00:08:15 -!- drmeiste_ is now known as drmeister 00:35:00 Is there some easy way to find the index of the lowest set bit in an integer? 00:36:29 in lisp? 00:37:31 In general, but my specific use case is in the middle of SBCL. 00:38:03 Looks like (integer-length (logxor x (logand x (1- x)))) does the trick. 00:38:07 it's ctz, which has an x86 instruction 00:38:40 i think (i-l (logand x (- x))) also works? 00:39:23 Good call, thank you. 00:39:37 https://en.wikipedia.org/wiki/Find_first_set#Properties_and_relations 00:48:33 Now to see if the compiler builds once more, and then to see how far host-2 gets. 00:48:45 -!- segv- [~mb@cpeB-151.mvcable.net] has quit [Ping timeout: 276 seconds] 00:59:02 Ugh. I'm finding myself rather disliking DO. 01:03:02 -!- ltbarcly_ [~textual@li94-204.members.linode.com] has quit [Quit: Computer has gone to sleep.] 01:23:41 -!- echo-are` [~user@114.254.107.157] has quit [Read error: Connection reset by peer] 02:04:25 -!- scymtym_ [~user@ip-5-147-115-29.unitymediagroup.de] has quit [Ping timeout: 272 seconds] 02:05:26 -!- brucem_ is now known as brucem 02:32:28 If I kill the -C SAP-REF VOPs, will that break anything, or just make the system a touch slower? 02:44:48 ... Judging by the VOPs generated by this macro, I'm going to go with "make the system a touch slower". 02:49:41 echo-area [~user@182.92.247.2] has joined #sbcl 02:50:51 prxq_ [~mommer@x2f64bcf.dyn.telefonica.de] has joined #sbcl 02:54:15 -!- prxq [~mommer@x2f6bfa8.dyn.telefonica.de] has quit [Ping timeout: 260 seconds] 02:59:21 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 03:02:03 echo-area [~user@182.92.247.2] has joined #sbcl 03:18:27 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 03:20:05 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 03:39:09 -!- christoph_debian [~christoph@ppp-46-244-239-69.dynamic.mnet-online.de] has quit [Ping timeout: 252 seconds] 03:40:13 -!- nyef [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 03:47:17 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 03:52:20 christoph_debian [~christoph@ppp-88-217-39-192.dynamic.mnet-online.de] has joined #sbcl 04:08:02 kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has joined #sbcl 04:21:56 yacks [~py@103.6.159.103] has joined #sbcl 04:59:13 -!- oleo [~oleo@xdsl-87-79-195-127.netcologne.de] has quit [Ping timeout: 265 seconds] 04:59:20 oleo [~oleo@xdsl-78-35-186-8.netcologne.de] has joined #sbcl 05:29:06 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 05:36:56 -!- LiamH [~none@pool-173-73-123-21.washdc.east.verizon.net] has quit [Quit: Leaving.] 06:09:29 -!- oleo [~oleo@xdsl-78-35-186-8.netcologne.de] has quit [Quit: Leaving] 06:38:06 attila_lendvai [~attila_le@5.76.185.228] has joined #sbcl 06:38:06 -!- attila_lendvai [~attila_le@5.76.185.228] has quit [Changing host] 06:38:06 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 06:44:35 -!- Bike [~Glossina@75-164-162-106.ptld.qwest.net] has quit [Ping timeout: 265 seconds] 06:46:15 Bike [~Glossina@75-164-162-106.ptld.qwest.net] has joined #sbcl 07:07:21 slyrus [~chatzilla@adsl-172-11-236-194.dsl.pltn13.sbcglobal.net] has joined #sbcl 07:34:59 -!- slyrus [~chatzilla@adsl-172-11-236-194.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 246 seconds] 07:48:18 |3b| [bbb@2600:3c00::f03c:91ff:fedf:5b65] has joined #sbcl 07:48:18 reb [user@nat/google/x-allzaxamjmqmdtkj] has joined #sbcl 07:48:18 foom [~jknight@2620:15c:6:14:be30:5bff:fedf:6db6] has joined #sbcl 07:48:18 gko [gko@2400:8900::f03c:91ff:fe70:e605] has joined #sbcl 07:56:24 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 08:10:22 -!- Bike [~Glossina@75-164-162-106.ptld.qwest.net] has quit [Quit: loud] 08:17:24 slyrus [~chatzilla@adsl-172-11-236-194.dsl.pltn13.sbcglobal.net] has joined #sbcl 08:46:08 ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has joined #sbcl 08:58:38 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 272 seconds] 08:59:20 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 09:03:52 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 246 seconds] 09:18:15 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:22:30 -!