00:00:00 --- log: started forth/12.09.22
01:17:53 --- quit: Nisstyre-laptop (Remote host closed the connection)
01:45:05 --- join: ttmrichter_ (~ttmrichte@61.184.205.44) joined #forth
01:45:05 --- mode: ChanServ set +v ttmrichter_
01:46:41 --- quit: ttmrichter (Disconnected by services)
01:46:43 --- nick: ttmrichter_ -> ttmrichter
01:49:53 --- quit: fantazo (Remote host closed the connection)
02:48:55 --- join: fantazo (~fantazo@213.129.230.10) joined #forth
02:48:56 --- mode: ChanServ set +v fantazo
03:50:49 --- join: MayDaniel (~MayDaniel@unaffiliated/maydaniel) joined #forth
03:50:50 --- mode: ChanServ set +v MayDaniel
03:53:46 --- quit: yiyus (Ping timeout: 268 seconds)
03:54:10 --- join: yiyus (~124271242@je.je.je) joined #forth
03:54:11 --- mode: ChanServ set +v yiyus
04:30:03 --- join: ncv (~quassel@89.123.35.0) joined #forth
04:30:04 --- quit: ncv (Changing host)
04:30:04 --- join: ncv (~quassel@unaffiliated/neceve) joined #forth
04:30:04 --- mode: ChanServ set +v ncv
06:30:31 --- quit: ncv (Ping timeout: 256 seconds)
06:43:38 --- quit: ttmrichter (Read error: Connection reset by peer)
06:44:04 --- join: ttmrichter (~ttmrichte@61.183.22.204) joined #forth
06:44:05 --- mode: ChanServ set +v ttmrichter
07:31:58 --- quit: ttmrichter (Quit: i got shot at today)
07:42:21 --- join: Onionnion (~ryan@adsl-68-254-170-33.dsl.milwwi.ameritech.net) joined #forth
07:42:21 --- mode: ChanServ set +v Onionnion
07:46:53 --- join: gnooth (~gnooth@108-249-104-182.lightspeed.sndgca.sbcglobal.net) joined #forth
07:46:53 --- mode: ChanServ set +v gnooth
10:20:05 --- join: Nisstyre-laptop (~yours@oftn/member/Nisstyre) joined #forth
10:20:05 --- mode: ChanServ set +v Nisstyre-laptop
10:32:31 --- join: swiley (~glen@199.191.58.39) joined #forth
10:32:31 --- mode: ChanServ set +v swiley
10:48:54 --- join: Onionnion|Eee (~ryan@adsl-68-254-170-33.dsl.milwwi.ameritech.net) joined #forth
10:48:54 --- mode: ChanServ set +v Onionnion|Eee
10:59:49 --- quit: fantazo (Remote host closed the connection)
11:46:15 --- join: fantazo (~fantazo@91-119-67-128.dynamic.xdsl-line.inode.at) joined #forth
11:46:15 --- mode: ChanServ set +v fantazo
12:53:36 --- quit: Onionnion|Eee (Ping timeout: 246 seconds)
12:53:36 --- quit: Onionnion (Ping timeout: 246 seconds)
13:05:34 --- join: Onionnion (~ryan@adsl-68-254-162-194.dsl.milwwi.ameritech.net) joined #forth
13:05:34 --- mode: ChanServ set +v Onionnion
13:16:02 --- quit: nottwo_ (Ping timeout: 264 seconds)
13:16:26 --- join: nottwo (~trannie@nottwo.org) joined #forth
13:16:26 --- mode: ChanServ set +v nottwo
13:20:13 --- quit: Onionnion (Ping timeout: 246 seconds)
13:31:55 --- join: Onionnion (~ryan@adsl-69-215-131-78.dsl.milwwi.ameritech.net) joined #forth
13:31:55 --- mode: ChanServ set +v Onionnion
14:07:45 --- quit: Onionnion (Ping timeout: 260 seconds)
14:21:00 --- join: Onionnion (~ryan@adsl-68-254-173-230.dsl.milwwi.ameritech.net) joined #forth
14:21:01 --- mode: ChanServ set +v Onionnion
