00:00:00 --- log: started forth/08.02.20 00:51:51 JasonWoof: consider that on modern processors the issue of accessing memory and keeping it in cache is your critical path - it becomes an interesting idea. following pointers around all over the place causes cache spills and is why these processors don't perform well in standard stack oriented software. 00:53:21 ahh. hadn't thought of that 00:55:05 proteusguy: otoh, if you write code to other places, you usually have to flush the caches (because of separate i/d caches) 00:55:37 --- quit: nighty^ ("Disappears in a puff of smoke") 00:57:12 oxygene: i'm just repeating what the paper said. it's intriguing but i'm curious if anyone has actually tried it and seen it work. i think their point is that routines are typically very short so the cost of moving the full routing into the stack is less than pushing its address to the stack then dereferencing it to execute. 00:58:57 actually with the relatively large L1 & L2 caches now, it's conceivable to fit your entire forth environment in cache. get your memory management down right and you should have execution speeds that far exceed the typical compiled C app. 01:00:07 yeah, I have a hard time imagining that I'd write more than 1MB of code 01:00:42 I just realized that if you have the return stack grow down you could conceivably just execute it 01:00:55 seems like a lot of extra moving memory around 01:01:31 there aught to be a way to make it so your often-run routines don't fall out of the cache too much 01:01:39 it does but i think its a consideration that moving one pointer address may cost the same as n words so might as well move those n words. 01:02:54 JasonWoof: doesn't appear cpu designers want to allow the programmers to hint too hard so just gotta profile a lot and learn how the system behaves. 01:03:24 And for just about any language except forth - it's probably the right decision. 01:03:32 hopefully if you call a word 4,000 times per second it's not going to spend much time out of the cache 01:03:44 and if it falls out then you weren't using it much, so it's ok 01:03:50 seems like a non-problem to me 01:04:59 yeah i think you're generally fine. memory arrangement, page boundaries and such are where you need more manual intervention. 01:05:10 forth isn't designed for being optomized well 01:06:19 my view of forth was that you could meaningfully optimize exactly for your environment and should. 01:07:08 yeah, but not on the level of page boundaries, instruction reordering and such 01:07:29 it's more that when building from scratch you can make the program well suited to its environment 01:07:31 why not? you actually put your code into explicit pages. 01:08:29 to me it seems to be the forth way to keep it simple 01:08:43 when the compiler is kept simple, it can be taken advantage of 01:09:41 example: you can optomize to some degree, by skipping the return stack push/pop on leaf functions 01:09:48 (ie functions which don't call any other functions) 01:10:31 forth often does _not_ do this. 01:11:25 but in not doing that, it keeps things simple, and programmers often find a way of taking advantage of the rstack always having a reference to the current word 01:12:20 rstack manipulation can do some very cool stuff 01:14:06 yep indeed. its just frustrating that "normal" cpus suck at stack-based architectures performance-wise. 01:16:07 : nth-execute: 4* r> + @ execute ; 01:16:09 something like that 01:17:02 : handle-0,1,2,3 ( x -- ) nth-execute: zero one two three ; 01:17:58 well, yeah, modern big CPUs aren't optomized for stack-based stuff 01:18:22 frequent branching/calling is tough on the pipelines 01:19:07 perhaps chip designers will put more emphasis on optomizing for stack stuff though, since many interpreted languages work on a stack vm 01:19:22 and those are the ones that could use some more speed :) 01:31:12 they usually get JITed, and lose the stack characteristics on the way these days 01:40:13 so why not do that with forth? 