00:00:40 stassats: i tried one week with clojure, i came back to cl because of slime XD 00:01:17 stassats: a precise gc probably needs a whole new design of compiler and code generation 00:29:33 at least sbcl 1.0.56 is working as usual, was a problem with project directory (.bak and .old dirs) 00:42:23 -!- homie [~levgue@xdsl-84-44-153-69.netcologne.de] has quit [Read error: Connection reset by peer] 00:43:01 homie [~levgue@xdsl-84-44-153-69.netcologne.de] has joined #sbcl 00:43:02 ckairaba [~ckairaba@39.104.100.84.rev.sfr.net] has joined #sbcl 00:56:20 -!- ckairaba [~ckairaba@39.104.100.84.rev.sfr.net] has quit [Quit: Quitte] 01:21:29 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 252 seconds] 01:32:19 attila_lendvai [~attila_le@178-164-242-247.pool.digikabel.hu] has joined #sbcl 01:32:19 -!- attila_lendvai [~attila_le@178-164-242-247.pool.digikabel.hu] has quit [Changing host] 01:32:19 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 01:48:34 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 02:09:52 LiamH [~healy@pool-74-96-18-66.washdc.east.verizon.net] has joined #sbcl 02:45:34 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 02:45:41 -!- specbot [~specbot@pppoe.178-66-69-164.dynamic.avangarddsl.ru] has quit [Disconnected by services] 02:45:43 specbot [~specbot@pppoe.178-66-10-110.dynamic.avangarddsl.ru] has joined #sbcl 02:48:49 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 276 seconds] 02:56:48 angavrilov [~angavrilo@217.71.227.190] has joined #sbcl 03:20:26 stassats [~stassats@wikipedia/stassats] has joined #sbcl 03:33:18 -!- stassats [~stassats@wikipedia/stassats] has quit [Read error: Operation timed out] 03:39:37 saschakb_ [~saschakb@p4FEA120F.dip0.t-ipconnect.de] has joined #sbcl 03:42:58 -!- saschakb [~saschakb@p4FEA0AC8.dip0.t-ipconnect.de] has quit [Ping timeout: 245 seconds] 04:01:34 plutoid [~pluto@61.170.248.215] has joined #sbcl 04:40:05 sdemarre [~serge@91.176.18.35] has joined #sbcl 04:40:44 -!- LiamH [~healy@pool-74-96-18-66.washdc.east.verizon.net] has quit [Quit: Leaving.] 05:40:00 cmm- [~cmm@bzq-79-182-209-72.red.bezeqint.net] has joined #sbcl 05:40:00 -!- cmm [~cmm@bzq-79-180-227-205.red.bezeqint.net] has quit [Ping timeout: 260 seconds] 06:15:41 yates [~user@nc-71-48-9-61.dhcp.embarqhsd.net] has joined #sbcl 06:15:54 is clisp 2.49 ok for building sbcl? 06:16:15 it *seems* to be running ok with clisp 06:16:20 INSTALL says "only some versions" 06:32:56 -!- plutoid [~pluto@61.170.248.215] has quit [Quit: Leaving] 06:49:17 SBCL is ok for building SBCL 06:49:34 you shouldn't use any other lisps unless you really need to 06:50:34 but, it should be doable, so report any regressions to launchpad 07:50:56 slyrus_ [~chatzilla@99-28-161-110.lightspeed.miamfl.sbcglobal.net] has joined #sbcl 08:58:01 -!- saschakb_ [~saschakb@p4FEA120F.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer] 09:40:11 plutoid [~pluto@114.82.61.120] has joined #sbcl 09:40:30 pluto_ [~pluto@114.82.61.120] has joined #sbcl 09:44:14 attila_lendvai [~attila_le@188-143-61-13.pool.digikabel.hu] has joined #sbcl 09:44:14 -!- attila_lendvai [~attila_le@188-143-61-13.pool.digikabel.hu] has quit [Changing host] 09:44:14 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 09:45:11 -!- pluto_ [~pluto@114.82.61.120] has quit [Client Quit] 10:12:06 homie` [~levgue@xdsl-87-79-197-9.netcologne.de] has joined #sbcl 10:14:51 -!