01:08:28 -!- hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has quit [Quit: hargettp] 02:46:34 moocow [~new@poco208-2.fredcanhelp.com] has joined #sbcl 02:46:50 -!- holycow [~new@poco208-2.fredcanhelp.com] has quit [Ping timeout: 264 seconds] 02:50:09 -!- moocow is now known as holycow 02:57:02 moocow [~new@poco208-2.fredcanhelp.com] has joined #sbcl 02:57:08 -!- holycow [~new@poco208-2.fredcanhelp.com] has quit [Ping timeout: 258 seconds] 03:06:28 holycow [~new@poco208-2.fredcanhelp.com] has joined #sbcl 03:07:41 -!- moocow [~new@poco208-2.fredcanhelp.com] has quit [Ping timeout: 276 seconds] 03:12:42 -!- holycow [~new@poco208-2.fredcanhelp.com] has quit [Ping timeout: 240 seconds] 03:14:31 holycow [~new@poco208-2.fredcanhelp.com] has joined #sbcl 04:04:55 -!- rbarraud [~rbarraud@118-92-10-101.dsl.dyn.ihug.co.nz] has quit [Read error: Connection reset by peer] 04:13:50 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 264 seconds] 04:14:14 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 05:54:28 rbarraud [~rbarraud@118-92-10-101.dsl.dyn.ihug.co.nz] has joined #sbcl 06:49:18 -!- slyrus [~chatzilla@adsl-75-36-215-204.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 245 seconds] 06:56:21 slyrus [~chatzilla@adsl-75-36-215-204.dsl.pltn13.sbcglobal.net] has joined #sbcl 07:51:07 -!- Krystof [~csr21@84-51-132-95.christ977.adsl.metronet.co.uk] has quit [Ping timeout: 240 seconds] 07:57:04 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 07:57:04 -!- ChanServ has set mode +o nikodemus 07:57:48 morning 08:40:56 morning 08:44:39 -!- rbarraud [~rbarraud@118-92-10-101.dsl.dyn.ihug.co.nz] has quit [Read error: Connection reset by peer] 08:48:01 rbarraud [~rbarraud@118-92-134-38.dsl.dyn.ihug.co.nz] has joined #sbcl 08:51:18 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 240 seconds] 08:52:26 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 09:32:27 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 240 seconds] 09:33:30 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 10:07:21 hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has joined #sbcl 10:12:29 Krystof [~csr21@howells.doc.gold.ac.uk] has joined #sbcl 10:12:29 -!- ChanServ has set mode +o Krystof 10:34:15 udzinari [~user@nat/ibm/x-uwvyhjojremwleaa] has joined #sbcl 10:35:37 -!- hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has quit [Quit: hargettp] 10:37:45 hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has joined #sbcl 10:58:03 -!- rbarraud [~rbarraud@118-92-134-38.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 245 seconds] 11:19:16 froydnj [~froydnj@gateway.codesourcery.com] has joined #sbcl 11:31:33 what's a fast way to create a bignum filled with numbers in certain ranges? (setf (ldb ) ) compiles down to calls to %dpb 11:33:30 -!- hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has quit [Quit: hargettp] 11:33:54 also, seems to me that bignum is a fine way to store a couple of words, judging from objdef.lisp it should consume less memory than a specialzed vector, shouldn't it? 11:36:19 if it's literally a couple: no, the same amount of memory (though one less active word) because of object alignment 11:36:31 if you want to store an odd number of words, you can save a double word 11:37:36 srcaddr, srcport, dstaddr, dstport 11:37:49 where each addr is a u32, and each port a u16 11:38:30 making it 3words on 32bit, although our servers are mostly 64bit 11:39:50 we need those to pluck into a hash-table, so far I've been using a vector specialized to u32, and some quickly-written hash functions on top of those 11:51:34 attila_lendvai [~attila_le@4d6f5d3b.adsl.enternet.hu] has joined #sbcl 11:54:12 hargettp [~anonymous@pool-71-184-181-149.bstnma.east.verizon.net] has joined #sbcl 13:16:06 stassats [~stassats@wikipedia/stassats] has joined #sbcl 13:25:54 -!