- prxq_ is now known as prxq 09:24:15 edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has joined #sbcl 09:31:38 stassats [~stassats@wikipedia/stassats] has joined #sbcl 09:45:16 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 09:49:43 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 246 seconds] 09:52:35 pnpuff [~harmonic@unaffiliated/pnpuff] has joined #sbcl 09:53:03 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Quit: Ping timeout: ] 09:57:13 (declare (optimize (safety 0))) in the end of late-type.lisp is what's causing a memory fault on (typep 1 '(and x . 2)) 10:00:06 the explanation is lacking and doesn't make much sense 10:09:26 -!- pnpuff [~harmonic@unaffiliated/pnpuff] has quit [] 10:10:40 -!- echo-area [~user@182.92.247.2] has quit [Remote host closed the connection] 10:18:09 -!- heddwch [~yoshi@76.8.3.189] has quit [Ping timeout: 276 seconds] 10:30:30 -!- Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has quit [Ping timeout: 276 seconds] 10:30:47 -!- kludge` [~comet@unaffiliated/espiral] has quit [Ping timeout: 265 seconds] 10:36:14 kludge` [~comet@unaffiliated/espiral] has joined #sbcl 10:54:23 -!- ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has quit [Remote host closed the connection] 11:34:01 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 11:38:38 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 264 seconds] 11:42:02 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 11:42:22 attila_lendvai [~attila_le@5.76.185.228] has joined #sbcl 11:42:23 -!- attila_lendvai [~attila_le@5.76.185.228] has quit [Changing host] 11:42:23 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 11:56:15 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 12:11:43 yacks [~py@103.6.159.103] has joined #sbcl 12:23:36 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 276 seconds] 12:39:12 eudoxia [~eudoxia@r179-24-23-152.dialup.adsl.anteldata.net.uy] has joined #sbcl 13:22:48 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 13:27:31 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Ping timeout: 260 seconds] 13:40:39 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 13:42:23 -!- ltbarcly [~textual@pool-108-42-99-156.snfcca.fios.verizon.net] has quit [Quit: Computer has gone to sleep.] 13:42:45 heddwch [~yoshi@76.8.3.189] has joined #sbcl 13:56:55 -!- xymox [lechuck@unaffiliated/contempt] has quit [Ping timeout: 272 seconds] 13:58:00 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 14:02:24 eudoxia_ [~eudoxia@r179-25-4-106.dialup.adsl.anteldata.net.uy] has joined #sbcl 14:06:20 -!- eudoxia [~eudoxia@r179-24-23-152.dialup.adsl.anteldata.net.uy] has quit [Ping timeout: 265 seconds] 14:09:25 nyef [~nyef@pool-70-109-134-244.cncdnh.east.myfairpoint.net] has joined #sbcl 14:09:36 G'morning all. 14:11:33 -!- eudoxia_ [~eudoxia@r179-25-4-106.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving] 14:11:45 eudoxia [~eudoxia@r179-25-4-106.dialup.adsl.anteldata.net.uy] has joined #sbcl 14:12:22 morning nyef 14:18:06 I had forgotten how quickly things can go when you're putting together small missing bits from a large piece of software. 14:19:05 I see that there is definite progress on arm-port-2. Enjoying the long-range kibitzing :) 14:19:55 my small contribution to the port is remembering how to log into my beaglebone 14:20:05 let me know as/when that could be useful 14:20:15 Last night my build got as far as src/code/unportable-float before choking on the fact that I killed all of the number-stack stuff. And if I disable that, it gets to src/code/kernel before choking on seventeen recursive definitions of known functions. 14:20:52 that is quite some way 14:21:01 Most of which will be easy to define, of course. 14:21:55 My current goal is to get target-alieneval built, and then to try a simple ALIEN-FUNCALL. 14:22:20 After that, I'm thinking to put the project down for a while to work on other stuff. 14:33:39 -!- drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has quit [Remote host closed the connection] 14:56:58 drmeister [~drmeister@166.170.40.83] has joined #sbcl 14:57:07 oleo [~oleo@xdsl-78-35-186-8.netcologne.de] has joined #sbcl 15:15:01 -!