- homie [~levgue@xdsl-84-44-153-69.netcologne.de] has quit [Ping timeout: 244 seconds] 10:17:52 nikodemus [~nikodemus@cs78186070.pp.htv.fi] has joined #sbcl 10:17:52 -!- ChanServ has set mode +o nikodemus 10:21:46 could there be a heuristically adaptable GC? 10:24:24 prxq [~mommer@mnhm-4d010c62.pool.mediaWays.net] has joined #sbcl 10:44:45 in principle 10:59:48 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 11:04:02 stassats`: there's an madvise flag to tell linux (unportably) that the address range can eb reclaimed if needed. 11:05:11 (mornin' from a friendly mcdonald's... seems universities close on weekends in frane) 11:05:14 pkhuong: MADV_DONTNEED? 11:05:34 i threw a simple test in C, and it's the same as unmap+mmap 11:06:41 and appears to be portable 11:09:08 i'm more surprised why RES grows beyond bytes-consed-between-gcs 11:13:45 stassats`: do you have both a well-behaving and a res-growing sbcl handy? 11:14:12 an old well-behaving 11:14:18 1.0.42 11:14:22 that's ok 11:15:24 http://paste.lisp.org/display/129216 11:15:48 (didn't test outside darwin yet, but linux top starts it's header line with PID, doesn't it?) 11:16:48 (except that time goes to *trace-output*, but that's just cosmetic) 11:18:03 hum. not sure if the command-line switches need to be adjusted for 1.0.42 11:21:06 re GC, I'd really like to see what we could do with s/w write barriers and a fork-based concurrent mark and sweep. 11:22:01 -!- Phoodus [~foo@wsip-68-107-217-139.ph.ph.cox.net] has quit [Ping timeout: 244 seconds] 11:24:14 nikodemus: http://paste.lisp.org/display/129216#1 bad sbcl 11:26:44 hm, it didn't return a pretty table for some reason 11:26:49 while it did for an old sbcl 11:27:10 i'll try again 11:28:24 ok, did correctly on the second try 11:28:49 nikodemus: http://paste.lisp.org/display/129216#2 11:29:22 thanks 12:20:30 attila_lendvai [~attila_le@188-143-61-13.pool.digikabel.hu] has joined #sbcl 12:20:31 -!- attila_lendvai [~attila_le@188-143-61-13.pool.digikabel.hu] has quit [Changing host] 12:20:31 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 12:26:59 -!- nikodemus [~nikodemus@cs78186070.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 12:35:38 -!- prxq [~mommer@mnhm-4d010c62.pool.mediaWays.net] has quit [Quit: Leaving] 12:44:56 attila_lendvai1 [~attila_le@188-143-61-13.pool.digikabel.hu] has joined #sbcl 12:44:56 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Disconnected by services] 12:45:06 -!- attila_lendvai1 is now known as attila_lendvai 12:45:07 -!- attila_lendvai [~attila_le@188-143-61-13.pool.digikabel.hu] has quit [Changing host] 12:45:07 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 12:59:14 attila_lendvai1 [~attila_le@178-164-242-115.pool.digikabel.hu] has joined #sbcl 12:59:14 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Disconnected by services] 12:59:15 -!- attila_lendvai1 is now known as attila_lendvai 12:59:21 -!- attila_lendvai [~attila_le@178-164-242-115.pool.digikabel.hu] has quit [Changing host] 12:59:21 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 13:11:31 -!- plutoid [~pluto@114.82.61.120] has quit [Quit: Leaving] 13:17:33 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Ping timeout: 245 seconds] 13:41:38 leuler [~user@p549023CB.dip.t-dialin.net] has joined #sbcl 13:43:22 nikodemus [~nikodemus@cs27100107.pp.htv.fi] has joined #sbcl 13:43:22 -!- ChanServ has set mode +o nikodemus 13:51:24 prxq [~mommer@mnhm-4d010c62.pool.mediaWays.net] has joined #sbcl 13:51:32 sbcl 1.0.56 already exhibits the RES creep 13:53:47 what is RES creep? 