01:41:05 well, AOT (ahead of time) might be more sensible - and I think that's exactly what the commercial forths are doing 01:42:08 the free ones stick to sentiments like "we ought to be portable to all 12 architectures linux supports, and better not treat one better than the other" - or maybe it's just lack of manpower and focus 03:12:34 --- join: nighty^ (n=nighty@p3047-adsau16honb13-acca.tokyo.ocn.ne.jp) joined #forth 05:42:11 --- join: forther (n=forther@c-67-180-150-67.hsd1.ca.comcast.net) joined #forth 05:49:42 --- join: Al2O3_ (n=Al2O3@c-76-120-54-133.hsd1.co.comcast.net) joined #forth 05:56:02 --- quit: Al2O3 (Connection timed out) 06:08:16 --- join: edrx (i=edrx@189.25.14.62) joined #forth 07:09:48 --- quit: nighty^ ("Disappears in a puff of smoke") 07:10:05 --- join: nighty^ (n=nighty@p3047-adsau16honb13-acca.tokyo.ocn.ne.jp) joined #forth 07:15:51 --- part: edrx left #forth 07:32:30 --- quit: nighty^ (Client Quit) 07:38:18 --- quit: ramkrsna ("Leaving") 08:01:12 --- quit: ygrek ("Leaving") 08:01:33 --- join: ygrek (i=user@gateway/tor/x-43ac22e859bd6188) joined #forth 08:38:36 --- join: tathi (n=josh@pdpc/supporter/bronze/tathi) joined #forth 08:38:36 --- mode: ChanServ set +o tathi 08:50:03 --- quit: forther ("Leaving") 09:01:35 --- join: Quartus__ (n=Quartus_@205.205.50.1) joined #forth 10:08:04 --- join: forther (n=forther@207.47.34.100.static.nextweb.net) joined #forth 10:08:30 Hi 10:10:01 someone on SEAforth forum mentioned, that on some #forth channel SEAforth chip is known as "vaporware". I wonder was it this channel ... 10:32:28 --- quit: sauvin (Remote closed the connection) 10:44:24 --- join: arke (n=arke@p57A746D0.dip.t-dialin.net) joined #forth 10:44:24 --- mode: ChanServ set +o arke 10:47:51 forther: that is a generic term... 10:56:51 SEAforth chips look like the heaviest elements on the periodic chart: known to exist, but damned hard to get a hold of. 11:10:21 hrm, first google hit on seaforth is for a sportfishing store in san diego 11:20:32 forther: well, the web indicates there are prototypes on silicon already, with some releases expected this year 11:20:46 looks mighty useful to me, so I'll keep my eyes out 11:40:21 well, it IS out. you can read about it on http://www.intellasys.net 11:41:15 if you can't order it for delivery, it's not out. 11:41:17 if you request the tools, docs and the development board my e-mail you may get one 11:42:52 begging for prototypes does not equal a released product. 11:45:22 how else will we know were to send it 11:46:22 ...? It's not released. The last word is that you have to apply and be considered. 11:47:05 if you want to call that 'out', you're using your own private dictionary. 11:48:50 And IIRC they were originally claiming that they were going to have chips available for sale last January. 11:49:57 probably have it in production, but too low yield to be viable yet 11:50:32 bga, hmmm 11:50:42 I have no interest in begging to be allowed to fiddle with a prototype of a product that may never achieve actual release. 11:50:56 They reportedly have one built into a USB drive. 11:51:16 it looks interesting, but i certainly wouldn't make plans around it 11:51:21 But yeah, they're a long way from achieving actual release, so I'll wait to be actually interested until they have an actual product. :) 11:52:24 thinking about requesting a data sheet though 11:53:24 watch out, they may not consider you worthy! 11:53:35 for a datasheet? heh 11:54:10 well, requesting a datasheet is tiny effort. If they don't send one, then they solved the problem of evaluating it all that much faster 11:54:53 it rather raises the question of why the datasheet isn't a downloadable pdf, right on the home page. 11:54:58 n/m, it IS 11:55:03 though you have to go into the forums 11:55:15 http://www.intellasys.net/phpBB/viewtopic.php?f=6&t=62 11:55:31 curious 11:55:50 datasheet and tools (simulator, etc.) are sent to everyone who asked 11:57:23 hm, the adcs are 18bit, but the dacs are 6bit... 11:59:05 you can build PWM and have up to 18 bits @ 96 kHz 11:59:53 which would be more than adequate 12:00:06 one of the application in the package is music synthesizer. You can learn how it is could be done from it 12:03:34 oh ... we just released the datasheet fo unconditional access. to get the tools you still have to send an e-mail 12:08:42 and who are you at intellisys, forther? 12:11:06 forther. why? 12:11:56 Nicks contain less info than names. 12:12:01 no, I mean who are you? You imply you're part of the organization. 12:12:21 yes, I'm working in there. I'm a programmer 12:13:06 who is unwilling to identify himself, yet sets out to do PR. 12:14:01 heh, take off yer foilhat and relax 12:14:25 what difference it would make? 12:14:32 I'm not tense; just struck by the unprofessionalism. 12:15:03 it doesn't inspire confidence. 