**** BEGIN LOGGING AT Fri Nov 04 02:59:56 2005
Nov 04 07:21:47 <vmaster> you guys discuss the interesting stuff at the worst time possible: 5am to 9am
Nov 04 07:27:53 <[g2]> 5am to 9am where ? :)
Nov 04 07:28:45 <vmaster> Germany, GMT+1
Nov 04 07:29:13 <[g2]> oh so you mean 11AM to 3PM :)
Nov 04 07:29:56 <[g2]> as it's 9:30AM here :)
Nov 04 07:31:32 <vmaster> mhh, 11PM to 3AM then for you
Nov 04 07:32:11 <prpplague> hehe
Nov 04 11:37:39 <ka6sox> vmaster, too early for Uni students?
Nov 04 11:44:00 <vmaster> heh, yeah, got up one hour after you went to bed
Nov 04 12:07:01 <lennert> tonight is hacking night!
Nov 04 12:10:58 <ka6sox-office> lennert, do you have a S3 board?
Nov 04 12:11:09 <ka6sox-office> never mind...You must have.
Nov 04 12:11:10 <lennert> ka6sox-office: yes, the digilent one
Nov 04 12:11:14 <lennert> yeah, for pong :)
Nov 04 12:11:59 <ka6sox-office> lets see about putting together something that we can use to make the JTAG faster.
Nov 04 12:12:07 <ka6sox-office> (and do some COMP stuffs.
Nov 04 12:12:20 <lennert> yes
Nov 04 12:12:24 <lennert> that's one of the things 'on the list'
Nov 04 12:12:49 <lennert> i have to look into how to build the interface circuitry
Nov 04 12:12:58 <ka6sox-office> okay since there are several of us we should work together on it.
Nov 04 12:13:04 <lennert> i have another xilinx CPLD board with jtag, so i could in theory use the s3 board to jtag the cpld board
Nov 04 12:15:15 <lennert> i've thought about vmaster's idea, and i can't poke any holes into it
Nov 04 12:15:21 <vmaster> i have working code to program coolrunner devices with xilinx xsvf files
Nov 04 12:15:32 <lennert> i.e. either you scan into IR, or you scan into DR, and other cases are 'exceptional
Nov 04 12:15:56 <lennert> vmaster: programming sp3 is trivial, see http://svn.wantstofly.org/vhdl/utils/
Nov 04 12:18:33 <vmaster> yeah, xcr is almost as easy, it's just a few command types that tell you what to do, like irscan, drscan, move the statemachine
Nov 04 12:18:46 <lennert> yeah
Nov 04 12:18:54 <lennert> i was looking at it too generally
Nov 04 12:19:07 <lennert> i.e. when you clock a sequence of bits, you want to be able to control tdi/tms and store tdo somewhere
Nov 04 12:19:17 <vmaster> yep
Nov 04 12:19:28 <lennert> but indeed, you're mostly either 1/ clocking tms to get into another state; or 2/ clocking tdi in shift-ir or shift-dr.
Nov 04 12:19:36 <lennert> that observation alone makes it a bunch simpler.
Nov 04 12:23:49 <ka6sox-office> yes I agree.
Nov 04 12:24:02 <ka6sox-office> we need to be able to use the onboard memory for that job.
Nov 04 12:24:28 <ka6sox-office> I was thinking of working on the memory code today/this weekend.
Nov 04 12:24:50 <vmaster> the low-level stuff is trivial - but to achieve high performance, you'll have to define a higher-level interface
Nov 04 12:25:20 <ka6sox-office> I wonder if we use the picoblaze we might get some performance help.
Nov 04 12:25:30 <vmaster> http://www.digilentinc.com/info/S3Board.cfm <- that's the board you have?
Nov 04 12:25:34 <ka6sox-office> (high lev el stuff)
Nov 04 12:25:41 <ka6sox-office> vmaster: yes
Nov 04 12:25:42 <lennert> vmaster: yup
Nov 04 12:27:10 <lennert> vmaster: it can do 180mhz des 8)
Nov 04 12:27:25 <ka6sox-office> lennert: what size S3 is on your board?
Nov 04 12:27:28 <vmaster> heh, yeah, saw your vhdl code
Nov 04 12:27:34 <lennert> xc3s200 :(
Nov 04 12:27:44 <lennert> i'd love an xc3s1000
Nov 04 12:29:06 <ka6sox-office> okay so I think we all have the same size box.
Nov 04 12:29:25 <ka6sox-office> I forget how many slices a Picoblaze takes.
Nov 04 12:29:31 <vmaster> i'll talk to my professor to get me one of these... there should be some money left for toys to play with
Nov 04 12:29:51 <ka6sox-office> vmaster: the 1000 or 200?
Nov 04 12:30:38 <vmaster> how much $$$ is the difference between the two?
Nov 04 12:30:43 <lennert> vmaster: $50
Nov 04 12:31:13 <vmaster> mhh, guess the S1000 then
Nov 04 12:31:25 <ka6sox-office> makes sense.