- drmeister [~drmeister@166.170.40.83] has quit [Read error: Connection reset by peer] 15:17:02 asedeno [~asedeno@66.102.14.24] has joined #sbcl 15:26:53 Hunh. WIDETAG-OF an instance is INSTANCE-POINTER-LOWTAG? 15:43:58 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Read error: Connection reset by peer] 15:53:40 drmeister [~drmeister@166.170.40.83] has joined #sbcl 15:59:49 rpg [~rpg@198-74-7-110.fttp.usinternet.com] has joined #sbcl 15:59:50 -!- drmeister [~drmeister@166.170.40.83] has quit [Read error: Connection reset by peer] 16:03:58 drmeister [~drmeister@166.170.40.83] has joined #sbcl 16:35:28 -!- yacks [~py@103.6.159.103] has quit [Quit: Leaving] 16:37:32 pranavrc [~pranavrc@122.174.20.195] has joined #sbcl 16:37:32 -!- pranavrc [~pranavrc@122.174.20.195] has quit [Changing host] 16:37:32 pranavrc [~pranavrc@unaffiliated/pranavrc] has joined #sbcl 16:41:58 WIDETAG-OF: Seven instructions, no branches. 16:42:13 Six instructions on little-endian systems. 16:50:05 The old flow was five instructions for an immediate value, seven instructions for a pointer, both flows having a taken branch, for ten instructions total. 16:55:50 -!- gko [gko@2400:8900::f03c:91ff:fe70:e605] has quit [Ping timeout: 240 seconds] 16:58:04 gko [gko@2400:8900::f03c:91ff:fe70:e605] has joined #sbcl 17:10:54 -!- drmeister [~drmeister@166.170.40.83] has quit [Read error: Connection reset by peer] 17:21:37 hlavaty [~user@friedrichstrasse.knowledgetools.de] has joined #sbcl 17:22:13 -!- oleo [~oleo@xdsl-78-35-186-8.netcologne.de] has quit [Ping timeout: 248 seconds] 17:23:03 oleo [~oleo@xdsl-84-44-152-188.netcologne.de] has joined #sbcl 17:24:18 -!- xymox [lechuck@unaffiliated/contempt] has quit [Ping timeout: 245 seconds] 17:32:43 drmeister [~drmeister@166.170.40.83] has joined #sbcl 17:32:59 -!- drmeister [~drmeister@166.170.40.83] has quit [Remote host closed the connection] 17:33:35 drmeister [~drmeister@155.247.96.196] has joined #sbcl 17:34:22 xymox [lechuck@unaffiliated/contempt] has joined #sbcl 17:41:20 alpha- [~silver@unaffiliated/alpha--] has joined #sbcl 17:42:14 is there a way to increase precision of calculations done by sbcl ? Like, increase the number of decimal places in results 17:43:04 use double-floats 17:44:10 and beyond that? arbitrary precision ? 17:44:26 This is a frightening question, because it suggests a lack of knowledge about how floating-point numbers work. 17:45:13 it depends on what computations you're doing 17:45:28 if your computations are rational ones, you can use rationals 17:45:46 if your computations are transcendental, you get to implement arbitrary precision or use clisp 17:45:51 there's computable reals too 17:45:56 true 17:46:09 -!- pranavrc [~pranavrc@unaffiliated/pranavrc] has quit [Ping timeout: 240 seconds] 17:47:22 can I somehow interface with GMP or PARI ? 17:47:42 but you probably don't need that. FP accumulates errors, continued fractions accumulate computation time. 17:48:08 alpha-: there's already arbitrary precision rational arithmetic in the CL standard. 17:51:39 -!- kanru [~kanru@118-163-10-190.HINET-IP.hinet.net] has quit [Ping timeout: 260 seconds] 17:53:37 fiveop [~fiveop@p5DDC42AE.dip0.t-ipconnect.de] has joined #sbcl 17:54:18 Hello fiveop. 17:54:59 hi 17:55:04 I pulled your last commits 17:55:21 From... a few minutes ago? 17:55:27 just now 17:55:33 don't know how old they are 17:55:35 Fair enough. 17:55:50 are VOPs one of the two IRs? 17:56:12 VOPs are some of the building blocks for IR2. 17:56:46 -!- alpha- [~silver@unaffiliated/alpha--] has left #sbcl 17:56:49 What comes first in the compilation process. IR1 or IR2? 17:56:53 IR1. 17:56:57 Thank you 17:57:06 enough questions for the next quarter hour 17:57:19 Heh. Okay. 17:58:32 I expect that there's going to come a point, fairly soon, where I drop this thing again for a while to work on other stuff. 18:02:00 What ARM Version do you target? 18:02:26 I understand, that not every version specifies an FPU. 18:02:53 (Though ARMv6, which runs on Raspberry Pi, does) 18:03:55 -!