- attila_lendvai [~attila_le@4d6f5d3b.adsl.enternet.hu] has quit [Ping timeout: 240 seconds] 13:30:07 attila_lendvai [~attila_le@4d6f5d3b.adsl.enternet.hu] has joined #sbcl 14:14:50 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 272 seconds] 14:27:53 -!- attila_lendvai [~attila_le@4d6f5d3b.adsl.enternet.hu] has quit [Ping timeout: 252 seconds] 14:30:31 down to 19 New bugs on launchpad \o/ 14:44:54 nyef [~nyef@pool-71-255-129-229.cncdnh.east.myfairpoint.net] has joined #sbcl 14:45:02 Hello all. 14:45:31 Does anyone know offhand how wide an unboxed complex-single-float is on x86-64? 14:46:10 on-stack? 14:46:13 *nyef* hopes the answer is 8 octets. 14:46:20 Yeah. 14:46:33 yeah, 2x4 bytes 14:46:46 I'm just trying to validate a fix for a CXF-arg brokenness. 14:48:42 ... And it looks good, so now I "just" need to rebase it and commit. 14:51:59 hi nyef! 14:52:08 very cool work on ppc threads! 14:53:03 Thanks. 14:56:08 I guess the next challenge in that direction is either an arm-port, or enabling threading on something else like sparc or mips. 15:02:06 -!- udzinari [~user@nat/ibm/x-uwvyhjojremwleaa] has quit [Remote host closed the connection] 15:04:35 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Ping timeout: 255 seconds] 15:09:17 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 15:09:17 -!- ChanServ has set mode +o nikodemus 15:19:42 I wish there were later darwin builds... it's at 1.0.29/30 now.. 15:20:06 Is it possible to contribute to this process? 16:08:17 nikodemus: your one-liner for 1.0.42.17 is a little odd 16:10:16 froydnj: just committed the follow-up 16:11:28 deepfire: yes. come freeze time, test. after release, build & sign a binary 16:12:38 (though there is a preference for binaries built by gpg connected people or developers, for obvious reasons) 16:18:08 ... Damnit, every time I see OpenBSD and FP exceptions I think "but OpenBSD doesn't -do- FP exceptions", and then it takes me a second to realize that it's for amd64, which is very different from the PPC situation. 16:20:44 -!- foom [~jknight@ita4fw1.itasoftware.com] has quit [Remote host closed the connection] 16:28:50 hey, i hadn't noticed: lp has a new status available: "Opinion" :) 16:40:52 christoph_debian [christoph@sf-ogame.de] has joined #sbcl 17:02:59 -!- Krystof [~csr21@howells.doc.gold.ac.uk] has quit [Ping timeout: 252 seconds] 17:24:12 foom [~jknight@ita4fw1.itasoftware.com] has joined #sbcl 18:17:52 ok, down to 12 New bugs -- i'm done for the day 18:18:18 -!- lnostdal_ [~quassel@222.80-202-49.nextgentel.com] has quit [Ping timeout: 240 seconds] 18:19:00 (i'm still hoping someone else will tackle the floating point format patches...) 18:25:20 There's no specialized 64bit arithmetics on a 32bit system? 18:26:05 no. lichtblau hacked something together a while back 18:27:52 regpair 18:29:39 http://repo.or.cz/w/sbcl/lichteblau.git/tree/761f496c0c1af016ab3fdf590a809bece9dbc07d:/contrib/sb-regpair 18:29:44 (defun u64+ (x y) 18:29:44 (declare (type (unsigned-byte 64) x y)) 18:29:44 (declare (optimize speed)) 18:29:44 (declare (sb-ext:muffle-conditions sb-ext:compiler-note)) 18:29:44 (+ x y)) 18:29:59 What am I missing that I still see notes? 18:30:39 which notes? 18:30:59 forced to do GENERIC-< (cost 10) 18:32:34 if you change it to -byte 65 you'll see them on x86-64 as well 18:32:43 i don't know 18:36:50 placing a locally muffle-conditions around it makes it go away 18:37:09 Does it generate good code, though? 