21:01:30 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 21:01:30 -!- names: ccl-logbot lisppaste2 mega1 kclifton Blkt tcr Fare nyef froydnj mbohun @nikodemus angavrilov svr deepfire The_Jon_Smith specbot rbarraud minion cmm- pkhuong_ slyrus__ Kaes chandler lnostdal @Xof fe[nl]ix foom vs slyrus_ _3b` gnooth jsnell `micro joshe luis tokenrove 21:04:07 attila_lendvai [~attila_le@apn-94-44-6-221.vodafone.hu] has joined #sbcl 11:29:15 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 11:29:15 -!- names: ccl-logbot stassats attila_lendvai @nikodemus tcr Kaer fe[nl]ix _3b` gnooth tsuru lnostdal svr1 lisppaste2 Fare mbohun angavrilov deepfire specbot minion cmm- pkhuong_ slyrus__ chandler @Xof foom vs tokenrove luis joshe `micro jsnell slyrus_ 11:39:32 minion: memo for pkhuong_: Hi, did you have time to look through the SSE patches? 11:39:32 Remembered. I'll tell pkhuong_ when he/she/it next speaks. 11:57:02 -!- attila_lendvai [~attila_le@adsl-89-132-3-59.monradsl.monornet.hu] has left #sbcl 11:58:05 mega1 [~quassel@pool-03710.externet.hu] has joined #sbcl 12:06:36 attila_lendvai [~attila_le@adsl-89-132-3-59.monradsl.monornet.hu] has joined #sbcl 12:09:02 hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has joined #sbcl 13:24:49 -!- hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has quit [Quit: hargettp] 13:37:44 hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has joined #sbcl 13:55:27 -!- hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has quit [Quit: hargettp] 13:55:48 hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has joined #sbcl 13:59:13 nyef [~nyef@pool-70-109-147-129.cncdnh.east.myfairpoint.net] has joined #sbcl 13:59:23 G'morning all. 14:00:08 -!- hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has quit [Ping timeout: 245 seconds] 14:00:59 morning 14:01:15 well, day anyways :) 14:01:54 i call your attention to the date: 10.10.10, and #b101010 => 42. this _has_ to be a good day 14:04:14 nikodemus: Have you at all considered adding a macro version of ensure-gethash with lazy default argument evaluation to alexandria? 14:06:04 hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has joined #sbcl 14:06:32 I don't see much use for the function, while such macro encapsulates the core of any simple hash-based caching scheme. 14:12:06 angavrilov: i haven't considered that, no, but sounds like a good idea 14:14:14 I've ended up making a variation on that in almost every project of mine. Plus a "with-memoize"-style wrapper that passes the hash as (load-time-value (make-hash-table ...)) 14:15:02 Like here, for instance: http://github.com/angavrilov/cl-gpu/blob/master/buffers/utils.lisp 14:16:36 One point of uncertainty here is whether the core lookup macro should be function-like or with-like too. 14:21:40 Indeed, it is the tenth day of the eighth month of 2010. 14:35:34 Or maybe it could be e.g. "with-memoize-hash ((key &rest make-args &key (table nil t-p) &allow-other-keys) &body code)" 14:36:46 (with-hash-value (value default-form table) ...) 14:37:07 dunno 14:38:05 angavrilov: do you have any examples of cl-gpu usage, by the way? 14:38:47 usage of what exactly? 14:39:02 that with-hash-value is very strange 14:39:25 it's missing key, yeah, just realized :) 14:39:41 the semantics of with-memoize is more clear to me: try lookup, if fail execute body and save the result 14:39:54 in both cases return the value 14:40:40 re. cl-gpu -- something trivial: i've never done any gpu programming, so an example of eg. summing a bunch of floats in a vector 14:41:41 hm. cl-gpu isn't in quicklisp yet... 14:41:47 that kind of thing is not trivial to parallelize 14:42:16 summing two arrays and storing the result in a third one is a lot simpler 14:42:34 right -- that works for me too as an example :) 14:47:40 test/cuda-driver/thread-idx here initializes a trivial array with some numbers in a parallelized way: http://github.com/angavrilov/cl-gpu/blob/master/test/cuda-driver.lisp 14:49:06 with-test-gpu-module is different from with-gpu-module only by the fact that it recompiles the module from scratch every time 14:51:43 interesting! 