- eudoxia [~eudoxia@r179-25-4-106.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving] 18:04:36 Originally, I was targeting ARMv5, though right now I really don't care. It's enough to get the system working on ONE version, and then we can discuss portability later. 18:05:26 Oh, and I'm considering getting a BeagleBone, since it looks like it would be a bit faster than the Pi for this, even though the Pi looks like the better choice for media stuff and interactivity. 18:06:10 any thoughts on the odroid u3 or whatever it's called? 18:07:29 Nope. No thoughts at all. 18:08:03 I was thinking of ordering one of these: http://hardkernel.com/main/products/prdt_info.php?g_code=G138733896281&tab_idx=2 18:08:39 nyef: beaglebone runs ARMv7 18:09:37 A hackable wifi/SD card might be cheaper ;) 18:10:25 Did you watch the talk? 18:10:33 Ooh. The U3 looks sweet. 18:10:45 ... what talk? 18:10:46 Vivitron [~Vivitron@c-50-172-44-193.hsd1.il.comcast.net] has joined #sbcl 18:12:00 I meant pkhuong. There was a talk on microcontrollers in SD-Cards at 30c3. 18:13:45 Right, bunnie's work. That's not quite practical yet. There's a brand of Wifi-enabled SD cards with a system(3) injection bug in their HTTP shell. 18:20:25 slyrus: I've seen tweets from R.Joswig stating his success with Clozure on an ODROID XU. It has an ARM big.LITTLE chip 18:25:28 ltbarcly [~textual@216.113.168.148] has joined #sbcl 18:35:28 -!- ltbarcly [~textual@216.113.168.148] has quit [Ping timeout: 260 seconds] 18:38:40 ltbarcly [~textual@216.113.168.148] has joined #sbcl 19:15:14 Bike [~Glossina@67-5-233-232.ptld.qwest.net] has joined #sbcl 19:17:57 -!- ivan``_ [~ivan@unaffiliated/ivan/x-000001] has quit [Ping timeout: 248 seconds] 19:20:25 ivan`` [~ivan@unaffiliated/ivan/x-000001] has joined #sbcl 19:26:56 -!- slyrus [~chatzilla@adsl-172-11-236-194.dsl.pltn13.sbcglobal.net] has quit [Remote host closed the connection] 19:28:13 slyrus [~chatzilla@adsl-172-11-236-194.dsl.pltn13.sbcglobal.net] has joined #sbcl 19:38:48 wifi-enabled sd cards 19:38:52 what were they thinking? 19:41:29 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Read error: Operation timed out] 19:45:14 scymtym_ [~user@ip-5-147-115-29.unitymediagroup.de] has joined #sbcl 19:46:11 where is other-pointer-lowtag defined? 19:46:19 compiler/generic/early-objdef. 19:46:56 Krystof: leave the SD card in your camera, download to your laptop via HTTP 19:51:12 thx 19:53:26 -!- foom [~jknight@2620:15c:6:14:be30:5bff:fedf:6db6] has quit [Quit: Leaving] 19:54:26 foom [~jknight@2620:15c:6:14:be30:5bff:fedf:6db6] has joined #sbcl 20:03:34 -!- ivan`` [~ivan@unaffiliated/ivan/x-000001] has quit [Ping timeout: 240 seconds] 20:06:48 ivan`` [~ivan@unaffiliated/ivan/x-000001] has joined #sbcl 20:07:47 there are years when I am glad that I am a long way behind the curve of technology 20:20:55 leuler [~user@p548F9D8F.dip0.t-ipconnect.de] has joined #sbcl 20:21:57 -!- drmeister [~drmeister@155.247.96.196] has quit [Ping timeout: 248 seconds] 20:46:15 and of course you'll want HDMI out from that card too... 20:46:24 nyef: what does LRA stand for? 20:46:29 Lisp Return Address. 20:46:53 crixus [~Rob@69.77.176.98] has joined #sbcl 20:47:20 thank you 20:48:07 so how about a summer of code project to hack a version of SBCL with decent JVM integration? 20:48:57 maybe the TLA file could be expanded by a few abbreviations consistently used in the code 20:49:20 like LRA/NFP (and I'm sure I'll find more, while reading) 20:50:15 slyrus: And here I was thinking that a summer of code project to create a new compiler backend, or finish up an existing-but-incomplete one might be a good idea. (-: 20:50:33 well, yes, that too :) 20:53:49 Oh, hell. The COMPUTE-LRA "instruction" can be passed a BACKWARDS label, completely breaking it. /-: 21:00:33 what is reg_CODE used for? 21:00:58 Most of the time, it points to the "code object" that contains the function currently executing. 21:02:07 It's used as the basis for finding boxed constants (there's a constant vector as part of the code object), and to help the GC know when and how to update the program counter / instruction pointer when a code object moves during garbage collection. 