13:55:34 http://paste.lisp.org/display/129216#2 compare the two last annotations 14:17:14 -!- sdemarre [~serge@91.176.18.35] has quit [Ping timeout: 246 seconds] 14:29:26 -!- siccegge` [~user@davinci.siccegge.de] has quit [Read error: No route to host] 14:38:31 -!- sshirokov [sshirokov@2600:3c02::f03c:91ff:fe93:e02d] has quit [Ping timeout: 260 seconds] 14:38:48 sshirokov [sshirokov@2600:3c02::f03c:91ff:fe93:e02d] has joined #sbcl 14:57:34 1.0.51.29-cd470a0 ; good... 15:03:23 -!- slyrus_ [~chatzilla@99-28-161-110.lightspeed.miamfl.sbcglobal.net] has quit [Ping timeout: 245 seconds] 15:06:01 1.0.53.87-79bc50a ; bad.. 15:07:48 could there be a run-time option on how many generations to release memory? 15:08:04 because i imagine for some workloads it can be better to never release memory 15:13:10 it would be easy to make small-generation-limit accessible from lisp, sure 15:13:19 just need a sensible name for it 15:14:15 and to change the condition from > to >=, so that 0 can be specified to mean "always" 15:15:58 -!- homie` [~levgue@xdsl-87-79-197-9.netcologne.de] has quit [Quit: ERC Version 5.3 (IRC client for Emacs)] 15:16:47 1.0.53.10-9cd0440 ; good ... tension mounts! 15:17:39 place your bets! 15:17:42 gen is 1 after a nursery collection 15:17:55 so 0 is already always 15:18:30 nursery is always remapped? 15:18:53 homie [~levgue@xdsl-87-79-197-9.netcologne.de] has joined #sbcl 15:18:56 um, no? some miscommunication here 15:19:07 attila_lendvai [~attila_le@178-164-242-115.pool.digikabel.hu] has joined #sbcl 15:19:08 -!- attila_lendvai [~attila_le@178-164-242-115.pool.digikabel.hu] has quit [Changing host] 15:19:08 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 15:19:26 by always i meant "always release memory" 15:19:35 nursery is zero, but when that test is run gen is one greater than the last collected generation 15:19:49 so setting the limit to zero would always release memory 15:20:23 oh, ok, i guess 15:20:38 my bet is that "default gen.bytes_consed_between_gc to same 5% of dynamic space size" is the culprit 15:21:10 but in the test you set the bytes_consed_between_gc explicitly 15:21:15 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Client Quit] 15:21:23 only for nursery, not other generations 15:21:42 oh, ok 15:22:10 yeah, 1.0.42 shows that memory grows, just not as much 15:22:33 so, i think it has always been so, just magnified 15:24:19 attila_lendvai [~attila_le@178.164.242.115] has joined #sbcl 15:24:19 -!- attila_lendvai [~attila_le@178.164.242.115] has quit [Changing host] 15:24:19 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 15:24:25 SBCL 1.0.53.10-9cd0440 ends up with 593m 15:24:43 1.0.53.87-79bc50a with 800m 15:26:58 1.0.53.48-b43491a bad, with 764m... 15:27:27 ~4 bisection steps left 15:27:32 *stassats`* misses version numbers in the log 15:27:41 me too, sort of 15:28:07 i think it would be possible to script git-log to generate them, actually 15:29:38 *stassats`* doesn't miss "did i forget to update version.lisp-expr?" 15:30:49 I think there's an option to generate number of commits since last tag 15:36:04 -!- gko [~gko@114-34-168-13.HINET-IP.hinet.net] has quit [Ping timeout: 252 seconds] 15:39:28 1.0.53.29-993f64e => 605m ... good, i think (based on the progression) 15:44:21 not that i can see (in git log) 15:44:46 need to roll it from git rev-list, log, and describe 15:46:55 *stassats`* made a building sbcl with safety 0, and it indeed shaves 40 seconds of the build time, now to 2:42 15:47:32 hm, we could add a commit note with the version, though... 