00:08:16 -!- tcr [~tcr@cpc1-bour2-0-0-cust414.15-1.cable.virginmedia.com] has quit [Quit: Leaving.] 00:17:30 -!- pkhuong_ is now known as pkhuong 00:25:39 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 00:43:36 -!- attila_lendvai [~attila_le@catv-89-133-171-82.catv.broadband.hu] has quit [Quit: Leaving.] 00:55:37 -!- hargettp [~anonymous@pool-71-174-140-140.bstnma.east.verizon.net] has quit [Quit: hargettp] 01:28:23 you should be able to do ultra fast octet matching with the insns already defined on x86-64, no? 01:28:34 blah, tcr is no longer here 01:29:14 *_3b`* votes to do it on a gpu :p 01:29:35 opencl ;) 01:30:35 *_3b`* would probably use cuda, just since that seems to have more CL bindings currently (and i have mostly nv hardware at the moment) 01:30:37 froydnj: yup, you should. 01:31:00 although, a SWAR version would also be doable. 01:34:29 -!- Blkt [~user@dynamic-adsl-94-34-31-251.clienti.tiscali.it] has quit [Quit: Error: do not makunbound t please] 01:34:39 <_3b`> hmm, now it's bugging trying to remember how that paper was talking about doing parsing in parallel on connection machines 01:35:49 *_3b`* really needs to figure out some better way to keep track of random papers i download... rename them to their title instead of 'paper.pdf' or whatever, or make an index or something 01:36:06 _3b`: I used to use gmail for that 01:36:09 now I use papers 01:36:28 with its semi automated matching function, entering bibliographical data takes an instant. 01:36:40 _3b`: regex parsing, right? 01:37:32 I guess CYK could be parallelised? 01:37:50 <_3b`> not sure what i'm remembering was regex specifically, but probably comparable if not 01:38:38 <_3b`> might have just been an overview in a paper with a bunch of unrelated algorithms 01:46:22 <_3b`> ah, looks like it was in 'data parallel algorithms' by hillis and steele 01:47:07 <_3b`> though google finds a few papers just about parsing on CM, but don't see any downloads for those 01:50:43 I read a lot fewer papers lately, not having access to the ACM DL :( 01:52:22 pkhuong: what is this "papers" program? 01:55:31 froydnj: the first hit on google, . 01:55:59 pkhuong: ah, thanks. boo, mac 01:59:40 -!- nyef [~nyef@pool-70-109-134-127.cncdnh.east.myfairpoint.net] has quit [Quit: G'night all.] 03:29:10 rbarraud_ [~rbarraud@118-93-72-48.dsl.dyn.ihug.co.nz] has joined #sbcl 03:29:16 -!- rbarraud [~rbarraud@118-93-72-48.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 272 seconds] 03:50:41 rbarraud__ [~rbarraud@118-93-72-48.dsl.dyn.ihug.co.nz] has joined #sbcl 03:51:47 -!- rbarraud_ [~rbarraud@118-93-72-48.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 265 seconds] 04:06:53 -!- slyrus [~chatzilla@adsl-75-36-215-204.dsl.pltn13.sbcglobal.net] has quit [Remote host closed the connection] 04:14:31 slyrus [~chatzilla@adsl-75-36-215-204.dsl.pltn13.sbcglobal.net] has joined #sbcl 04:16:48 -!- rbarraud__ [~rbarraud@118-93-72-48.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 240 seconds] 05:00:56 rbarraud__ [~rbarraud@121.90.24.151] has joined #sbcl 05:43:50 -!- rbarraud__ [~rbarraud@121.90.24.151] has quit [Ping timeout: 264 seconds] 06:45:03 tcr [~tcr@cpc1-bour2-0-0-cust414.15-1.cable.virginmedia.com] has joined #sbcl 06:47:58 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 06:47:58 -!- ChanServ has set mode +o nikodemus 06:56:21 rbarraud__ [~rbarraud@118-93-163-163.dsl.dyn.ihug.co.nz] has joined #sbcl 07:20:17 -!- rbarraud__ [~rbarraud@118-93-163-163.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 276 seconds] 07:50:04 -!- Krystof [~csr21@84-51-132-95.christ977.adsl.metronet.co.uk] has quit [Ping timeout: 265 seconds] 08:18:43 attila_lendvai [~attila_le@catv-89-133-171-82.catv.broadband.hu] has joined #sbcl 08:37:54 tcr: re sse: fyi, angavrilov is also working on it 09:07:09 rbarraud [~rbarraud@118-93-163-163.dsl.dyn.ihug.co.nz] has joined #sbcl 09:37:48 Krystof [~csr21@158.223.51.76] has joined #sbcl 09:37:48 -!- ChanServ has set mode +o Krystof 10:12:31 hargettp [~anonymous@pool-71-174-140-140.bstnma.east.verizon.net] has joined #sbcl 10:22:57 stassats` [~stassats@wikipedia/stassats] has joined #sbcl 10:47:27 -!- Krystof [~csr21@158.223.51.76] has quit [Ping timeout: 265 seconds] 11:18:29 Krystof [~csr21@158.223.51.76] has joined #sbcl 11:18:29 -!- ChanServ has set mode +o Krystof 11:31:19 -!- rbarraud [~rbarraud@118-93-163-163.dsl.dyn.ihug.co.nz] has quit [Ping timeout: 240 seconds] 11:33:10 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: This computer has gone to sleep] 11:35:17 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 11:35:17 -!- ChanServ has set mode +o nikodemus 11:36:56 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Client Quit] 11:42:00 -!- stassats` [~stassats@wikipedia/stassats] has quit [Ping timeout: 272 seconds] 11:54:15 -!