Topic for #milkymist is now Radical Tech Coalition :: Milkymist One, Migen, Milkymist SoC & Flickernoise :: Logs: http://en.qi-hardware.com/mmlogs
wolfspraul has joined #milkymist
pablojavier has joined #milkymist
pablojavier has quit [#milkymist]
xiangfu has joined #milkymist
kilae has joined #milkymist
azonenberg has joined #milkymist
wolfspraul has joined #milkymist
cladamw has joined #milkymist
kilae_ has joined #milkymist
wolfspraul has joined #milkymist
DJTachyon has joined #milkymist
<wpwrak> hmm, i wonder if we should have a perl-like <assignment> if <condition>; construct
<wpwrak> i notice that - particularly with midi controls - i'm doing a lot of foo = cond ? complicated_expression : foo;
<wpwrak> that would then become foo = complicated_expression if cond;
<kristianpaul> easy to read :)
methril has joined #milkymist
DJTachyon has joined #milkymist
Gurty has joined #milkymist
DJTachyon has joined #milkymist
lekernel_ has joined #milkymist
<GitHub152> [milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/UkVStA
<GitHub152> [milkymist-ng/master] Send SDRAM initialization sequence and answer PHY read/write requests. Obstinately refuses to work, unfortunately. - Sebastien Bourdeauducq
<kristianpaul> wpwrak: AWESOME !!
<kristianpaul> animations :)
<kristianpaul> getting ready for the next party? ;-D
<wpwrak> heh :)
<wpwrak> yeah, i think with good image support, we can do quite a lot
<kristianpaul> indeed
<wpwrak> i.e., fake lots of effects m1 isn't capable of ;-)
* kristianpaul needs a midi controller too
<kristianpaul> fake, lol :)
<kristianpaul> finally mid and bass make more sense for me :)
* kristianpaul thinks in a fake matrix lets said 20x 20 pixels
<kristianpaul> but how i could draw ont it..
<kristianpaul> nah too complicated..
<wpwrak> you may want to look at how i've done the rotating disc. it's a multi-layered xfig file. then i labeled the elements according to the frame in which i want them, then filter the xfig and convert the filtered result
<wpwrak> for sequences with more frames you'd probably want some kind of keyframe animation. but for simple things this works quite nicely
<lekernel> yeah, well done wpwrak! thanks!
LmtdAt has joined #milkymist
<wpwrak> shall i merge it into master ?
<lekernel> shall we push out a web-update?
<lekernel> yes, sure
<wpwrak> ;-))
km2 has joined #milkymist
DJTachyon has joined #milkymist
<wpwrak> merged
* lekernel loves to fight DRAM problems. I wish I had a magical internal logic analyzer in the FPGA which samples at 400MHz, takes only a few seconds to attach to a signal, has a reasonably deep memory...
cjdavis1 has joined #milkymist
<lekernel> actually that would be a nice application of bitstream format reverse engineering
jonand_ has joined #milkymist
<wpwrak> i have one code generation / scheduler regression to track down: in a different version of the animation patch, some of the variables aren't updated properly. need to see what's causing that. back when i wrote the scheduler, i didn't fully understand the lifecycle of variables, so maybe i got something wrong there.
<wpwrak> luckily, it's pretty difficult to hit that one :)
Scopeuk has joined #milkymist
cxadams has joined #milkymist
<GitHub7> [migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/d8d4e81b6e0624cb7055dbb147a967f5975f294b
<GitHub7> [migen/master] bank/csrgen: fix RE generation - Sebastien Bourdeauducq
Gurty has joined #milkymist
<wpwrak> lekernel: you think the fpga could sample at 400 MHz ?
<lekernel> it's sampling the DRAM pins at 333MHz atm (or at least it's supposed to)
<lekernel> and of course, no fast scope/LA around ...
<lekernel> I was hoping the NWL PHY would work, but it does not. in fact, it seems rather crippled with bugs.
<wpwrak> hm. 300 Mhz would be quite respectable already
<wpwrak> particularly if you can DMA to DRAM
<lekernel> and there's absolutely no documentation about how it calibrates the data paths, which is what is failing atm
<lekernel> RTFS ...
<kristianpaul> what S stand for? Sources? ;)
<wpwrak> silicon ?
<lekernel> yes, sources
<lekernel> they have some system that writes LFSR-generated data to the DRAM and attempt to read it back
<kristianpaul> sparv_8 soc https://github.com/teeshina/soc_leon3
<lekernel> but in case of problems, I have quite no idea how it determines whether the problem is on read or write
<kristianpaul> calibrate data paths sounds something, you are usually forbid to ask ;)
<whitequark> ok, finally got time to look at my new m1
<whitequark> it looks awesome! too bad I don't have a VGA anything at the moment
<whitequark> I was especially impressed by overall build quality, and the huge collection of socket plugs. I thought there were less of them in the world
hypermodern has joined #milkymist
<hypermodern> Greetings do any of you have some links to videos of you using the Milkymist? I'd like to have your permission to use them in a presentation about the hardware.
<GitHub48> [milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/5r1XPw
<GitHub48> [milkymist-ng/master] bios: fix function prototypes - Sebastien Bourdeauducq
lekernel has joined #milkymist
Technicus has joined #milkymist
<lekernel> hi hypermodern
<lekernel> sure, http://vimeo.com/30842451
<lekernel> the latest one from werner http://www.youtube.com/watch?v=NrCbg6HNb5A
<lekernel> according to our DRAM specs we have a data output valid window of 2.1 ns worst case, and the data timing relative to the clock is +/- 0.7 ns
<lekernel> I wonder if this little calibration mess is really necessary anyway
<lekernel> or if we can just treat it as a normal synchronous circuit
<lekernel> hmm... it's tight, but it seems we can