00:04:55 dto [~user@pool-100-0-107-54.bstnma.fios.verizon.net] has joined #sbcl 00:29:18 -!- hargettp [~hargettp@c-76-119-232-173.hsd1.ma.comcast.net] has quit [Quit: Linkinus - http://linkinus.com] 00:34:55 ASau` [~user@p54AFF123.dip0.t-ipconnect.de] has joined #sbcl 00:38:11 -!- ASau [~user@p54AFE9E8.dip0.t-ipconnect.de] has quit [Ping timeout: 252 seconds] 01:23:55 psilord [~psilord@c-69-180-173-249.hsd1.mn.comcast.net] has joined #sbcl 01:31:40 -!- macdice [~user@46-65-10-191.zone16.bethere.co.uk] has left #sbcl 01:31:56 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: happening terminated by computer explosion] 01:41:38 hi psilord1 01:41:44 and psilord 01:55:06 i wrote my first :around methods the other day. i'll be very careful, though :) 02:26:44 -!- ASau` is now known as ASau 02:52:34 Krystof: re shark not being fish, the example might come from a (once well known, I think) predator/prey simulation with shark feeding on fish. 03:19:52 yacks [~py@103.6.159.103] has joined #sbcl 03:38:38 -!- christoph_debian [~christoph@ppp-88-217-35-202.dynamic.mnet-online.de] has quit [Ping timeout: 264 seconds] 03:46:04 -!- Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has quit [Ping timeout: 252 seconds] 03:50:31 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 03:52:06 christoph_debian [~christoph@ppp-88-217-48-119.dynamic.mnet-online.de] has joined #sbcl 03:55:41 pranavrc [~pranavrc@122.164.159.83] has joined #sbcl 03:55:41 -!- pranavrc [~pranavrc@122.164.159.83] has quit [Changing host] 03:55:41 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 04:05:49 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 04:25:00 dtp: Hello. 04:32:38 hello dto. I'm afk for a bit though. :| 04:32:45 okay. 04:32:47 hi :) 05:06:02 andreh [~andreh@201.22.190.64.dynamic.adsl.gvt.net.br] has joined #sbcl 05:06:31 -!- andreh [~andreh@201.22.190.64.dynamic.adsl.gvt.net.br] has quit [Client Quit] 05:09:31 andreh [~andreh@201.22.190.64.dynamic.adsl.gvt.net.br] has joined #sbcl 05:18:34 edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has joined #sbcl 06:10:54 -!- LiamH [~none@96.231.217.60] has quit [Quit: Leaving.] 06:26:53 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 06:45:31 -!- Quadrescence [~quad@unaffiliated/quadrescence] has quit [Quit: This computer has gone to sleep] 07:24:18 danlentz_ [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has joined #sbcl 07:25:07 -!- danlentz_ [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has quit [Remote host closed the connection] 07:25:54 -!- Posterdati [~kvirc@host208-231-dynamic.2-79-r.retail.telecomitalia.it] has quit [Ping timeout: 264 seconds] 07:26:03 -!- danlentz [~danlentz@c-68-37-70-235.hsd1.nj.comcast.net] has quit [Ping timeout: 264 seconds] 07:26:39 Posterdati [~kvirc@host208-231-dynamic.2-79-r.retail.telecomitalia.it] has joined #sbcl 07:32:42 -!- dto [~user@pool-100-0-107-54.bstnma.fios.verizon.net] has quit [Remote host closed the connection] 07:47:50 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 07:58:42 jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has joined #sbcl 08:18:34 pnpuff [~harmonic@unaffiliated/pnpuff] has joined #sbcl 08:20:27 -!- yacks [~py@103.6.159.103] has quit [Read error: Operation timed out] 08:21:54 yacks [~py@103.6.159.103] has joined #sbcl 09:36:45 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 09:41:29 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 252 seconds] 09:55:18 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Read error: Connection reset by peer] 10:04:19 sdemarre [~serge@91.176.254.125] has joined #sbcl 10:28:11 -!- oleo [~oleo@xdsl-87-79-255-130.netcologne.de] has quit [Ping timeout: 260 seconds] 10:28:43 oleo [~oleo@xdsl-78-35-175-183.netcologne.de] has joined #sbcl 10:37:41 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 252 seconds] 10:40:56 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 10:50:47 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: mental process ended into paranoid blackout] 11:25:41 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 11:30:25 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 272 seconds] 12:20:37 -!