Nov 04 12:31:46 <ka6sox-office> okay I still think with the 200 we should be able to make this go.
Nov 04 12:32:08 <vmaster> yeah, of course
Nov 04 12:32:18 <lennert> you don't need much logic for jtag
Nov 04 12:32:49 <vmaster> not for low-level stuff
Nov 04 12:32:49 <ka6sox-office> its mostly about cranking up the memory bandwidth.
Nov 04 12:33:37 <ka6sox-office> I would like to see something that has a pattern matching system so that we can program things with the S3 doing all the heavy lifting.
Nov 04 12:33:48 <lennert> even if you define higher-level primitives (such as "read the word at xscale memory location X using icache jtag"), you still don't need much logic
Nov 04 12:34:16 <vmaster> mhh, that's probably too high - would be very target specific
Nov 04 12:34:30 <lennert> vmaster: yes, it's just for illustration
Nov 04 12:34:31 <ka6sox-office> that is what I would like to do but use some of the RAM for that kind of matching.
Nov 04 12:34:44 <ka6sox-office> so that the box would be "universal
Nov 04 12:41:43 <vmaster> what host interface do you want to use? usb 2.0?
Nov 04 12:44:19 <lennert> the digilent sp3 doesn't really have any suitable interface
Nov 04 12:45:18 <ka6sox-office> I was thinking about using the usb2.0 interface with the uC interface.
Nov 04 12:45:25 <ka6sox-office> (not serial but parallel)
Nov 04 12:45:28 <vmaster> the new xilinx board is going to have usb 2.0, iirc?
Nov 04 12:45:37 <ka6sox-office> the new one will yes.
Nov 04 12:46:08 <ka6sox-office> so for now what I intend to do is do development on the S3 and eventually switch to the S3e
Nov 04 12:46:37 <ka6sox-office> I can't remember if its going to be Dram or Sram.
Nov 04 12:46:39 <lennert> i guess i should sell my s3 and get an s3e too
Nov 04 12:46:51 <lennert> (but i have a v2pro board on order as well..)
Nov 04 12:46:54 <ka6sox-office> lennert, that is a good choice.
Nov 04 12:47:05 <ka6sox-office> the v2pro can have the PPC cores.
Nov 04 12:47:57 <lennert> yeah
Nov 04 12:48:01 <lennert> i also have an ml310 to work with now
Nov 04 12:50:15 <ka6sox-office> lennert, do we have armeb gnat?
Nov 04 12:50:20 <lennert> ka6sox-office: nope
Nov 04 12:50:27 <lennert> ka6sox-office: regular arm doesn't have gnat either
Nov 04 12:50:31 <lennert> ka6sox-office: gnat needs gnat to build, apparently
Nov 04 12:50:41 <lennert> ka6sox-office: but drow said there's no reason why arm doesn't have gnat
Nov 04 12:50:46 <lennert> ka6sox-office: so it shouldn't be that hard to add
Nov 04 12:51:07 <ka6sox-office> that would be useful to me.
Nov 04 12:52:14 <ka6sox-office> the S3e board with 32mb of RAM would be quite a good box to look at using for a debugger.
Nov 04 12:52:44 <lennert> i'd have to take [g2] up on his offer :)
Nov 04 13:03:24 <[g2]> heh :)
Nov 04 13:04:24 * [g2] thinks lennert is going to be doing FPGA work for someone :)
Nov 04 13:04:37 <lennert> [g2]: for you :)
Nov 04 13:05:03 <[g2]> hey I'd love to have you on my team
Nov 04 13:05:18 <[g2]> but I think you are already doing work
Nov 04 13:05:40 <lennert> i'm doing lots of stuff all over the place
Nov 04 13:05:51 <lennert> the current day job isn't too bad
Nov 04 13:05:56 <[g2]> I know :) I can barely keep up
Nov 04 13:06:02 <lennert> hehe
Nov 04 13:06:51 <[g2]> so you've got the UART going right ?
Nov 04 13:07:05 <lennert> yeah
Nov 04 13:07:32 <[g2]> you guys know of any modem chips ?
Nov 04 13:08:13 <lennert> nope, sorry
Nov 04 13:08:45 <ka6sox-office> not anymore..all the ones I know are obsolete.
Nov 04 13:09:07 <[g2]> are there current gen ones ?
Nov 04 13:09:38 <ka6sox-office> winmodems only now :P
Nov 04 20:05:01 <ka6sox-office> 4am in Europe...sleeping people.
Nov 04 21:44:55 <key2> 5:47 u mean :)
Nov 04 21:45:36 <ka6sox-office> ah..20:47 here
Nov 04 21:47:13 <key2> where are you ? Cali ?
Nov 04 21:50:30 <ka6sox-office> ya
Nov 04 21:50:52 <key2> where about ?
Nov 04 21:51:05 <ka6sox-office> Santa Barbara
**** ENDING LOGGING AT Sat Nov 05 02:59:57 2005