21:36:03 ok, we don't have it yet 21:36:42 sync 21:36:49 it's coarser grained too, iirc. 21:37:12 mbar is coarser? 21:37:27 sync is. 21:37:49 *stassats* misses intel docs, IBM's are not as easy to read 21:42:32 isn't sync for lwarx and stwcx? 21:45:14 -!- daimrod [daimrod@sbrk.org] has quit [Ping timeout: 268 seconds] 21:47:42 and allocate-vector-on-heap is an assembly routine, can't test that fast 21:47:46 Bike [~Glossina@67-5-226-71.ptld.qwest.net] has joined #sbcl 21:49:20 The PowerPC instruction, sync, provides an ordering function that ensures that all instructions initiated prior to the sync instruction complete, and that no subsequent instructions initiate until after the sync instruction completes. When the sync instruction completes, all storage accesses initiated prior to the sync instruction are complete. 21:53:20 top with 64 threads doesn't fit on my screen when broken done on by threads 21:53:30 hahaha 21:54:43 htop does a better job 21:56:32 64GB of memory is nice too 21:56:33 I have the same issue with 24 threads here. I wound up breaking it up in 4 columns, with htop. 21:56:58 first world problems 21:57:15 -!- rpg [~rpg@23-25-144-217-static.hfc.comcastbusiness.net] has quit [Quit: rpg] 21:57:57 so, turns out ppc64 is just ppc32 with some wider registers, so porting shouldn't be hard 22:02:21 i think i can test it quicker if i just insert some vops around make-array 22:02:30 with primitive 22:03:51 this POWER7 is still slower than my laptop 22:04:11 maybe because of other processes 22:10:54 no need for vop. we already have membar functions 22:11:12 (%memory-barrier) 22:11:58 or (sb-thread:barrier (:memory) (make-array ...)) 22:13:19 that gives me 0 22:14:08 and it uses SYNC 22:14:24 so, success? 22:14:40 but there are different flavors of SYNC, then there's MSYNC, MBAR 22:14:45 each with different parameters 22:14:48 pkhuong: right 22:15:00 we only have the heavyweight one, iirc. 22:15:34 the finer grained variants and instructions are recent extensions. 22:15:50 and which ppc are we targetting? 22:15:56 no idea. 22:16:14 Bike_ [~Glossina@75-175-73-51.ptld.qwest.net] has joined #sbcl 22:16:16 -!- shafire [2e05dbbe@gateway/web/freenode/ip.46.5.219.190] has quit [Quit: Page closed] 22:16:51 -!- Bike [~Glossina@67-5-226-71.ptld.qwest.net] has quit [Disconnected by services] 22:16:56 -!- Bike_ is now known as Bike 22:17:13 so, adding sync after the body of pseudo-atomic? 22:17:53 I'd say at the end of pseudoatomic. 22:18:47 i imagine that wouldn't be the only problem with ordering 22:18:50 I don't think it matters if we do it before or after leaving PA, but before is likely safer. 22:19:36 probably not, but the criterion for PA and for a publish fence are similar, so it'll take of a few. 22:21:03 -!- milosn_ [~milosn@user-5af5015d.broadband.tesco.net] has quit [Remote host closed the connection] 22:22:57 Bike_ [~Glossina@174-25-37-243.ptld.qwest.net] has joined #sbcl 22:24:35 -!- Bike [~Glossina@75-175-73-51.ptld.qwest.net] has quit [Ping timeout: 245 seconds] 22:24:52 trying sync after ,@forms 22:25:25 milosn [~milosn@user-5af5015d.broadband.tesco.net] has joined #sbcl 22:25:32 -!- Bike_ [~Glossina@174-25-37-243.ptld.qwest.net] has quit [Client Quit] 22:25:50 Bike [~Glossina@174-25-37-243.ptld.qwest.net] has joined #sbcl 22:27:34 *stassats* is not ready for such ordering after x86 22:28:01 -!- LiamH [~none@pdp8.nrl.navy.mil] has quit [Quit: Leaving.] 