15:00:08 --- join: SHUPFS (~user@S0106001111de1fc8.cg.shawcable.net) joined #forth
15:00:08 --- mode: ChanServ set +v SHUPFS
15:27:54 --- quit: MayDaniel (Read error: Connection reset by peer)
16:32:53 --- join: RodgerTheGreat (~rodger@71-13-215-128.dhcp.mrqt.mi.charter.com) joined #forth
16:32:54 --- mode: ChanServ set +v RodgerTheGreat
18:49:41 --- join: kumul (~kumul@adsl-72-50-90-224.prtc.net) joined #forth
18:49:41 --- mode: ChanServ set +v kumul
19:25:35 --- quit: obobo (Remote host closed the connection)
19:27:59 --- join: obobo (~chatzilla@dyn-dsl-mb-216-168-118-197.nexicom.net) joined #forth
19:27:59 --- mode: ChanServ set +v obobo
21:31:42 --- join: ttmrichter (~ttmrichte@221.234.38.61) joined #forth
21:31:42 --- mode: ChanServ set +v ttmrichter
21:54:59 --- join: KipIngram (~KipIngram@85.17.255.186) joined #forth
21:54:59 --- mode: ChanServ set +v KipIngram
21:55:13 <KipIngram> Hey hey.  Been a while...
21:55:32 <RodgerTheGreat> howdy
21:56:01 <KipIngram> Howdy.  Whazzup?
21:56:25 <RodgerTheGreat> just writing a video game. Usual weekend stuff.
21:56:29 <RodgerTheGreat> you?
21:57:58 <KipIngram> Waiting to get sleepy enough to sleep.  Saw "End of Watch" earlier tonight.
21:58:08 <RodgerTheGreat> cool
21:58:27 <KipIngram> It was pretty good.  Not quite what I expected, but alright.
21:58:43 <RodgerTheGreat> in what way did it deviate from your expectations?
21:58:51 <KipIngram> Member list looks different from when I was here in the pat.
21:58:53 <KipIngram> past
21:59:27 <KipIngram> Hmmm.
21:59:28 <RodgerTheGreat> people seem to drift in and out
22:00:26 <KipIngram> I was expecting it to be "major" in terms of the drug war - turns out it was more a "sidebar battle" in the drug war.  Mostly about the lives of those few people - not the big picture.
22:00:35 <KipIngram> That didn't make it bad - just not what I expected.
22:01:46 <RodgerTheGreat> mm
22:02:02 <RodgerTheGreat> so more a drama than a police action film
22:02:21 <KipIngram> No - lots of action.
22:02:29 <KipIngram> Just "local action".
22:02:33 <RodgerTheGreat> ah
22:02:39 <KipIngram> It was very action packed and violent.
22:02:44 <KipIngram> Probably very, very realistic.
22:03:00 <KipIngram> It's worth the price of the ticket.
22:03:57 <RodgerTheGreat> I'll keep it in mind
22:04:34 <RodgerTheGreat> the last film I saw was "the expendables 2", which was a very fun, dumb action movie
22:04:47 <RodgerTheGreat> it delivered pretty much exactly what was advertised
22:05:02 <KipIngram> Yes.  The jokes were superb.
22:05:12 <RodgerTheGreat> The Chuck Norris Jokes.
22:05:15 <KipIngram> Especially the Chuck Norris joke.
22:05:17 <KipIngram> :-)
22:05:21 <KipIngram> Precisely.
22:05:31 <RodgerTheGreat> and the whole intro sequence that was like 8 minutes straight of explosions
22:05:32 <KipIngram> Though there were jokes all over the map.
22:05:44 <KipIngram> You called it - exactly what it promised.
22:06:00 <KipIngram> And just enough Charisma Carpenter eye candy to please.
