Topic for #milkymist is now Milkymist One, Milkymist SoC & Flickernoise development channel (LLHDL/Antares are welcome too) :: Logs: http://en.qi-hardware.com/mmlogs :: JFDI
dvdk joined #milkymist
<dvdk> huh, just trying to get the latest ISE web pack, and the web page asks me for "company name" and "shipping address" for export regulation checks.
<dvdk> what do you guys enter there for download? just wondering whether I should insert "qi-hardware.com" or "sharism", which would even be partiall true...
<kristianpaul> dvdk: oh, hi !
<dvdk> hi
<kristianpaul> got a M1 too seems :-)
<kristianpaul> just lie
<kristianpaul> i enter real data, but i think other will not matter much...
<dvdk> (about M1) no, not yet. maybe sometime next year. just trying to get a working toolchain setup first.
<kristianpaul> sure
<kristianpaul> well, i think you can put qi-hw
<kristianpaul> why not? ;-)
<dvdk> lying would be easy. just feels discriminating, that they won't hand out theyr software to free software enthusiasts in a honest way.
<dvdk> yeah, think i will just do that.
<dvdk> they even want my phone number. yuck.
<kristianpaul> ha, just be sure disable webtalk (http://lekernel.net/blog/2010/05/xilinx-ise-12-phoning-home/)
<dvdk> maybe firewall the whole process
<kristianpaul> yeah, good idea too
<dvdk> ah, download started.
<dvdk> now i'm working at "qi-hardware" in the "free software development" department
<dvdk> ... located in germany
<kristianpaul> wellcome !
<wpwrak> ;-))
<wpwrak> they must be wondering what megacorp has recently become their customer
<kristianpaul> for the record i register in the name of the local hacklab :)
<kristianpaul> wpwrak: =)
Gurty joined #milkymist
[florian1 joined #milkymist
elldekaa joined #milkymist
<kristianpaul> 2 minutes of advantage computer from new laptop (synthesis), not bad :)
<wpwrak> huh ? you synthesized M1 in 2 minutes ?
<kristianpaul> erghh, sorry 16 minutes :-)
<kristianpaul> plus 2 in laptop
<wpwrak> still fast. but what exactly did you do ?
<kristianpaul> ah, well basically disable everything i could in the soc (tmu,fpu,usb,eth,pfpu,vide,midi,dmx,irac97), add namuru (just one channel)
<kristianpaul> and yes, i finally upgrade to 13.3 ise,
<kristianpaul> and no else i think
wolfspraul joined #milkymist
<wpwrak> kristianpaul: bah, and then they chickened out and didn't implement the AVR as well
<kristianpaul> i cant found sources...
<kristianpaul> lekernel: now milkymist.org redirect to the wordpress thing, wher is the old web archive??
<wpwrak> this has some stuff: http://milkymist.org/wp/for-developers/
<wpwrak> but not everything of the old material
<kristianpaul> np, archive.org have it :)
<wpwrak> well, some of the old things were useless. like quasi-empty PDFs. maybe the cleanup is still in progress
<wpwrak> hehe ;-))
<kristianpaul> cleaning is good
<roh> narf.
<roh> breaking links is bad. he should add rewrites
<kristianpaul> pages still there, just no acesible from the wp
<roh> e.g. where are the schematics?
<kristianpaul> milkymist.org/mmone.html
<kristianpaul> anyway, changes are good, i just was used to get schametics and other links from thre :)
xiangfu joined #milkymist
gbraad joined #milkymist
gbraad joined #milkymist
cjdavis joined #milkymist
xiangfu joined #milkymist
xiangfu joined #milkymist
sh4rm4 joined #milkymist
cjdavis1 joined #milkymist
lekernel_ joined #milkymist
r33p joined #milkymist
yizhang joined #milkymist
qwebirc20051 joined #milkymist
dvdk joined #milkymist
lars_ joined #milkymist
togi joined #milkymist
topi` joined #milkymist
zer0her0 joined #milkymist
ScopeukL joined #milkymist
roh joined #milkymist
whitequark joined #milkymist
guyzmo joined #milkymist
Fallenou joined #milkymist
juliusb joined #milkymist
azonenberg joined #milkymist
Gurty joined #milkymist
[florian1 joined #milkymist
cjdavis1 joined #milkymist
lekernel_ joined #milkymist
ximian joined #milkymist
aeris joined #milkymist
mwalle joined #milkymist
scrts joined #milkymist
ChanServ joined #milkymist
scrts joined #milkymist
xiangfu joined #milkymist
elldekaa joined #milkymist
<lekernel_> wpwrak: I wasn't planning to use any wacom stuff, but to clone their technology
<lekernel_> the sensor is a mere PCB
<lekernel_> with printed coils
<lekernel_> it's like an array of printed and multiplexed RFID readers
<lekernel_> you have an array for the X axis, and another one for the Y axis
<lekernel_> nothing hard to source, it's just work ...
<wpwrak> cloning is expensive for us :)
<wpwrak> we're low volume, so even "cheap" things are costly
lekernel__ joined #milkymist
<lekernel__> it's just a 4-layer PCB to stick behind the screen, some mosfets, and a medium-low range ADC. the rest can be done by the FPGA...