18:37:27 Oh, register pairing on an already starved architecture.. I guess that proper activation heuristics are not trivial to find.. 18:37:48 nyef: as it's generic-<, it might come from arg-parsing code, my suspicion anyway 18:37:58 how can I disassemble the arg parsing code? 18:38:15 Use sb-disassem:disassemble-code-component, or produce a trace-file and look in that. 18:38:17 tcr, try lowering safety? 18:39:08 I'll leave that to you guys :-) 18:41:51 I've actually been seeing some code that looks really bad, but I'm not sure that it truly is... or how to tell the compiler to make it better. 18:43:08 tcr, I'm trying to compile-file the snippet you pasted and 1.0.41.56 doesn't give me notes 18:43:12 nyef: like what? 18:43:43 It's a case of a (loop for i below 10 do ...) that maintains I in a register, but spills and reloads it for a funcall in the loop... 18:43:56 deepfire: you're on a 32bit system? 18:44:00 doh 18:44:08 tcr, oh, sorry! 18:45:47 Now, if I is unreferenced beyond the iteration control itself, it becomes cheaper to invert it (from 10 above 0), and then you can store it on the stack and hit it directly with sub [rbp-N], (fixnumize 1) / jnz. 18:46:24 Another fun one I noticed is that setting up a function call frame appears to involve moving RBP to a temp reg, then storing the temp reg to the new frame. 18:47:00 I wonder if there's a vop boundary for the RBP issue 18:47:45 It's more that there was a VOP for obtaining the frame pointer, and then it could be MOVE-ARG'ed into the new frame. 18:48:55 Yeah, I'm seeing VOP CURRENT-FP => t29[RBX], four more VOPs, then VOP MOVE-ARG t29[RBX] t31[RAX] => t35[S1]. 18:49:17 A peephole pass would kill that no problem, but it shouldn't even come up, you know? 18:50:04 Actually, I've got an idea on how to kill that one. 18:50:19 A bit hacky, though. 19:00:58 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: Leaving] 19:14:27 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 19:14:27 -!- ChanServ has set mode +o nikodemus 19:15:02 Have they invented paybeer (beerpal?) yet so that we can all send nikodemus a pint? 19:19:11 lnostdal [~quassel@56.84-48-233.nextgentel.com] has joined #sbcl 19:29:08 -!- cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has quit [Ping timeout: 260 seconds] 19:29:49 cmm [~cmm@bzq-79-181-203-193.red.bezeqint.net] has joined #sbcl 20:01:07 Krystof [~csr21@84-51-132-95.christ977.adsl.metronet.co.uk] has joined #sbcl 20:01:07 -!- ChanServ has set mode +o Krystof 21:18:28 -!- lnostdal [~quassel@56.84-48-233.nextgentel.com] has quit [Ping timeout: 245 seconds] 21:19:12 -!- tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has quit [Quit: Leaving.] 21:19:39 tcr [~tcr@81-233-246-97-no37.tbcn.telia.com] has joined #sbcl 21:49:53 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Ping timeout: 258 seconds] 21:53:56 what is register pairing 21:55:12 oh new vops 21:55:16 hm 21:58:29 -!- The_Jon_Smith [~The_Jon_S@ip24-250-13-137.ri.ri.cox.net] has left #sbcl 22:01:56 *nyef* sighs. 22:02:25 My SBCL hacking environments are desynced. 22:02:49 Only one of them has the prototype useful git->cvs export script. 22:34:12 -!- foom [~jknight@ita4fw1.itasoftware.com] has quit [Remote host closed the connection] 22:53:05 attila_lendvai [~attila_le@4d6f5d3b.adsl.enternet.hu] has joined #sbcl 23:14:28 rbarraud [~rbarraud@202-180-88-252.callplus.net.nz] has joined #sbcl 23:14:54 -!- attila_lendvai [~attila_le@4d6f5d3b.adsl.enternet.hu] has quit [Quit: Leaving.] 23:21:54 foom [~jknight@ita4fw1.itasoftware.com] has joined #sbcl