21:03:22 compiled code is in objects handled by the gc? that sounds interesting 21:03:31 Doesn't it, though? 21:03:45 that sounds like a stupid gc ;) 21:04:21 pkhuong: You'd rather have compiled code be stuck in memory permanently, refcounted, or manually managed? 21:04:36 not copied around, for sure. 21:04:37 Oh, wait, refcount still counts as a gc, even if it is a joke... 21:05:12 Oh yeah, speaking of. Did I mention that I found a way to make CODE_PAGE_FLAG not a horrible joke? 21:05:39 better than your recent commits? 21:06:11 Yeah, my recent commits remove the dependency that conservative roots had on it. 21:06:21 But I now know how to actually enforce it. 21:06:48 Problem is, genesis doesn't enforce it, so we'd need a full non-conservative GC before it fully kicks in. 21:08:19 We'd need to do something clever to arrange to pin code_page_p pages when they're found to contain live objects, and to disable that noise when doing a gc_and_save, but... 21:14:43 genesis could be taught to be cleverer 21:15:39 Also on my list of possible projects, somewhere after forward-porting http://repo.or.cz/w/sbcl/nyef.git/shortlog/refs/heads/radical-refactoring/genesis and organizing the commits (and possibly the code) a bit better. 21:16:26 But that's a large enough bit of infrastructure damage that I'd like someone else to agree with me that it's the right direction to go in before I start. 21:16:59 But genesis outputting a page table means that we can ditch gencgc_partial_pickup. 21:17:04 ltt_ [~ltt_@201.75.216.14] has joined #sbcl 21:24:59 -!- ltt_ [~ltt_@201.75.216.14] has quit [Quit: Textual IRC Client: www.textualapp.com] 21:35:04 -!- foom [~jknight@2620:15c:6:14:be30:5bff:fedf:6db6] has quit [Quit: Leaving] 21:46:49 foom [~jknight@2620:15c:6:14:be30:5bff:fedf:6db6] has joined #sbcl 21:57:15 -!- mood [~mood@146.185.164.46] has quit [Ping timeout: 272 seconds] 22:01:25 -!- Posterdati [~kvirc@host208-231-dynamic.2-79-r.retail.telecomitalia.it] has quit [Excess Flood] 22:01:53 Posterdati [~kvirc@host208-231-dynamic.2-79-r.retail.telecomitalia.it] has joined #sbcl 22:03:17 mood [~mood@146.185.164.46] has joined #sbcl 22:10:20 drmeister [~drmeister@155.247.96.196] has joined #sbcl 22:10:58 Oh, hell. I'm catching a failed-aver in the middle of processing a back-patch for a load/store instruction. I know which function is being compiled, but not which instruction is being assembled, nor which VOP generated the offending function. 22:11:51 I do know, however, that there is a label involved. 22:16:43 nyef: too large offset, like we have on SPARC? 22:18:05 Yes, that much is a given. 22:18:15 The question is, where? 22:20:55 The trace file doesn't seem to help, either. /-: 22:21:29 I'm more worried about how to fix this. a load immediate pseudo instruction that expands into two instructions on demand? 22:22:21 Okay, such a load immediate instruction can't really be used in a back-patch or a chooser without knowing exactly what's wrong with it... 22:22:44 What's the situation that you're looking at? 22:23:59 -!- drmeister [~drmeister@155.247.96.196] has quit [Remote host closed the connection] 22:25:08 actually, no, I think our SPARC issue doesn't need that much smartness. We should just be able to look at tn-offset. 22:27:49 ... Oh. I know what's causing this. I've been stashing various odd bits in the ELSEWHERE segment, and with the size of this function that delta has overflowed. 22:29:07 when you talk about 'boxed registers', do you mean registers the gc should look at? 22:29:42 Sortof. They're registers that the GC should look at and must contain boxed object references or immediate data. 22:29:52 (Boxed immediate data, that is.) 22:30:02 So there's a type-tag on anything in those registers. 22:30:10 -!- leuler [~user@p548F9D8F.dip0.t-ipconnect.de] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 22:30:24 The other kind of register the GC should look at is an "interior pointer". 