22:06:05 <RodgerTheGreat> Say what you want about Stallone, but the man knows how to make that kind of movie
22:06:37 <KipIngram> I don't have anything bad to say about him.  He knows what he's doing.
22:06:52 <KipIngram> I've taken a whole lot of pleasure from his work over the years.
22:07:21 <RodgerTheGreat> have you noticed they've started doing a bunch of remakes of goofy over-the-top action movies from the 80s as gritty, super-serious blockbusters?
22:07:38 <KipIngram> Yes.  Totally.
22:07:41 <RodgerTheGreat> apparently they're remaking robocop
22:07:47 <RodgerTheGreat> seems like a really bad idea
22:07:49 <KipIngram> Yes they are.
22:07:55 <KipIngram> They just remade Total Recall.
22:07:59 <RodgerTheGreat> yep
22:08:14 <RodgerTheGreat> what's next, big trouble in little china?
22:08:15 <KipIngram> I am seeing the orignal RoboCop at the Houston Alamo Drafthouse in a week or so.
22:08:22 <KipIngram> Part of their "Bad Ass 101" series.
22:08:28 <RodgerTheGreat> the original RoboCop is a fantastic movie
22:08:34 <KipIngram> Yup.
22:08:50 <KipIngram> I just saw "Serenity" there a week or so ago.
22:08:54 <KipIngram> Sold out a whole theater.
22:08:59 <RodgerTheGreat> I believe it
22:09:34 <KipIngram> We were hoping my wife's cousin and his wife would join us, but they didn't get their tickets in time
22:09:38 <KipIngram> Oh well.
22:09:42 <RodgerTheGreat> I just can't see RoboCop working at all if they played it straight. The zaniness and tongue-in-cheek aspects of it really made the original.
22:09:51 <KipIngram> Totally.
22:10:19 <KipIngram> Paraphrase, but: "How can I help you, officer?"...  "By resisting arrest."
22:10:27 <KipIngram> So cool.
22:10:33 <KipIngram> "You called for backup?"
22:10:36 <RodgerTheGreat> I'd buy that for a dollar!
22:10:52 <KipIngram> :-)
22:11:04 <RodgerTheGreat> "Would somebody call a goddamned paramedic?"
22:11:08 <KipIngram> So is there much Forth traffic here these days?
22:11:25 <RodgerTheGreat> ugh I don't even want to think about what they're going to do to ED-209
22:11:32 <RodgerTheGreat> some days
22:11:43 <KipIngram> I don't know what ED-209 is.  :-(
22:11:47 <RodgerTheGreat> the game I'm writing is made in forth
22:11:59 <KipIngram> Good.  What kind of game?
22:12:05 <RodgerTheGreat> http://3.bp.blogspot.com/-uIgeAk2DuFU/Tuh1yJ2W3pI/AAAAAAAAAy8/K7sC89bQonQ/s1600/ED-209+with+Dick+Jones.jpg
22:12:51 <RodgerTheGreat> I'm doing a randomly-generated dungeon adventure game in the style of "The Binding of Isaac", except you'll be controlling a squad rather than a single person
22:13:03 --- quit: nighty^ (Read error: Connection reset by peer)
22:13:43 <RodgerTheGreat> here's a current screenshot: http://i.imgur.com/AhKEg.png
22:15:27 <RodgerTheGreat> last time I showed my forth code here everyone thought it was hideous, though
22:17:04 <KipIngram> Well, Forth is a very personal thing.
22:17:28 <KipIngram> Which means you have a right to your own Forth code, and everyone else has a right to judge...  :-)
22:17:38 <RodgerTheGreat> it makes me happy when I write programs with it and that's all that really matters
22:17:39 <KipIngram> If it runs then that's the ultimate test, isn't it?
22:17:43 <RodgerTheGreat> true
22:17:47 <KipIngram> Exactly.
22:18:01 <KipIngram> It's yours - do with it what you will.
