#milkymist IRC log for Tuesday, 2012-03-13

kristianpaulso.. migen is used at cern?00:55
wpwrakkristianpaul: it'll be blamed for the black holes eating the planet they're producing these days01:41
wpwrakof course, it won't be blamed for very long01:42
ThihiI just visited a local hackerspace. They had a box for higg's bosons.02:12
wolfsprauloops, seems emotions are flying high on the myhdl list04:26
wolfspraulseems everybody's nerves are a bit irritated :-)04:27
wolfspraulis either migen or myhdl a superset of each other, or are they trying to solve different problems?04:37
wolfspraulsuperset conceptually, not necessarily today...04:37
cladamwwpwrak, sorry. I don't understand your U18(ADV7125)'s central pad's connection to ground ? It's LQFP, if it's LFCSP_VQ then needs a central pad to ground. Did you mean this ? or I misunderstood ?06:00
wpwrakcladamw: oh, i see. right, it's not QFN. i was confused by the data sheet that shows the pin configuration only for QFN. didn't realize we had LQFP instead. sorry.08:01
cladamwwpwrak, R142, since 530 Ohm is not standard value but I checked http://www.yageo.com.tw/, there's 523 Ohm and 536 Ohm: http://octopart.com/partsearch#search/requestData&q=RC0402FR-07536RL08:25
cladamwso a 536 Ohm is okay to you ?08:25
wpwrak(R142) do you think it matters whether we have 530 or 560 ? it's almost within the 5% tolerance band anyway :)08:39
cladamwwpwrak, could you check "REFERENCE INPUT" on page 12 of ADV7125 ? there's a 560 Ohm mentioned: "Use of a fixed 560 © RSET resistor yields the analog output levels quoted in the Specifications section"08:40
wpwrakmy reasoning is like this: a different value for Rset affects the VGA current/voltage. it should this change the intensity of the respective color channel. now, given that the receiving device (screen, etc.) has brightness and contrast controls anyway, i would expect that you can compensate for such issues easily08:40
cladamwi see your points, since tso i think 530 Ohm is deriving from Equation 1 for IOG (mA), so AD also knows a 560 Ohm is the most popular  value in market that can get easily, so i think AD mentions a 560 Ohm. but true that if we change to 536 Ohm to close it. ;-)08:44
Fallenou19:26 < sh4rm4> Fallenou, once the MMU is finished i'd be interested if musl libc could be ported to milkymist-linux < I haven't head about musl libc before, looks interesting though :) MMU is far from finished it needs tests, work, bugfix. I just implemented a few (16) tests, which are not yet passing for a reason I don't yet understand. I have been quite busy those days and I didn't work much on the MMU unfortunately. :/08:48
lekernelkristianpaul: not as far as I know08:53
lekernelrhino != cern08:53
lekernelwolfspraul: yeah, amazing heh? "let's keep our tiny group closed and oblivious of what others are doing, because they might exploit us..." :)09:14
lekernelanyway, don't feed the troll ...09:14
larscI bet some people consider you to be the troll ;)09:27
lekernelprobably :)09:29
Fallenouhe's tall, he's the troll !09:32
wolfspraulwell, let's work on bringing people together12:33
wolfspraulit's not easy [talking about migen & myhdl]12:33
wolfspraulstart with diversity, some are old, some young, some students, some industry veterans, some have 2h/week, some have nothing else to do, some know a lot, some know little, some want a lot, some want little. and so on and so on.12:34
wolfspraulmaybe it's a miracle that any collaboration happens in this mess :-)12:34
GitHub125[rtems-yaffs2] sebhub pushed 2 new commits to master: https://github.com/milkymist/rtems-yaffs2/compare/1a9b2b8...cbe749213:14
GitHub125[rtems-yaffs2/master] Change mode of root and lost+found directory - Sebastian Huber13:14
GitHub125[rtems-yaffs2/master] Update due to API changes - Sebastian Huber13:14
lekernelah, great :)13:18
lekernelAPI changes at last13:20
rohwolfspraul: in opensource there ALWAYS need to be 2 slightly different implementations for the same problem15:14
rohgcc and llvm... kde and gnome ... qt and gtk .... emacs and vim...15:14
roh ;)15:14
qi-botThe firmware (using branch) build was successful, checkout the VERSIONS for detail, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120313-1456/15:36
Action: wpwrak hates half-day blackouts17:49
larscdon't party so hard ;)17:54
wpwraki'll tell that to my electricity company ...17:55
larscah, a power blackout and not a wpwrak blackout17:57
kristianpaulbut, whas it scheduled?17:57
kristianpaulmay be adverticed on local the radio, by 4AM ;)17:58
kristianpauloh hard words..18:01
Action: lekernel wonders what Buenos Aires is like. should travel there someday... :)18:04
wpwrakkristianpaul: scheduled ? you're kidding, right ? ;-)18:07
kristianpaulwpwrak: perhaps18:07
kristianpaulbut half-day sounds serious for not been schedule/warned perhaps??18:07
wpwraklekernel: it's definitely worth a visit :) and we don't always have blackouts. although we'll probably see more of them for a while18:08
wpwrakkristianpaul: i doubt it was planned. i don't think they do any scheduled maintenance. if it fails, they fix it. maybe.18:09
Action: kristianpaul want vist la patagonia18:09
kristianpaulso you may consider arrange energy alternative sources considering the wind aroud :-)18:10
qi-botThe firmware build was successful, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-20120313-1746/18:26
wolfspraulroh: oh sure [several alternatives for same problem] I was trying to understand migen & myhdl better23:41
lekernelyou'll probably learn on this topic as much from this thread as from your local pub's counter :-)23:43
wolfspraulthat's why I'm asking here23:43
wolfspraulI could go through the myhdl archives, seems the clue is there - but I'm a little lazy :-) if no answer now, it will become clearer over time23:44
wolfspraulwill there be a big 'switch-day' where the m1 soc switches to migen?23:44
lekernelin a few words, myhdl is very close to what verilog/vhdl can already do, while migen allows you to use more python features, and also comes with some library components23:46
lekernelalso, myhdl can model asynchronous designs while migen cannot (but it can be interfaced easily to asynchronous bits)23:46
lekerneland up until recently, migen had no way to write test benches in python, while this was always an interesting feature of myhdl23:48
lekerneland yes, there will be a big switch-day.23:48
lekernelwpwrak: yeah, they're not consistent about that resistor value23:51
wpwrakthey're not even consistent about their voltage reference ;-)23:51
wpwrakat least the general configuration (termination resistances, and single vs. doubled ended) seems to be correct23:52
wolfspraullekernel: ah nice, that's very helpful. thanks!23:56
lekernelinteresting ethernet findings... btw, some rc1 boards have the original Molex P/N, and ethernet works on them23:56
wolfspraulhuh? problem with Ethernet?23:57
wpwrakhmm. the SN75HVD12 is a bit weird. what on earth is that little NAND gate doing there ? the function tables are remarkably silent about it23:57
lekernelwolfspraul: well, the chip deals with it... see Werner's email23:58
wpwraklekernel: i think ethernet is extremely robust. so "works" may mean a number of things :)23:58
wolfspraulcan we improve something in R4? (will read mails of course...)23:58
wpwrakwhat i'm not sure about is how those transformers work. we'll need joerg for that :)23:59
wpwrakwolfspraul: dunno. it could also be an incorrect data sheet23:59
--- Wed Mar 14 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!