12:15:23 hrm, it's a different way of doing business, but fairly normal for many newer companies 12:15:38 not the ones I deal with. 12:15:40 forther: if we don't know your name, we can't validate your employment. If we can't know you work there, the things you say about it carry less weight. 12:16:04 Where less might go to zero. 12:16:05 I thought it was gavin from the first couple comments, actually. :) 12:16:17 But he wouldn't have bothered actually reading the web site... 12:16:23 heh 12:17:50 forther: well in any case thanks for pointing it out 12:18:08 you are welcome! 12:20:13 forther: if the tools were available for download, instead of people having to email some guy who tells some guy who asks some other guy for permission who might email them to you, then we'd say the tools are "out" 12:20:57 but in either case the chips remain "vaporware" until i can go to the website and order them in bulk quantity and expect to receive a shipment in a timely manner 12:21:56 "vaporware" = something that's been announced but is not available to the general public 12:28:52 No, "never will be available" is a key part of "vaporware" definition 12:28:55 no. 12:28:57 that's not true. 12:29:10 paul graham's 'arc' was vaporware for 6 years, then he finally released a trivial lisp interpreter named 'arc'. 12:29:22 http://www.google.com/search?client=safari&rls=en&q=define:+vaporware&ie=UTF-8&oe=UTF-8 12:29:39 arc was more useful when it Was vaporware 12:29:43 most of the definitions have that "never" part 12:30:14 lucca: heh 12:30:14 forther: customers can't tell the difference until it is in their hands 12:30:37 a big part of vaporware is exxagerated claims 12:30:39 this one may also qualify for 'never', at any rate 12:30:58 http://en.wikipedia.org/wiki/Vaporware#Overambitious_hype 12:31:05 jeff fox goes on c.l.f and starts ranting about how evil C programmers will bow down before SEAforth which is 100000 times faster and 1000000 times more efficient and 100000000 less power-hungry than a "pentium" 12:31:17 that's really bad PR from my point of view 12:31:35 well, it's jeff fox. 12:31:47 if you guys tone down the rhetoric a bit and focus on getting your products out of the door and communicating exactly what's going in w.r.t the development process, people would take you a lot more seriously 12:31:49 maybe 'forther' is fox, too? 12:32:23 Quartus__: that would be plain wierd. 12:32:43 see above. 12:33:03 a downloadable simulator and dev tools are about 10000x more valuable than newsgroup postings, or coming into an IRC channel to dispute the definiton of 'vaporware', IMO 12:33:11 but I could be wrong; I don't own a semicondutor design business 12:33:26 forther: for me, that url you posted brings up a list of seven definitions, five of which don't include 'never' as a necessary part of the definition. 12:33:39 though many of them say "MAY never be released", or something of that sort 12:33:42 I'm just speaking as ap otential customer; I'd loe to get my hands on a forth chip or 20. 12:33:51 potential, love* 12:37:02 --- join: snoopy_1711 (i=snoopy_1@dslb-084-059-101-166.pools.arcor-ip.net) joined #forth 12:39:16 unfortunately, the only way to get the real chip is to send the e-mail to forth@intellasys.net I wish it was available for general public, but it is not there yet 12:39:43 so, vapourware. Not released. Not out. 12:41:00 what's the difference between your "released" and "out" ? 12:41:00 it means that Joe Sixpack can go and order the chips 12:41:04 and it's not simply a matter of sending an email; you have to shine the apple for teacher, and hope you're a winner. 12:41:07 could something be "out" and not "released"? 12:41:09 and receive them in a timely manner together with tools and docs 12:41:10 and start coding 12:42:01 i don't care if its out or released or whatever 12:42:12 because i can't do anything with it until i can order a chip and receive it 12:42:33 point is, you can't bloody buy it yet! 12:43:03 Quartus__: there's also companies which offer products for sale but don't ship them immediately... 12:43:06 well ... we distribute it for free. not for everyone just yet 12:43:27 Quartus__: oh I see. you have to be part of an elite club. We're just Outsiders, not worthy at all. 12:43:43 I consider that an asinine method of doing business, frankly. 12:43:44 Quartus__: its not like we could code anything decent anyway. We're just C programmers in disguise. 12:44:03 right. ANSI-coders. 12:44:33 I hear werty is taking orders for NewForth. 