22:18:26 <KipIngram> I definitely recommend "Thinking Forth," if you haven't read it yet.
22:18:38 <KipIngram> It's available for free online - google it.
22:18:43 <RodgerTheGreat> I've read both that and Starting Forth several times
22:18:49 <RodgerTheGreat> I own dead tree copies actually
22:19:01 <KipIngram> Me too.  :-)
22:19:10 <RodgerTheGreat> I'm dying to get my hands on a copy of "Scientific Forth" but it is exceedingly rare
22:19:31 <KipIngram> :-|  Never heard of it.  I want it now that you've mentioned it.
22:19:52 <RodgerTheGreat> it describes using Forth for scientific computation
22:19:54 <KipIngram> Jesus.
22:20:02 <KipIngram> First results on Amazon are over $200.
22:20:10 <RodgerTheGreat> along the way it develops a bunch of DSLs so that that isn't insane in Forth
22:20:30 <RodgerTheGreat> including DSLs for state machines which are in turn used to build a FORTRAN extension in Forth
22:20:56 <RodgerTheGreat> KipIngram: yep. Totally sucks.
22:21:50 <RodgerTheGreat> I actually did some digging around to try to get ahold of the guy to see if he'd do another printing or release the manuscript or something but he died a while back and the trail goes cold
22:22:13 <RodgerTheGreat> reading book reviews about it just makes me salivate
22:22:38 <KipIngram> Wow.  Do you know if it takes a scaled fixed point approach or a floating point approach?
22:22:54 <RodgerTheGreat> actually the $203 copy is new
22:23:09 <RodgerTheGreat> hmm
22:23:27 <RodgerTheGreat> I'm unsure
22:23:29 <KipIngram> I may just have to spring for it.  The company I work for is being acquired - I have a few options.  Going to make a few $$$.
22:23:47 <RodgerTheGreat> the fixed-point approach is idiomatically forth
22:23:57 <KipIngram> CM was definitely a scaled fixed point kind of guy - with the approach being driven by the application at hand.
22:24:03 <KipIngram> Right.
22:24:47 <RodgerTheGreat> if you buy it and you have access to a book-scanner I would be in your debt for a PDF copy
22:26:12 <KipIngram> I will let you know if I buy it.
22:27:34 <RodgerTheGreat> godspeed
22:27:45 <KipIngram> At the very least we could discuss the details.
22:27:52 <KipIngram> That would be strictly legal.  ;-)
22:28:24 <KipIngram> My own special interest is the design of FPGA-based processors optimized for Forth.
22:28:44 <KipIngram> I focus on the very low-level aspects of FPGA circuitry.
22:28:46 <RodgerTheGreat> like the J-1?
22:28:58 <KipIngram> Yes, but as I said, low level.
22:29:15 <KipIngram> Not so much "how few lines of VHDL," but "how fast."
22:29:37 <KipIngram> I'm all about maximizing clock speed and getting things to work in single-cycle mode.
22:29:44 <RodgerTheGreat> cool
22:30:00 <RodgerTheGreat> what sorts of applications are your designs used for?
22:31:10 <KipIngram> Well, this is a hobby, and I've never actually gotten one into an FPGA yet.  But I know enough about FPGAs to know (for sure) that what I'm thinking about would work.
22:31:26 <KipIngram> Apps of interest to me include bitcoin mining, massively parallel processing, and stuff like that.
22:32:01 <KipIngram> I can easily see making a 1024 or so count parallel processor built around this design.
22:32:20 <KipIngram> Running at about 500-600 MIPS per processor.
22:32:37 <KipIngram> Xilinx Spartan 6 based.
22:32:52 <RodgerTheGreat> are they F18-style cores with extremely constrained stacks and program space?
22:33:45 <RodgerTheGreat> ah, here's Noble's "FORTRAN": http://www.taygeta.com/fsl/docs/551.jvn.fall01/ftran202.htm
22:33:59 <RodgerTheGreat> seems to use floats
22:34:19 <KipIngram> No - this design is entirely my own.