- pnpuff [~harmonic@unaffiliated/pnpuff] has quit [] 12:27:11 -!- milosn [~milosn@94.12.79.143] has quit [Ping timeout: 260 seconds] 12:27:57 -!- milosn_ [~milosn@94.12.79.143] has quit [Ping timeout: 272 seconds] 12:43:44 fiveop [~fiveop@p5DDC4465.dip0.t-ipconnect.de] has joined #sbcl 12:45:51 when taking code from MIPS, I always forget that is has a branch delay slot 12:47:43 milosn [~milosn@94.12.79.143] has joined #sbcl 12:56:33 -!- foom [~jknight@2620:15c:6:14:be30:5bff:fedf:6db6] has quit [Ping timeout: 272 seconds] 13:03:24 -!- andreh [~andreh@201.22.190.64.dynamic.adsl.gvt.net.br] has quit [Quit: Quitte] 13:14:11 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 13:18:50 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 252 seconds] 13:29:11 could be worse. you could forget that your port target has them ;) 13:29:41 -!- sdemarre [~serge@91.176.254.125] has quit [Ping timeout: 248 seconds] 13:32:11 in the first iteration of the ppc port, gbyers had carefully annotated all the branch instructions with their latency in the :delay slot 13:38:37 wow. 14:12:46 LiamH [~none@96.231.217.60] has joined #sbcl 14:21:46 nyef [~nyef@pool-108-7-220-91.bstnma.fios.verizon.net] has joined #sbcl 14:21:55 G'morning all. 14:36:39 hi nyef 14:38:21 At some point, I hope to have the project bandwidth to at least sit down and assess what still needs doing on SBCL/ARM at least in terms of "bulk lifting". Things like the floating-point support, static-functions, full-call assembly-routines, and so on. But right now I'm swamped. /-: 15:02:58 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 15:07:55 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 272 seconds] 15:18:48 stassats [~stassats@pppoe.178-66-85-194.dynamic.avangarddsl.ru] has joined #sbcl 15:18:49 -!- stassats [~stassats@pppoe.178-66-85-194.dynamic.avangarddsl.ru] has quit [Changing host] 15:18:49 stassats [~stassats@wikipedia/stassats] has joined #sbcl 15:30:00 sdemarre [~serge@91.176.254.125] has joined #sbcl 15:31:02 leuler [~user@p548FA3D6.dip0.t-ipconnect.de] has joined #sbcl 15:41:13 -!- slyrus [~chatzilla@107.201.5.56] has quit [Ping timeout: 272 seconds] 15:43:42 -!- stassats [~stassats@wikipedia/stassats] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 15:56:05 slyrus [~chatzilla@107.201.5.56] has joined #sbcl 16:06:21 nyef: I continue to go through build-order and active files 16:06:55 there are a few things I can not fix without understanding a lot more about how things work, I write comments about each of those cases to the files in build-order 16:07:20 Stuff like, fixing ':frob-nfp' in define-full-call for the tail case :) 16:12:31 fiveop: is your branch published somewhere? 16:12:40 Hunh. I thought I'd done most of the NFP stuff already? 16:13:14 Ah, guess not. 16:13:15 Krystof: fiveop/sbcl-arm on github 16:13:32 the stuff of the last two days isn't however 16:15:01 fiveop: frob-nfp should be easy. Look at the PPC port, see what it does, find another instance of it, look that up in the ARM port, and see what I did. 16:15:29 You'll find an example in KNOWN-RETURN. 16:16:43 -!- sdemarre [~serge@91.176.254.125] has quit [Ping timeout: 260 seconds] 16:18:13 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 16:20:41 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 16:48:01 pnpuff [~harmonic@unaffiliated/pnpuff] has joined #sbcl 16:56:19 I'll do that 16:56:31 That might fix lots of files :) 16:58:19 sdemarre [~serge@91.176.254.125] has joined #sbcl 17:27:42 jdz [~jdz@212.36.34.246] has joined #sbcl 17:28:07 -!- jdz is now known as Guest35956 17:32:58 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 17:53:31 Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has joined #sbcl 17:54:01 can I deactivate style-warnings for the whole host-2 process? 