14:52:04 do you have any gpu programming resources you can recommend? 14:56:05 well, there always is the programming guide :) http://developer.download.nvidia.com/compute/cuda/3_2/toolkit/docs/CUDA_C_Programming_Guide.pdf 14:57:48 hm, i have NVIDIA GeForce 8600M GT -- am i good to go or out of luck, and how do i find out? 14:59:47 Well, table A-1 of the guide says it has capability 1.1; that's what cl-gpu currently requires 15:00:23 "0000:f0:10.0 VGA compatible controller: ATI Technologies Inc RV350 AP [Radeon 9600]", according to lspci. I think the kernel might say something differently... 15:03:06 ... it's not in the table? 15:03:32 From nvidia.com? :) 15:04:00 Looking at the programming guide now. 15:04:34 I can find the 9600 GT and the 9600M GT, but not a bare 9600. 15:05:30 i have 8400 GS, and CUDA on it is not much faster than the CPU 15:05:34 X says "ATI Radeon 9600 AP (AGP)"... 15:05:55 well, ATI doesn't look like nvidia 15:06:03 Oops. 15:06:07 *nyef* blinks. 15:06:19 Right, clearly I blanked out there. 15:06:43 And my last nVidia machine melted down last February. 15:08:51 ... and doesn't look to be in the table, either. 15:09:08 stassats: Yeah, for some real gain over modern CPUs you need at least something like GTX270 15:09:50 i only bought it because of video acceleration, and it's doing well 15:12:46 Well, naturally, it's optimized to do video, and not for computing huge expressions with heavy register usage, unaligned data access and so on. 15:15:05 In addition to more computation power, later models also have more registers and better memory access controllers. 15:20:18 why did my fix stop working? this is unfair 15:31:12 Failure: compiler.pure.lisp / BUG-654289 15:31:41 if I change 10 to 20 it passes; maybe the test is assuming too fast of a machine 15:34:11 this is 1.0.43.26 on Linux x86-64 15:34:28 lp #654289 15:34:28 https://bugs.launchpad.net/bugs/654289 15:34:30 rather, 1.0.43.36 15:37:28 kclifton [~kclifton@s198-166-45-245.ab.hsia.telus.net] has joined #sbcl 15:43:43 nyef: I guess I'll unfix the x86 VOP and commit? 15:43:44 pkhuong_, memo from angavrilov: Hi, did you have time to look through the SSE patches? 15:43:59 minion: memo for angavrilov: nope. 15:43:59 Remembered. I'll tell angavrilov when he/she/it next speaks. 15:44:22 pkhuong_: Does the x86 VOP need unfixing, though? 15:44:29 gnooth: thanks for the heads-up 15:44:39 i'll adjust the test into something more sensible 15:45:54 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 15:46:44 nyef: well, it special cases > word-length shifts 15:46:56 nikodemus: thanks 15:47:38 Yes, because shift instructions on x86 are taken mod-32, aren't they? 15:47:44 right. 15:47:48 nyef: -ve shifts still need fixing though. 15:48:32 "-ve"? 15:48:39 negative 15:48:43 Ah. 15:49:09 but fixnum=>fixnum with (constant) positive shift > word length shouldn't happen at all anymore. 15:50:47 Right, because either the value being shifted is zero, or the result will overflow a fixnum. 15:50:49 Okay. 15:52:46 -!- mega1 [~quassel@pool-03710.externet.hu] has quit [Ping timeout: 240 seconds] 15:59:49 mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has joined #sbcl 16:22:18 -!- mbohun [~mbohun@ppp115-156.static.internode.on.net] has quit [Quit: Leaving] 16:33:28 -!- hargettp [~anonymous@pool-71-174-128-10.bstnma.east.verizon.net] has quit [Quit: hargettp] 16:38:22 -!