22:34:23 <KipIngram> From scratch.
22:34:33 <RodgerTheGreat> ah
22:35:00 <RodgerTheGreat> I was just asking whether you made similar design tradeoffs
22:35:04 --- part: SHUPFS left #forth
22:35:23 <KipIngram> The stacks are in hardware, but they're asynchronous and I can have them as deep as I like.
22:35:31 <KipIngram> Deeper just takes more of the chip.
22:35:44 <KipIngram> 16-bit address space for the programs.
22:35:57 <RodgerTheGreat> that's quite a lot
22:36:09 <KipIngram> 32-bit address space for data and 32-bit ALU operations, but that's variable as well.
22:36:23 <KipIngram> It's "enough."  :-)
22:36:25 <RodgerTheGreat> harvard architecture, check
22:36:44 <KipIngram> Yes, but FPGAs generally support dual port memory.
22:37:04 <KipIngram> So the two address spaces share the same memory.
22:37:34 <RodgerTheGreat> ok, makes sense
22:39:05 <KipIngram> The thing I spent the most time wrangling with was the conditional stuff.
22:39:33 <KipIngram> I organize program memory in cells.
22:39:56 <KipIngram> Each can contain either a jump or call instruction (conditional or unconcitional) or three five-bit opcodes.
22:40:20 <KipIngram> How to execute the conditional stuff without slowing down the machine turns out to be a big deal.
22:40:37 <RodgerTheGreat> yeah, pipelines suck
22:40:51 <KipIngram> No piplelining in this design.
22:41:09 <KipIngram> Well...
22:41:22 <RodgerTheGreat> well most of the problems I'm familiar with resulting from branches have something to do with prefetching
22:41:25 <KipIngram> The conditional execution is based on three-opcode cells.
22:41:47 <KipIngram> So if *any* of the three drive a conditional "don't execute" then the entire next cell is skipped.
22:42:00 <KipIngram> So that means if the first opcode in a cell does that, the next to execute.
22:42:03 <KipIngram> No matter what.
22:42:26 <KipIngram> So that's a little strange, but in Forth philosophy it's really up to the developer to keep up with stuff like that.
22:42:45 <RodgerTheGreat> if I understand correctly that sounds like a delay slot
22:42:46 <KipIngram> You are in control of the silicon.
22:43:10 <RodgerTheGreat> oh the horrors of the MIPS architecture
22:43:26 <KipIngram> Yes, but on the other hand if the third opcode is the one with the conditional then there's not a delay slot.
22:43:43 <KipIngram> I know.
22:44:05 <KipIngram> I don't like it completely, but alternatives slow the machine down on EVERY SINGLE CYCLE.
22:44:52 <KipIngram> And this approach forces 2 delay slots on you 1/3 of the time, 1 delay slot on you 1/3 of the time, and 0 delay slots on you 1/3 of the time.
22:45:10 <KipIngram> That averages out to 1 delay slot.  And if you program cleverly you take advantage of it.
22:45:31 <KipIngram> Earlier architectures I considrered avoided this at the architectural level but ran slwoer.
22:45:33 <KipIngram> slower
22:46:08 <RodgerTheGreat> having multiple instructions per cell is one of those things that's hairy to expose in an assembler
22:47:04 <KipIngram> Yes - I envision an assembler in which you lay out three instruction per line.
22:47:12 <KipIngram> So that part of things is highly visible.
22:47:15 <RodgerTheGreat> that's one good solution
22:47:46 <KipIngram> In a Forth-like coding style you might just insert some extra spaces between code "phrases."
22:48:00 <KipIngram> A B C   D E F   G H I   ...
22:48:15 <RodgerTheGreat> I wrote a softsim of the F18 and an associated assembler and I found myself frequently writing machinecode by hand because the assembler hid information I needed
22:48:16 <KipIngram> Haven't tried it yet.