12:44:41 it will be out any day now; after all, he just has to sit down for 2 days and code i. 12:44:46 and you can knock off the 'we' prattle, Forther, if you don't have the stones to name yourself. 12:44:52 --- quit: Snoopy42 (Read error: 145 (Connection timed out)) 12:45:02 --- nick: snoopy_1711 -> Snoopy42 12:45:46 My name is Michael Montvelishsky. I don't see how it will make you happy 12:46:11 and in what regard do you speak for Intellisys? 12:46:36 ...I think we're creating a great deal of unnecessary drama here 12:46:51 I don't. 12:47:11 why does he Need to be speaking in any capacity other than someone who is interested in these things? 12:47:38 because he claims to be speaking for the company! 12:47:49 nope 12:48:10 when you say 'we', we mentally substitute 'IntellaSys', at which point you are speaking for your employer 12:48:21 ... I don't agree with that 12:48:25 unless by 'we' you mean 'me, my teddy bear and imaginary friend' 12:48:27 I do. 12:48:50 I guess it's a small startup company thing 12:49:16 tell you what, I'll forward the log to intellisys, see what they think. 12:49:19 which may further color your views of them, but not all of us consider it a bad thing to hear what non-PR employees think of their work 12:50:28 I'm not sure how you think that will benefit anyone, but okay. 12:56:15 forther: didn't you write the synthesizer in the tools package? 12:57:09 not me, the other guy did. I just did some tiny parts of it 12:57:42 Huh. But you demoed it at the November Forth Day? 12:58:34 Do you have anything that you have worked on that you can talk about? 12:58:35 Yes, I did 12:59:27 are you sure, that this channel is good place to talk about it? 13:00:28 Just curious. Most of my objection with what you've said so far is that it's kind of pointless. 13:00:54 I mean, anyone who has spent five minutes googling SEAForth and Intellasys already knows everything you've said so far. 13:01:15 I'd be curious to hear about someone's experience actually using the thing. 13:01:30 ooh, me too 13:01:33 go ahead, ask 13:01:50 what's it like using the thing? 13:02:09 is it mind-boggling or useful to have several cores at your disposal? 13:02:28 JasonWoof: ever played with erlang? 13:02:54 parallel programming is very different 13:03:18 do the cores execute in lock-step off the same clock? 13:03:28 no, it's asynchronous all the way, isn't it? 13:03:44 right, asynchronous 13:04:09 Actually you can execute off a stream of code coming from an adjacent core, I believe. 13:04:38 yes, and it is very useful feature 13:04:57 and there are two reset pins, so they don't necessarily have to all reset at the same time 13:04:59 what kind of synchronization and communication primitives do you use? 13:05:48 our port are very close to CSP channels 13:06:10 lots of power pins... can you have parts of it powered on/off independently? 13:06:42 sleeping node virtually doesn't consume 13:07:58 we better do not substitute actual manual with IRC talk (which tend to get emotional) 13:08:37 datasheet is on http://www.intellasys.net/templates/trial/content/SEAforth-24A_Data_Sheet.pdf 13:09:32 You just don't have a clue, do you? 13:09:38 Oh well, back to coding 13:14:08 I think we banned the people that randomly get emotional 13:14:28 three, to date 13:18:32 i wouldn't call gavino 'emotional'. 13:18:59 That's not his predominant characteristic, true. 13:38:07 --- quit: ygrek (Remote closed the connection) 13:38:19 --- join: ygrek (i=user@gateway/tor/x-441f5361fead865c) joined #forth 14:06:27 --- join: LOOP-HOG (n=jasondam@c-76-105-172-75.hsd1.or.comcast.net) joined #forth 14:15:13 --- quit: ygrek (Remote closed the connection) 14:15:29 --- join: snoopy_1711 (i=snoopy_1@dslb-084-059-120-233.pools.arcor-ip.net) joined #forth 14:22:39 --- quit: Snoopy42 (Read error: 145 (Connection timed out)) 14:22:43 --- join: Snoopy42 (i=snoopy_1@dslb-088-068-203-195.pools.arcor-ip.net) joined #forth 14:28:29 --- quit: snoopy_1711 (Read error: 145 (Connection timed out)) 14:31:58 --- join: foxchip (n=fox@207.47.34.100.static.nextweb.net) joined #forth 14:38:32 hi foxchip 14:40:12 I think that you already answered your own questions, but I may post the transcript date for others to read 14:47:19 --- join: snoopy_1711 (i=snoopy_1@dslb-088-068-194-026.pools.arcor-ip.net) joined #forth 14:47:42 I was not I but thank you anyway. 14:47:54 IT was not I but thank you anyway. 