17:56:11 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 260 seconds] 17:57:11 Probably, but I'm not quite sure how. Have a look in src/cold/ for anything in the way of a "target-*-stem" function that looks suitable. 17:59:33 Umm... I'm looking at your sbcl-memory-layout repository, and... have you read doc/internals/objects-in-memory.texinfo yet? 18:00:55 The ALPHA backend runs on a 64-bit machine, but the heap is forced to the low two gigs or so and heap words are 32-bit. 18:02:22 no I haven't 18:03:26 src/code/room might be an interesting read as well, though you might want to read a modern version, not the one in the arm-port tree. 18:04:44 I've checked out the latest version as well 18:05:13 Also see is_lisp_pointer and is_lisp_immediate in src/runtime/runtime.h for more lowtag/widetag games. 18:10:22 n-lowtag-bits is defined based on the constants then available to be the number of address bits that must be zero for aligned access to an "allocation unit", which is defined to be double-word aligned. 18:10:38 I'm not quite sure how best to express that, TBH. 18:12:40 but that explains the formula 18:13:53 so essentially. n-lowtag-bits is defined to be as large as possible, when "allocations units" are assumed to be aligned to two words 18:14:08 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 18:14:29 Yup. 18:15:07 There was a proposal for 31-bit fixnums on 32-bit platforms, but it required overlapping two pointer lowtags and doubling the allocation granularity for non-CONS objects. 18:15:14 What I wonder, now that I understand your fixnum change in 64 bit, what did one get from having 4 bit lowtags on 64 bit platforms before that change? 18:15:47 Not much? 18:16:03 nothing? 18:16:25 Umm... I think the answer is "spare lowtags". /-: 18:16:35 Oh, wait. I know. 18:16:44 It's the NIL/CONS thing again. 18:17:13 I don't know how nil is represented so far 18:17:31 that one's an interesting hack. 18:17:31 list-pointer-lowtag and other-pointer-lowtag have to differ by the address-delta for a word. 18:17:35 is a CONS a word with a list-pointer-lowtag 18:17:53 Have a look at objects-in-memory, it has a writup for NIL/CONS. 18:18:02 I'll read that 18:18:08 when I get stuck again :) 18:18:23 with frob-nfp fixed for tail calls, there is progress to be made again :) 18:18:37 Okay, have fun. 18:19:17 unix-open, no SAP-reg available 18:19:32 VOP? 18:20:51 pack-load-tns: start-vop ALLOC-NUMBER-STACK-SPACE, last-vop FAST-IF-<-C/SIGNED 18:20:59 I had that before in misc-alien 18:21:12 the last-vop was different (dealloc-number-stack-space or something similar) 18:21:19 (and haven't fixed it) 18:24:03 Oh, lovely. 18:24:37 ... Are you keeping a log, like I've been doing for my work on this, btw? 18:26:49 I have my commit messages 18:26:53 and notes in build-order 18:27:36 So... not the same thing at all? 18:27:40 not really 18:27:47 Okay, was just wondering. 18:28:10 I don't note from which arch I take the 'inspiration' for my vops 18:28:12 +down 18:28:56 Which is pretty much everything I did so far (fixed some bugs in inst usage as well) 18:29:58 (But when writing (inst eor temp temp temp) or similar, I admit, it's easy to forget a register :)) 18:30:10 One of the things that I was trying to do with the arm-port was to lay out a history of what it takes to write a new backend, what the design considerations are, workable approaches, and so on. Basically, provide a roadmap for how to port to yet another CPU type. 18:31:38 But then, as I said: All I did so far is porting vops from other arches. 18:32:22 What might be interesting is a list of all vops that have to be provided in compiler/target, maybe separated by file they should appear in 18:49:15 -!- pnpuff [~harmonic@unaffiliated/pnpuff] has quit [] 19:03:46 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 19:22:25 Quadrescence [~quad@unaffiliated/quadrescence] has joined #sbcl 19:27:09 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Remote host closed the connection] 19:33:19 poglesbyg_ [~poglesbyg@80.202.97.213] has joined #sbcl 19:35:37 can I put unsigned-num stuff into any other registers, when I run out of unsigned-reg registers? 