22:48:29 <RodgerTheGreat> of course when I say "assembler
22:48:35 <RodgerTheGreat> " I mean a simple forth compiler
22:48:35 <KipIngram> Right - I don't like hiding stuff.
22:48:52 <KipIngram> It's not what Forth teaches us.
22:49:50 <RodgerTheGreat> the F18 has four ops per word, and if one of the first three slots is a branch the remainder of the current word is used
22:50:35 <RodgerTheGreat> this leads to some really thought-provoking layout optimization problems
22:51:20 <KipIngram> Yeah.
22:51:23 <KipIngram> I've read all that.
22:51:39 <KipIngram> Those variable length effective addresses complicate the logic.
22:51:45 <KipIngram> I only have one EA size: 16 bits.
22:51:57 <KipIngram> I do support micro-looping, though.
22:52:05 <RodgerTheGreat> the micronext functionality is very nice though
22:52:07 <KipIngram> That turns out to be trivially simple.
22:52:13 <KipIngram> Right.
22:52:36 <KipIngram> The really neat thing is that you can implement u-next in a way that's "free."
22:53:00 <KipIngram> You can see the u-next opcode coming and execute it ahead of time, so you wind up with an "opcode machine gun."
22:53:11 <KipIngram> No overhead whatsoever until you terminate the loop.
22:53:17 <RodgerTheGreat> nice
22:53:43 <RodgerTheGreat> although in your case you only have two ops in the loop to play with to the F18's three
22:54:05 <KipIngram> I set mine up so that I could use either one cell or two cells for that.
22:54:13 <RodgerTheGreat> ah ok
22:54:15 <KipIngram> So I can have up to five.
22:54:20 <RodgerTheGreat> that could be very flexible
22:55:22 <KipIngram> I've tinkered with this for years, but I've really made a lot of progress in the last few months, so hopefully I'll have it in an FPGA soon.
22:55:30 <RodgerTheGreat> the F18's addressing modes (through A for example) make it very easy to do fast copies and such. Did you add offset registers like that?
22:55:36 <KipIngram> I have a Spartan 6 eval kit under my coffee table right now.  :-)
22:55:58 <KipIngram> The architecture does support A and B addressing registers.
22:56:06 <RodgerTheGreat> great
22:56:27 <KipIngram> How flexible you want to be on that front turns out to affect opcode count more than anything else.
22:56:40 <KipIngram> And since I am restricted to 5-bit opcodes I have to be a bit conservative.
22:56:45 <RodgerTheGreat> naturally
22:56:46 <KipIngram> Can't waste them...
22:58:45 <KipIngram> Ok - got to sleep.  Hope to chat again soon...
22:58:49 <RodgerTheGreat> other neat thing about the F18 is how the operations are clearly chosen so that the propagation delay from the ALU is roughly equivalent. Do you have a "delay slot" in your addition operator?
22:58:55 <RodgerTheGreat> yeah I should sleep too
22:59:16 <KipIngram> buffer 1
22:59:33 <KipIngram> Ooops...
22:59:36 <KipIngram> :-)
22:59:46 <KipIngram> Primitive console-based client.
23:00:05 <RodgerTheGreat> haha
23:00:45 <KipIngram> Yeah - I run my irc on a virtural server based in the Netherlands.  Always on, 24/7.
23:00:51 <KipIngram> I ssh into it with finch.
23:01:08 <KipIngram> No - not finch.
23:01:12 <KipIngram> weechat-curses.
23:01:23 <KipIngram> finch is for my yahoo messenger stuff.
23:01:57 <RodgerTheGreat> I like my fancy GUI accoutrements 
23:02:12 <RodgerTheGreat> ...even if I turn most of them off
23:02:34 <RodgerTheGreat> 'night all
23:02:40 --- quit: RodgerTheGreat (Quit: RodgerTheGreat)
23:03:52 <KipIngram> Night.
23:52:50 --- quit: swiley (*.net *.split)
23:59:59 --- log: ended forth/12.09.22