14:48:17 I just did 14:48:39 I am still excited about the chip even though I wouldn't know how the heck to use it 14:48:58 I can imagine that it would be useful in space based apps if rad hardened 14:50:05 there have also been advances in solar power material which make RTG units less needed in the outer solar system 14:50:40 maybe I should grab a book on basic electronics 14:55:18 --- quit: Snoopy42 (Read error: 145 (Connection timed out)) 14:55:38 --- nick: snoopy_1711 -> Snoopy42 15:00:59 --- join: crest_ (n=crest@p5B104F01.dip.t-dialin.net) joined #forth 15:08:40 --- quit: LOOP-HOG () 15:11:27 man, it's hard not to go through my crap 15:13:00 I'm moving, and determined not to take those boxes of stuff I've hardly opened since three moves ago 15:17:43 --- quit: crest__ (Read error: 110 (Connection timed out)) 15:29:38 --- part: forther left #forth 15:34:46 --- join: forther (n=forther@207.47.34.100.static.nextweb.net) joined #forth 16:15:15 --- join: nighty^ (n=nighty@210.188.173.246) joined #forth 16:26:38 JasonWoof: What is the url to your javascript shadow? 16:33:35 Nevermind, found it. 16:38:31 --- quit: foxchip () 16:50:38 --- quit: proteusguy (Connection timed out) 16:51:34 --- join: proteusguy (n=proteusg@ppp-124-120-223-238.revip2.asianet.co.th) joined #forth 17:00:36 woh, people actually looking at my work 17:05:28 * crc watches some of your stuff 17:06:28 heh, glad Deformative found it, I don't know where I put it 17:06:57 JasonWoof: It was at home.jasonwoof.org:2201/html 17:07:03 http://home.jasonwoof.org:2201/html/shadow/ 17:07:05 right 17:07:12 don't see it on my real website 17:07:27 home.jasonwoof.org won't be up much longer 17:07:38 I am working on an animated site. 17:07:46 Not like your shadow, but I wanted to see how you did it. 17:07:53 cool 17:08:00 I would paste it here, but it has some info on it, and I am too lazy to remove the info. 17:08:13 eek! info! 17:08:32 info = other peoples contact information and I do not know if they would be man. 17:08:37 I think I need to add a section to my website for my projects 17:08:40 mad rather 17:08:42 ahh 17:09:13 I'm starting to have web-based tools that are useful 17:09:47 right now my css font picker is ready 17:09:52 and some things are close, like the shadow thing 17:10:28 --- quit: tathi ("leaving") 17:11:36 I've got a pretty sophisticated thing for auto-generating code for processing forms 17:11:56 it makes the html template, the php code, the sql code, even an e-mail template 17:12:06 it is forever not finished though 17:12:20 I've added all kinds of stuff, made it more flexible, etc 17:12:24 but it's always rough around the edges 17:12:40 saves me loads of time, but I don't feel it's even close to ready to take over the world 17:13:07 Hmm, your shadow depends on opaque images. 17:13:10 :( 17:16:28 yep 17:16:39 doesn't work otherwise 17:17:30 I plan to build something where you give the background color and the width and it generates the opaque images for you. 17:18:08 well, I did build such a thing mostly, but before I finished I realized that the program I was using to render won't run on my server 17:18:15 (it requires gnome) 17:18:28 I had to do nasty things to get it to run locally 17:18:44 some day I'll probably rewrite it to use imagemagick 17:19:33 to use transparent images you'd need a 5x5 table 17:33:02 Ugh, damnit. 17:33:08 Safari does not like my javascript. 17:33:12 * Deformative = pissed 17:35:12 --- join: skas (n=skas@c-24-61-243-42.hsd1.ma.comcast.net) joined #forth 18:05:29 --- join: arke_ (n=arke@p57A746D0.dip.t-dialin.net) joined #forth 18:08:30 --- quit: arke (Read error: 104 (Connection reset by peer)) 18:19:13 --- quit: forther ("Leaving") 18:19:16 --- quit: Al2O3_ (Remote closed the connection) 21:28:05 --- quit: skas ("Ex-Chat") 21:33:33 --- join: forther (n=forther@c-67-180-150-67.hsd1.ca.comcast.net) joined #forth 21:40:40 --- quit: forther ("Leaving") 22:31:51 --- quit: timlarson (kornbluth.freenode.net irc.freenode.net) 22:32:20 --- join: timlarson (n=timlarso@user-12l37rb.cable.mindspring.com) joined #forth 22:34:00 --- join: ramkrsna (n=ramkrsna@unaffiliated/ramkrsna) joined #forth 23:04:25 --- join: ramkrsna_ (n=ramkrsna@59.160.127.177) joined #forth 23:16:20 --- join: sauvin (n=sauvin@74-136-191-181.dhcp.insightbb.com) joined #forth 23:34:10 --- quit: ramkrsna (Read error: 110 (Connection timed out)) 23:45:38 --- quit: arke_ (Read error: 110 (Connection timed out)) 23:59:59 --- log: ended forth/08.02.20