15:49:44 CampinSam [~Sam@24-176-98-217.dhcp.jcsn.tn.charter.com] has joined #sbcl 15:50:28 1.0.53.38-6cfdd9e ok with 604 15:53:08 on an unrelated note, https://github.com/sbcl/sbcl/blob/master/src/runtime/thread.c#L386 it compares th->tls_cookie>=0, but tls_cookie is a lispobj, which is unsigned long 15:55:09 hunh 15:55:20 can you put that on lp? 15:55:37 ok 15:56:01 how do i title it? 15:59:06 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Quit: Leaving.] 16:03:52 stassats`: re build time: that large a reduction surprises me. The last time I measured the difference (with 1.0.50.54) I had 35 seconds less, but from 6:00 to 5:25 min, so half the relative gain. Which versions is your comparison based on? 16:04:11 1.0.42 and the latest 16:04:43 i set speed 3, everything else 0 16:05:22 can you get timings for everything before make-target-2 and everything after it? 16:06:02 on next rebuild, i'll try 16:06:03 So you compared how long an unchanged 1.0.42 took to compile itself, and how long a speed-3-latest took to compile the latest version? 16:06:12 (and the same when using the unsafe sbcl as the host) 16:06:21 leuler: not exactly 16:06:39 1.0.42 to compile the latest version, and the fast latest version to compile the latest version 16:09:23 current building current / current building fast / fast building fast, with times for make-target-2 and later accounted separately would be interesting 16:10:16 stassats`: OK, understood. 16:12:52 1.0.53.45-3cd198e: good, with 598m 16:14:04 stassats`: I believe since 1.0.42 the compiler has become measurably faster, so not all of the reduction you see can be attributed to the change in optimization qualities. 16:14:45 well, i had no other reason to change the building compiler, so attribute it to it too! 16:25:38 tsuru` [~charlie@adsl-74-240-217-174.bna.bellsouth.net] has joined #sbcl 16:49:31 attila_lendvai [~attila_le@178-164-242-115.pool.digikabel.hu] has joined #sbcl 16:49:31 -!- attila_lendvai [~attila_le@178-164-242-115.pool.digikabel.hu] has quit [Changing host] 16:49:31 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 17:13:43 tried profiling the runtime when running a gc test, most of the time appears to be spent in bzero 17:20:45 wow 17:32:59 it depends a lot on the load 17:33:24 stassats`: fix to the RES creep pushed 17:33:32 it's binary-trees from the shootout 18:09:38 -!- nikodemus [~nikodemus@cs27100107.pp.htv.fi] has quit [Ping timeout: 246 seconds] 18:17:07 nikodemus [~nikodemus@cs78186070.pp.htv.fi] has joined #sbcl 18:17:07 -!- ChanServ has set mode +o nikodemus 18:22:05 milanj [~milanj_@178-223-150-244.dynamic.isp.telekom.rs] has joined #sbcl 18:23:12 nikodemus: seems somewhat better, but still, i think adding a customizable small-generation-limit would be the way to go to really address this 18:24:32 stassats`: yes -- or accurate "n pages availabe for release to the OS" bookkeeping, or another heuristic 18:25:26 i left an incomplete bug-report on the "consing nothing but garbage" incarnation of this open in a tab at the office... 18:26:00 christoph_debian [~user@oteiza.siccegge.de] has joined #sbcl 18:50:16 tcr [~tcr@84-72-21-32.dclient.hispeed.ch] has joined #sbcl 18:54:00 -!- nikodemus [~nikodemus@cs78186070.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 19:02:58 nikodemus [~nikodemus@cs78186070.pp.htv.fi] has joined #sbcl 19:03:05 -!