22:32:54 so in a boxed (descriptor) register, a fixnum has to be encoded, and in non-descriptor registers, you could encode it raw? 22:33:04 Yes, exactly. 22:34:18 Also, in a boxed register, you can have a pointer to an object in the heap and the GC will update the pointer if it moves the object, but if that pointer was in a non-descriptor register then the GC would leave it alone. 22:34:54 are our gcs conservativ? 22:35:15 On x86oids, yes. On every other platform, no. 22:35:40 ARM will be precisely GC'd, not conservatively GC'd. 22:36:00 which is good, as far as I understand it 22:36:21 it's a trade off 22:36:28 It is and it isn't. Having to partition our register set simply Does Not Help with the register pressure. 22:36:59 You'll see several parts in my port log where I'm desperately scrambling to try and find some way to free up a register. 22:37:00 If a GC ever encounters a reference to an object, where he is not sure, that it's actually a reference, he cannot move the object, correct? 22:37:05 Correct. 22:37:24 how can that happen 22:37:31 And there's a mechanism whereby a program can specify that a reference to an object must not be moved. 22:38:02 The GC is only ever unsure about references on x86 and x86-64, as they're the conservative platforms. 22:38:03 (so far I only see you using more and more registers :)) 22:38:58 but isn't being able to move stuff around crucial to using your heap well (assuming worst case strain on heap/gc) 22:39:16 Yup! 22:39:33 Oh, and the first register pressure looks to start at 2012-Oct-14. 22:40:24 I'm currently on that day in the log and day 16 in the commits (they don't match up that well even with respect to the order) 22:41:07 Yeah, I'll work on a bunch of stuff, get it mostly working, then break out individual commits. 22:46:39 you are currently only using ARM instructions and no THUMB instructions, correct? 22:46:47 Correct. 22:47:11 I have no plans to use THUMB at this point, either. 22:48:12 how come the pc value changes when pc is read? 22:48:18 (or something along those lines) 22:48:59 Per the ARM architecture, reading from the PC register returns a value that is two instructions further along than the current instruction, due to pipeline issues. 22:49:09 -!- slyrus [~chatzilla@adsl-172-11-236-194.dsl.pltn13.sbcglobal.net] has quit [Ping timeout: 248 seconds] 22:49:39 I don't remember if there was anything in there about doing so in the immediate aftermath (or foremath) of a control-flow instruction. 22:50:39 drmeister [~drmeister@166.170.40.83] has joined #sbcl 22:52:00 thanks again for all the answers, good night (according to the scrollbar I'm about 1/3 through your log :)) 22:52:08 -!- fiveop [~fiveop@p5DDC42AE.dip0.t-ipconnect.de] has quit [Quit: humhum] 22:52:28 Yeah, 1/3 sounds about right. 22:54:32 slyrus [~chatzilla@adsl-172-11-236-194.dsl.pltn13.sbcglobal.net] has joined #sbcl 23:05:24 One "good" thing is that all of those bits that I knew were going to be fragile due to limited offset width and whatnot are all starting to break now, so I'm having to find solutions rather than leaving them as nasty little trip mines for the unwary. 23:06:53 ehaliewicz [~user@50-0-51-28.dsl.static.sonic.net] has joined #sbcl 23:22:26 -!- crixus [~Rob@69.77.176.98] has quit [Ping timeout: 264 seconds] 23:27:27 -!- drmeister [~drmeister@166.170.40.83] has quit [Read error: Connection reset by peer] 23:29:05 isn't it recommended these days to basically just always use thumbv2? 23:29:38 (not that it probably matters much except code-size, since the instruction sets are basically isomorphic, right?) 23:46:39 -!- edgar-rft [~GOD@HSI-KBW-109-193-013-113.hsi7.kabel-badenwuerttemberg.de] has quit [Quit: continuation disconnected by sudden death] 23:48:49 crixus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 23:50:04 drmeister [~drmeister@pool-71-175-2-214.phlapa.fios.verizon.net] has joined #sbcl 23:50:29 -!- scymtym_ [~user@ip-5-147-115-29.unitymediagroup.de] has quit [Ping timeout: 248 seconds] 23:57:37 crixxus [~Rob@135-23-80-105.cpe.pppoe.ca] has joined #sbcl 23:58:22 rpg_ [~rpg@198-74-7-110.fttp.usinternet.com] has joined #sbcl