19:37:02 fiveop: bad idea. 19:37:11 so I need to rewrite 19:37:24 unless you know for sure it's suitably aligned/tagged to look like non-pointer data. 19:37:41 you can push stuff on the number stack. 19:38:02 Pushing stuff on the number stack is painful, though. 19:38:06 non-pointer data means fixnum or other-immediate? 19:38:35 Basically, yeah. 19:38:44 So there are limits to what you can do. 19:39:59 -!- poglesbyg_ [~poglesbyg@80.202.97.213] has quit [Quit: Textual IRC Client: www.textualapp.com] 19:40:13 bignum-floor uses 7 unsigned-regs 19:40:34 Now I have to actually understand what's happening there in order to fix it. 19:40:35 We only have five. 19:41:09 Which file is this in? 19:41:42 Ah, arith. 19:42:20 Which backend did you copy from, and have you compared other possible backends for a lighter version? 19:42:47 that's an optional VOP if you don't have native 64x32 divide 19:43:49 or not, never mind. 19:44:09 just a useless feature 19:44:21 nyef: took it from MIPS (as always) and will look at others now 19:52:02 fiveop: seems to do it in 6 registers (xh is in/out) 19:53:20 GPL code, so careful ;) But it looks like russian peasant division. 19:53:48 Yeah, if you look out our VOPs, they also operate in terms of six registers. 19:53:59 s/out/at/ 19:54:13 Unless, of course, there's a machine instruction for it. 19:55:00 Wait... fiveop: Check out the SPARC version of BIGNUM-FLOOR. 19:55:35 that code would also be an obvious candidate for assembly routine... except that it's only used once ;) 19:55:43 Heh. 19:55:57 Oh, that's ano... wait, no it isn't. Hrm. 19:56:40 I was thinking that we could just set P-A and just clear the boxed registers that we use in the end, but we need to hold the pa-flag. 19:56:58 Still workable, actually, since we can arrange for the lifetimes not to overlap. 19:57:11 (JFTR, the pa-flag is required to be a specific unboxed register.) 19:58:46 fiveop: And the MIPS version looks to be using TEMP as a condition flag. 20:01:31 nyef: I'm looking at the sparc version right now 20:01:42 currently pondering how to realise addxcc in arm :) 20:02:05 oh wait 20:03:09 sparc version should work 20:03:10 you want sub/sbc. 20:03:18 addxcc is adcs 20:06:27 Once you get rid of TEMP, you should have enough registers. 20:07:24 I used the sparc version. It uses only five registers 20:07:25 . 20:18:25 -!- jewel [~jewel@105-237-68-160.access.mtnbusiness.co.za] has quit [Ping timeout: 246 seconds] 20:21:57 pnpuff [~harmonic@unaffiliated/pnpuff] has joined #sbcl 20:41:32 Are there anywhere expectations that carry flags or similar are carried over VOP boundaries? 20:43:48 not by default. 20:45:02 You can use the new form of conditionals if you can make some flags carry over between VOPs that only compute the condition and others that jump/select values. 20:45:18 but that's off by default and always opt-in. 21:22:37 -!- pnpuff [~harmonic@unaffiliated/pnpuff] has quit [Remote host closed the connection] 21:28:08 -!- leuler [~user@p548FA3D6.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 21:54:26 davazp [~user@14.Red-79-152-116.dynamicIP.rima-tde.net] has joined #sbcl 21:55:01 enough for today 21:56:07 nyef: in the next few days, I'll tidy up my changes, commit them and write a list of notes, why I ported certain vops 21:57:48 Cool. I'll look forward to it. 21:59:34 reached target-format 21:59:50 (though I have a lot of #+(or)'s in src/code dealing with floats) 22:00:30 -!- fiveop [~fiveop@p5DDC4465.dip0.t-ipconnect.de] has quit [Quit: humhum] 22:00:56 prxq [~mommer@x2f6a1c1.dyn.telefonica.de] has joined #sbcl 22:46:07 -!- sdemarre [~serge@91.176.254.125] has quit [Ping timeout: 246 seconds] 23:03:15 -!- oleo [~oleo@xdsl-78-35-175-183.netcologne.de] has quit [Ping timeout: 260 seconds] 23:03:52 oleo [~oleo@xdsl-78-35-139-202.netcologne.de] has joined #sbcl