<lekernel__> (of course, this is a perfect case of the "let's stick an FPGA in there and worry later about what it does" problem *g*)
<GitHub30> [milkymist] none pushed 3 new commits to master: https://github.com/milkymist/milkymist/compare/dbcce91...36b1e07
<GitHub30> [milkymist/master] softusb: accept everything but non-MIDI and system messages - Werner Almesberger
<GitHub30> [milkymist/master] bios: make the LIBS as .PHONY - Xiangfu Liu
<GitHub30> [milkymist/master] libfpvm: use C operator precedence - Werner Almesberger
<wpwrak> hehe :)
<wpwrak> i'm also not so sure it works that easily through an LCM. don't you have to have the coils on top ? so that would mean to put them into the glass or some have transparent plastic circuit above the display
<lekernel__> no, it seems to work OK with the coils behind
<wpwrak> does the VJ really want to wield a pen ?
<wpwrak> the pen is mightier than the sword, but ...
<lekernel__> those things have a range of a dozen centimeters anyway
<lekernel__> the thickness of the LCM isn't important
<lekernel__> what can cause more problems are interferences with the LCM electronics
<wpwrak> i was thinking more of shielding effects of all the metal in there
<lekernel__> (the pressure is sensed on the pen, which sends the information wirelessly to the tablet)
<wpwrak> can you source pens ?
<lekernel__> yes... or just make them. not excessively hard either...
<wpwrak> also, in some cases, a simple touch screen can emulate pressure via duration
<lekernel__> the most simple form is a capacitor, a coil and a switch
<wpwrak> there are also fancier techniques that give you pressure directly from a touch screen. might be patented, though
<wpwrak> (make them) without them looking "cheap" ? :)
<kristianpaul> most of branded pens looks cheap afaik
<wpwrak> but i can't really imagine a pen-based interface to be suitable in the heat of a performance
<wpwrak> that's more a studio thing
<kristianpaul> open source vision in the other hand ;-)
<wpwrak> and i don't think most studios will be very interested in anything that's not 1080p and so on. there's always the tendency to use the best technology available, to be able to reuse the material in a few years
<lekernel__> did I say it wouldn't be 1080p? :)
<wpwrak> yes ! vision ! where's that ?
<wpwrak> 1080p ? not in M2 :)
<lekernel__> well, not that wacom thing either :)
<lekernel__> that's for the very long term
<wpwrak> in M3, with terahertz RAM with optical interconnect, hey, why not ;-)
<lekernel__> DDR3 would be sufficient already ;)
<wpwrak> well, let's focus on filling the exoskeleton for now. enough work waiting there already.
<wpwrak> but vision is a good point. that's something that would add a lot of value and should be possible. visikord-style
<kristianpaul> lekernel__: is it too hard add more instructions to pfu?
<kristianpaul> i havent check source yet, just asking quick and dirty here :)
<kristianpaul> i noticed your apoint about a second lm32 is just another slow cpu, so i want reconsider the power of multiple pfu or how customize it for specific task
<kristianpaul> s/apoint/point
<lekernel__> I'm for the dataflow HLS approach, not layers of reconfigurable stuff over layers of reconfigurable stuff
<kristianpaul> ah, you made the java thing to work finally? :-)
<kristianpaul> or get back to myHDL ?
<wpwrak> for vision, how about a camera pixel path element that "listens" to an NxM matrix in the pixel stream. for each pixel, it differentiates, does a bandpass filter, followed by a per-pixel hysteresis and maybe an overall quorum. that would then become a signal to the LM32 and from there it could go into a PFPU variable
<lekernel__> neither are suitable for what I want to do, so I'm building my own
<lekernel__> wpwrak: what do you mean "differentiate"? compare with the previous value?
<wpwrak> yes
<wpwrak> well, in a first step, it could be just an array of thresholds
<wpwrak> one issue would be slow changes, like ambient illumination
<wpwrak> (array of thresholds) with 1x1 being an option ;-)
<lekernel__> the whole thing would certainly work well with dataflow synthesis :)
<wpwrak> ;-)
<kristianpaul> btw xula/xess seems to be working on a spartan6 board, so here is other coming alternative for milkymist soc development hack and fun
[florian] joined #milkymist
r33p joined #milkymist
cocoadaemon joined #milkymist
errordeveloper joined #milkymist
Gurty joined #milkymist
elldekaa joined #milkymist
<GitHub173> [migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/cd8544c7581d8df8f513ef4b920e79a4ef3ba8be
<GitHub173> [migen/master] Verilog generator - Sebastien Bourdeauducq
<GitHub143> [migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/e099f4d52f60511a9016d7e9489124956d3a73da
<GitHub143> [migen/master] Reset insertion - Sebastien Bourdeauducq
<GitHub198> [migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/5acf2e169fe963092a1ab26d9ec6f7a341cc3311
<GitHub198> [migen/master] Examples folder - Sebastien Bourdeauducq
<GitHub61> [migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/458cfc862325e256706d70eb7ee6f5fa012668a9
<GitHub61> [migen/master] CSR bus definitions - Sebastien Bourdeauducq