- ChanServ has set mode +o nikodemus 19:09:53 -!- nikodemus [~nikodemus@cs78186070.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 19:10:00 -!- CampinSam [~Sam@24-176-98-217.dhcp.jcsn.tn.charter.com] has quit [Read error: Connection reset by peer] 19:10:41 CampinSam [~Sam@24-176-98-217.dhcp.jcsn.tn.charter.com] has joined #sbcl 19:18:18 nikodemus [~nikodemus@cs78186070.pp.htv.fi] has joined #sbcl 19:18:18 -!- ChanServ has set mode +o nikodemus 19:23:23 sdemarre [~serge@91.176.18.35] has joined #sbcl 19:30:03 -!- nikodemus [~nikodemus@cs78186070.pp.htv.fi] has quit [Quit: Leaving] 19:39:39 *stassats`* read the commit message as redesign exiSting SBCL 19:39:43 was quite confusing 19:40:59 -!- tcr [~tcr@84-72-21-32.dclient.hispeed.ch] has quit [Ping timeout: 246 seconds] 19:41:44 tcr [~tcr@84-72-21-32.dclient.hispeed.ch] has joined #sbcl 19:43:07 saschakb [~saschakb@p4FEA120F.dip0.t-ipconnect.de] has joined #sbcl 20:09:25 -!- angavrilov [~angavrilo@217.71.227.190] has quit [Ping timeout: 248 seconds] 20:10:23 -!- yates [~user@nc-71-48-9-61.dhcp.embarqhsd.net] has quit [Quit: rcirc on GNU Emacs 24.0.95.1] 20:25:04 -!- sdemarre [~serge@91.176.18.35] has quit [Ping timeout: 276 seconds] 20:32:49 -!- milanj [~milanj_@178-223-150-244.dynamic.isp.telekom.rs] has quit [Quit: Leaving] 20:41:27 gunter [~user@ip68-231-216-62.oc.oc.cox.net] has joined #sbcl 20:46:02 Hello, this is my first attendance and I ask you kindly for advice. 20:46:25 this works in the repl: (defun prompt-read (prompt) 20:46:27 (format t "~a:" prompt) 20:46:27 (finish-output nil) 20:46:27 (read-line)) 20:46:58 but does not work in the case of "sbcl --script " 20:48:10 esp. the output flush doesn't work when starting sbcl with the option --script 20:49:08 I'm using current sbcl version for windows. 20:49:14 Thanks a lot 21:08:58 -!- gunter [~user@ip68-231-216-62.oc.oc.cox.net] has quit [Ping timeout: 260 seconds] 21:18:18 -!- saschakb [~saschakb@p4FEA120F.dip0.t-ipconnect.de] has quit [Quit: Verlassend] 21:59:39 -!- tcr [~tcr@84-72-21-32.dclient.hispeed.ch] has quit [Quit: Leaving.] 22:07:52 -!- prxq [~mommer@mnhm-4d010c62.pool.mediaWays.net] has quit [Quit: Leaving] 22:11:37 -!- leuler [~user@p549023CB.dip.t-dialin.net] has quit [Quit: ERC Version 5.1.2 $Revision: 1.796.2.6 $ (IRC client for Emacs)] 22:28:37 homie` [~levgue@xdsl-78-35-156-194.netcologne.de] has joined #sbcl 22:31:30 -!- homie [~levgue@xdsl-87-79-197-9.netcologne.de] has quit [Ping timeout: 252 seconds] 22:56:46 Fare [~Fare@173-9-65-97-NewEngland.hfc.comcastbusiness.net] has joined #sbcl 23:00:06 stassats`: what's confusing? 23:28:22 attila_lendvai1 [~attila_le@178-164-240-36.pool.digikabel.hu] has joined #sbcl 23:28:22 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Disconnected by services] 23:33:50 attila_lendvai [~attila_le@178-164-240-135.pool.digikabel.hu] has joined #sbcl 23:33:50 -!- attila_lendvai [~attila_le@178-164-240-135.pool.digikabel.hu] has quit [Changing host] 23:33:50 attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has joined #sbcl 23:34:58 attila_lendvai2 [~attila_le@188-143-56-181.pool.digikabel.hu] has joined #sbcl 23:34:59 -!- attila_lendvai [~attila_le@unaffiliated/attila-lendvai/x-3126965] has quit [Disconnected by services] 23:35:13 gunter [~user@ip68-231-216-62.oc.oc.cox.net] has joined #sbcl 23:36:24 -!- attila_lendvai1 [~attila_le@178-164-240-36.pool.digikabel.hu] has quit [Ping timeout: 245 seconds]