- hargettp [~anonymous@pool-71-174-140-140.bstnma.east.verizon.net] has quit [Quit: hargettp] 11:57:06 hargettp [~anonymous@pool-71-174-140-140.bstnma.east.verizon.net] has joined #sbcl 12:13:25 -!- _3b` [foobar@cpe-72-179-19-4.austin.res.rr.com] has quit [Ping timeout: 252 seconds] 12:14:51 _3b` [foobar@cpe-72-179-19-4.austin.res.rr.com] has joined #sbcl 12:46:09 I wonder what other CL implementations' compilers can conceivably support SSE intrinsics... 13:07:27 -!- Krystof [~csr21@158.223.51.76] has quit [Read error: Operation timed out] 13:09:48 -!- attila_lendvai [~attila_le@catv-89-133-171-82.catv.broadband.hu] has quit [Quit: Leaving.] 13:14:41 CCL could probably be adapted 13:32:07 Does it already have register allocation for xmm? When I last looked it didn't appear to unbox FP computations. 13:35:29 What does returning a closure boil down to at runtime? 13:39:27 Sorry, I actually want to know what's done to allocate a closure 13:40:09 Basically allocating a chunk of memory with the appropriate header, pointing to the preallocated code component, filling in values in the constant table? 13:49:29 in the constant table? 13:50:06 stassats [~stassats@wikipedia/stassats] has joined #sbcl 13:50:06 anyway, yeah: chunk of memory, header - code pointer - n boxed values (or value cells for mutable bindings) 13:50:09 and that's it 13:54:12 what about returning to an outside block? 14:00:25 How does sbcl represent objects of type simple-array? 14:01:26 angavrilov: I'd guess it's in src/compiler/generic/objdef.lisp 14:06:45 Do one-dimensional arrays have an indirect header, or are they represented as raw vectors? 14:07:32 Currently my sbcl sse contrib is missing only 3 obscure intrinsics and support for arrays, compared to the ecl version. 14:23:14 angavrilov: one-dimensional non-simple arrays (fill-pointer/displacement/etc.) have an indirect header 14:23:14 froydnj, memo from gigamonkey: It's moved to http://gigamonkeys.wordpress.com/ (yeah, yeah, I should set up redirects.) 14:23:25 ooo, minion is cross-channel 14:26:32 Is there any heuristics in sbcl that governs which branch is emitted right next to the test instruction? 14:27:20 like which block gets the fallthrue? 14:27:42 yes 14:28:11 I don't think so. there's a bb reordering for loops, but that's it 14:29:11 I mean I want to tell it somehow 14:29:32 at the moment it makes a block the fallthrough which will almost never be executed 14:30:05 froydnj: But since non-simple may imply a displacement chain, they are not very good for performance. So I guess that make-sse-array would have to disallow displacement offset for the one-dimensional shape. 15:13:23 nikodemus [~nikodemus@cs181058025.pp.htv.fi] has joined #sbcl 15:13:23 -!- ChanServ has set mode +o nikodemus 15:32:42 tcr: well yeah, there is one, but no repair afterwards 15:33:45 not for branch prediction purposes, but the DFO thingy does reduce the number of unconditional branches, iirc. 15:33:55 but, again, no jump tensioning to avoid jmp to jmp 15:34:24 angavrilov: you can just use with-array-data, with-pinned-object and vector-sap 17:36:28 Krystof [~csr21@84-51-132-95.christ977.adsl.metronet.co.uk] has joined #sbcl 17:36:28 -!- ChanServ has set mode +o Krystof 17:44:08 -!- nikodemus [~nikodemus@cs181058025.pp.htv.fi] has quit [Quit: Leaving] 18:54:37 nikodemus [~nikodemus@cs181199216.pp.htv.fi] has joined #sbcl 18:54:37 -!- ChanServ has set mode +o nikodemus 18:56:18 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 272 seconds] 19:35:43 stassats [~stassats@wikipedia/stassats] has joined #sbcl 19:38:33 -!- hargettp [~anonymous@pool-71-174-140-140.bstnma.east.verizon.net] has quit [Quit: hargettp] 19:53:00 hargettp [~anonymous@pool-71-174-140-140.bstnma.east.verizon.net] has joined #sbcl 19:56:34 -!- hargettp [~anonymous@pool-71-174-140-140.bstnma.east.verizon.net] has quit [Client Quit] 20:27:54 Blkt [~user@dynamic-adsl-94-34-31-251.clienti.tiscali.it] has joined #sbcl 20:29:31 cmm- [~cmm@bzq-79-181-202-240.red.bezeqint.net] has joined #sbcl 20:32:47 -!- cmm [~cmm@bzq-79-180-201-1.red.bezeqint.net] has quit [Ping timeout: 276 seconds] 20:44:22 -!- nikodemus [~nikodemus@cs181199216.pp.htv.fi] has quit [Quit: Leaving] 21:02:18 -!- Blkt [~user@dynamic-adsl-94-34-31-251.clienti.tiscali.it] has quit [Read error: Connection reset by peer] 21:31:06 kclifton [~kclifton@s198-166-45-245.ab.hsia.telus.net] has joined #sbcl 21:36:40 rbarraud [~rbarraud@118-93-163-163.dsl.dyn.ihug.co.nz] has joined #sbcl 22:33:55 -!- stassats [~stassats@wikipedia/stassats] has quit [Ping timeout: 240 seconds] 22:58:20 stassats [~stassats@wikipedia/stassats] has joined #sbcl 23:16:31 -!- tcr [~tcr@cpc1-bour2-0-0-cust414.15-1.cable.virginmedia.com] has quit [Quit: Leaving.]