22:28:12 too used to "i don't know what it is, but it works, so it's alright" 22:34:35 that works as well 22:39:58 thread.impure now not only finishes, but finishes successfully 22:40:36 but still can't run threaded slime 22:40:52 stassats: the issues might be in slime ;) 22:41:24 right at connecting and Memory fault at 10f (pc=0x4f0d779c, sp=(nil)), unlikely 22:41:30 straight to ldb 22:41:45 and fatal error encountered in SBCL pid 38771(tid 4123128976): gc signals blocked 22:42:09 backtrace might be interesting 22:43:13 it's full of question marks 22:43:45 attach in gdb to get the remainder. 22:43:54 http://paste.lisp.org/display/138367#1 22:47:33 looks like something after getaddrinfo 22:48:48 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 22:51:34 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 22:56:52 -!- drmeister [~drmeister@farnsworth.chem.temple.edu] has quit [Remote host closed the connection] 22:57:39 (sb-thread:make-thread (lambda () (print (sb-bsd-sockets:host-ent-addresses (sb-bsd-sockets:get-host-by-name "localhost"))) (finish-output))) 22:58:54 simplified: (sb-thread:make-thread #'sb-bsd-sockets:get-host-by-name :arguments '("localhost")) 22:59:34 is that a #+sb-bsd-sockets-addrinfo build? 22:59:50 yep 23:01:09 maybe bad variable types 23:01:39 but, it works on the main thread 23:04:38 long shot, but can you try wrapping getaddrinfo in a barrier? 23:05:10 so, on the one hand, this points to craziness in libc. 23:05:13 with the sync, now only (SLEEP NON-CONSING) fails on threads too 23:05:27 and some more unexpected successes 23:06:01 don't know how many failed before sync, since it couldn't finish the tests 23:06:25 on the other, i'd almost expect that kind of decision from drepper: "posix says it's not thread-safe. derp." 23:08:17 wrapping sb-thread:barrier around getaddrinfo didn't help 23:09:06 http://sourceware.org/bugzilla/show_bug.cgi?id=13271 ha-ha 23:09:12 "Just go away, you really know nothing. Guess who of the two of us is in the POSIX committee?" 23:09:14 guess the author 23:10:02 probably not related, just funny 23:10:19 I know the author ;) 23:10:36 *who the author is, that is. 23:18:44 maybe call_into_c is problematic 23:19:16 but everything happens in a single thread, so i don't expect any ordering issues 23:20:52 davazp [~user@92.251.185.254.threembb.ie] has joined #sbcl 23:24:08 another long shot is to try and see if disabling GC helps. 23:24:57 nope 23:32:15 if you have time, you can try and rip out some of the logic to figure out where exactly it fails. 23:33:36 well, i'm in no hurry, if all else fails, eventually i'll be resorting to re-compiling glibc and looking at what getaddrinfo does 23:35:59 puny bugs can't scare me 23:38:36 10f really looks like an offset from NULL. Maybe you can look at the grovelled offsets or the layouts and work back from there. 23:39:15 what makes it strange is that it works from the main thread, and works correctly 23:39:25 so the layout should really be correct as well then 23:40:01 sure. 23:40:46 -!- segv- [~mb@95-91-241-45-dynip.superkabel.de] has quit [Remote host closed the connection] 23:42:17 well, there are differences between the main thread and not 23:42:20 like stack location 23:42:40 maybe in one case it's filled with 0, and in other it's something else 23:42:50 which could mask a problem with layouts 23:42:59 I don't think it's a layout issue 23:43:20 although we might be getting bitten by an unexpected uninitialised slot 23:44:59 ruling out things never hurts 23:55:16 edgar-rft [~GOD@HSI-KBW-149-172-63-75.hsi13.kabel-badenwuerttemberg.de] has joined #sbcl