00:00:00 --- log: started forth/05.05.20 00:55:34 --- join: snoopy_1711 (snoopy_161@dsl-084-058-031-060.arcor-ip.net) joined #forth 01:04:26 --- quit: Snoopy42 (Read error: 145 (Connection timed out)) 01:04:46 --- nick: snoopy_1711 -> Snoopy42 01:05:30 --- quit: Raystm2 (Read error: 131 (Connection reset by peer)) 03:37:21 --- join: Raystm2 (~vircuser@adsl-69-149-32-57.dsl.rcsntx.swbell.net) joined #forth 03:38:48 --- quit: bbls () 05:18:10 --- quit: Robert ("brb") 05:34:43 --- join: Robert (~snofs@c-f778e055.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 06:10:59 --- join: madwork (~madgarden@derby.metrics.com) joined #forth 06:11:39 Hi, madman. 06:18:26 Hi there. 07:51:15 --- join: virl (anonymous@chello062178085149.1.12.vie.surfer.at) joined #forth 09:38:06 --- join: bbls (~bbls@80.97.121.4) joined #forth 09:38:07 hello 09:38:10 Hi, bbls 09:38:16 hi Robert 10:16:51 --- join: Herkamire (~jason@c-24-218-95-147.hsd1.ma.comcast.net) joined #forth 10:16:51 --- mode: ChanServ set +o Herkamire 10:34:14 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 10:34:46 Hi, 10:35:33 yo 11:52:57 --- nick: KB1FYR -> KB1FYR|Camping 12:11:15 --- quit: bbls () 12:44:06 hmm, what are good optmizations to speed up slow dictionary lookups? 12:44:31 Hashing. 12:45:23 Maybe using fixed name lenghts to avoid repeated conditionals and use block compares. Hm... making it cache friendly. 12:45:34 Assuming you're working with a complex CPU. 12:47:19 I'm working with a RISC CPU, so it's not so complex. 12:48:19 Well, hashing is the most general I can think of. 12:48:34 Next to that, read the optimization guidelines for your CPU. 12:50:47 --- nick: Raystm2 -> nanstm 12:57:25 hmm, are there any? I don't think so. 12:58:09 What are you coding for? 12:58:11 but it is fully deterministic and so quite easy 12:58:30 for a SX 12:58:48 Ah, OK. But where is the dictionary stored? 13:00:43 in an extern memory(SRAM) writing/reading is very fast but storing a memory adress into it is damm slow, because it uses a shift register to select the memory region. 13:07:09 --- quit: virl (Read error: 104 (Connection reset by peer)) 14:00:32 fully deterministic? either i forgot what that term stands for, or that was improperly used :\ 14:01:10 As opposed to cache magic making timings semi-chaotic :) 14:01:27 ehm 14:01:36 that only happens wehn YOU code a forth... ;) 14:07:33 but i have to admit, that is one great skill ;) 14:10:31 * Robert is a little confused. 14:12:18 --- join: tathi (~josh@pcp01375108pcs.milfrd01.pa.comcast.net) joined #forth 14:46:16 --- quit: onetom ("Reconnecting") 14:46:24 --- join: onetom (~tom@ns.dunasoft.com) joined #forth 15:01:03 thats ok 15:01:06 happens to the best of us 15:01:08 :p 15:01:41 --- quit: qFox ("this quit is sponsored by somebody!") 15:06:54 --- join: virl (anonymous@chello062178085149.1.12.vie.surfer.at) joined #forth 15:26:44 yeah, especially when you're talking to qFox. :P 15:27:13 Heh. 16:27:40 --- join: Topaz (~top@sown-89.ecs.soton.ac.uk) joined #forth 16:28:23 Hi. 17:03:32 --- nick: nanstm -> Raystm2 17:06:13 hi all 17:10:34 Hi, crc. 17:10:53 * Robert has been working with the same funcion for 24 hours now. 17:11:03 function* 17:16:14 ! 17:16:17 in what language? 17:16:30 P4 assembly. 17:16:39 SSE2 and that mess? 17:16:43 Yeah. 17:16:52 Trying to draw an entire polygon at once. 17:16:53 hmhm, bed 17:16:57 Night, Topaz 17:17:00 :D 17:17:02 * crc glances at his nice, printed copies of the P4 manuals 17:17:59 --- quit: Topaz (Remote closed the connection) 17:18:34 Robert: wouldn't that be a GPU operation as opposed to a P4 one? 17:19:05 Silence! 17:19:15 no. 17:19:17 :) 17:19:34 * crc is working a glossary for retro8 17:19:39 You're right, but I do this mostly to learn about SSE2 17:59:53 * tathi is really excited about this optimizing PPC forth compiler he's designing. 18:00:06 things just keep working out. :) 18:01:10 Cool. 18:01:45 I've wanted to do this for so long, but I kept putting it off because I thought it would be too complicated and hard. 18:02:06 What kind of optimizations does it do? 18:03:00 just generates sensible asm code for a forth word, really. 18:03:19 caching top of stack elements in registers during a word. 18:03:34 combining literals with an operation where that makes sense. 18:03:47 yeah, that's about it. 18:04:16 Cool :) 18:04:32 Does it perform reasonably well at branches as well? 18:05:58 I haven't worked out anything cool to do there yet. 18:52:29 --- quit: saon (Read error: 110 (Connection timed out)) 18:52:48 --- quit: saon|smgl (Read error: 110 (Connection timed out)) 18:53:02 --- quit: tathi ("leaving") 19:07:02 --- join: Sonarman (~cleetus@adsl-64-160-164-106.dsl.snfc21.pacbell.net) joined #forth 19:43:47 --- join: saon (1000@c-66-177-224-130.hsd1.fl.comcast.net) joined #forth 19:44:43 --- join: saon|smgl (~saon@c-66-177-224-130.hsd1.fl.comcast.net) joined #forth 21:00:05 --- quit: saon ("Lost terminal") 21:00:48 --- join: saon (1000@c-66-177-224-130.hsd1.fl.comcast.net) joined #forth 21:12:18 --- quit: Raystm2 ("User pushed the X - because it's Xtra, baby") 21:24:34 --- join: Raystm2 (~vircuser@adsl-69-149-32-57.dsl.rcsntx.swbell.net) joined #forth 21:38:28 --- quit: Raystm2 ("User pushed the X - because it's Xtra, baby") 22:15:32 --- join: arke (apache@84.218.22.246) joined #forth 22:16:48 teh hies 22:44:06 --- quit: Sonarman (Read error: 145 (Connection timed out)) 23:03:28 --- join: Raystm2 (~vircuser@adsl-69-149-32-57.dsl.rcsntx.swbell.net) joined #forth 23:32:43 --- join: aum (~aum@60-234-138-239.bitstream.orcon.net.nz) joined #forth 23:35:14 --- quit: arke ("CGI:IRC") 23:37:17 --- quit: Herkamire ("off to bed") 23:59:59 --- log: ended forth/05.05.20