- attila_lendvai [~attila_le@adsl-89-132-3-59.monradsl.monornet.hu] has left #sbcl 08:01:13 ccl-logbot [~ccl-logbo@setf.clozure.com] has joined #sbcl 08:01:13 -!- names: ccl-logbot attila_lendvai flip214 rbarraud Kaer Fare tsuru mega1_ cmm lnostdal froydnj fe[nl]ix _3b` gnooth lisppaste2 angavrilov deepfire specbot minion pkhuong slyrus__ chandler @Xof foom vs slyrus_ jsnell joshe tokenrove 08:07:08 -!- flip214 [~marek@unaffiliated/flip214] has quit [*.net *.split] 08:07:08 -!- lisppaste2 [~lisppaste@common-lisp.net] has quit [*.net *.split] 08:07:08 -!- minion [~minion@common-lisp.net] has quit [*.net *.split] 08:07:08 -!- attila_lendvai [~attila_le@catv-89-133-169-124.catv.broadband.hu] has quit [*.net *.split] 08:07:09 -!- Kaer [b@c-6bcfe253.97-16-64736c12.cust.bredbandsbolaget.se] has quit [*.net *.split] 08:07:09 -!- froydnj [~froydnj@gateway.codesourcery.com] has quit [*.net *.split] 08:07:09 -!- tsuru [~charlie@c-174-50-217-160.hsd1.tn.comcast.net] has quit [*.net *.split] 08:07:10 -!- gnooth [~test@ip98-176-79-151.sd.sd.cox.net] has quit [*.net *.split] 08:07:12 -!- slyrus_ [~chatzilla@adsl-75-36-215-204.dsl.pltn13.sbcglobal.net] has quit [*.net *.split] 08:07:12 -!- mega1_ [~quassel@catv4E5CABA2.pool.t-online.hu] has quit [*.net *.split] 08:07:14 -!- Xof [~crhodes@158.223.51.79] has quit [*.net *.split] 08:07:14 -!- tokenrove [~julian@miranda.org] has quit [*.net *.split] 08:07:14 -!- lnostdal [~Lars@167.80-203-136.nextgentel.com] has quit [*.net *.split] 08:07:15 -!- deepfire [~deepfire@80.92.100.69] has quit [*.net *.split] 08:59:31 flip214 [~marek@unaffiliated/flip214] has joined #sbcl 08:59:31 Kaer [b@c-6bcfe253.97-16-64736c12.cust.bredbandsbolaget.se] has joined #sbcl 08:59:31 tsuru [~charlie@c-174-50-217-160.hsd1.tn.comcast.net] has joined #sbcl 08:59:31 mega1_ [~quassel@catv4E5CABA2.pool.t-online.hu] has joined #sbcl 08:59:31 lnostdal [~Lars@167.80-203-136.nextgentel.com] has joined #sbcl 08:59:31 froydnj [~froydnj@gateway.codesourcery.com] has joined #sbcl 08:59:31 gnooth [~test@ip98-176-79-151.sd.sd.cox.net] has joined #sbcl 08:59:31 lisppaste2 [~lisppaste@common-lisp.net] has joined #sbcl 08:59:31 deepfire [~deepfire@80.92.100.69] has joined #sbcl 08:59:31 minion [~minion@common-lisp.net] has joined #sbcl 08:59:31 Xof [~crhodes@158.223.51.79] has joined #sbcl 08:59:31 slyrus_ [~chatzilla@adsl-75-36-215-204.dsl.pltn13.sbcglobal.net] has joined #sbcl 08:59:31 tokenrove [~julian@miranda.org] has joined #sbcl 08:59:31 -!- barjavel.freenode.net has set mode +o Xof 09:07:26 nikodemus [~nikodemus@cs181058025.pp.htv.fi] has joined #sbcl 09:07:26 -!- ChanServ has set mode +o nikodemus 09:17:06 good morning 09:21:28 -!- mega1_ [~quassel@catv4E5CABA2.pool.t-online.hu] has quit [Ping timeout: 240 seconds] 09:22:38 mega1 [~quassel@catv4E5CABA2.pool.t-online.hu] has joined #sbcl 09:57:04 tcr [~tcr@cpc5-bour5-2-0-cust340.15-1.cable.virginmedia.com] has joined #sbcl 10:42:34 Krystof [~csr21@cpc5-bour5-2-0-cust340.15-1.cable.virginmedia.com] has joined #sbcl 10:42:34 -!- ChanServ has set mode +o Krystof 11:10:14 mbohun [~mbohun@ppp115-156.static.internode.on.net] has joined #sbcl 12:06:29 Blkt [~user@dynamic-adsl-94-37-236-50.clienti.tiscali.it] has joined #sbcl 12:17:43 good day everyone 12:26:40 stassats [~stassats@wikipedia/stassats] has joined #sbcl 12:30:21 nyef [~nyef@pool-64-222-178-106.man.east.myfairpoint.net] has joined #sbcl 12:30:28 G'morning all. 12:32:03 hi nyef 12:36:02 blabla [~lisps@stargate.fs.uni-lj.si] has joined #sbcl 13:59:28 Feedback request time. Does http://repo.or.cz/w/sbcl/nyef.git/commitdiff/545af54ff8ac6bb2298aedb23e2245e78b4baa06 look good/bad to anyone? 14:00:02 (It builds, for both x86oids, but that's really all I can say about it. I certainly didn't performance-test it.)