#milkymist IRC log for Tuesday, 2011-10-11

wpwrakwolfspraul: hmm, a bit tricky then. not very expensive but still.01:36
wpwrakone consideration would also be how many of those already made would actually be sellable. e.g., if you're thinking of making a new revision, some of the boards in stock may actually have zero commercial value.01:54
wpwrak(and could be used to sweeten the current M1 offer, my declaring them as a free extra for early adopters)01:55
wpwrakif you can lower the price of the whole system in the future, dropping the jtag would also justify that change to existing customers.01:56
wolfsprauljtag-serial is included with every m1 for the time being02:13
wolfspraulwe need to learn much more about who buys it, where, etc. before optimizing this02:13
wpwrakgood. so we're on the same page :)02:14
wpwrakbtw, did you try to plant M1 at any other news outlet than The Reg ?02:18
wolfspraulno02:19
wolfsprauloverloaded02:19
wolfspraulbut I work on all this, and of course will continue02:19
wpwrak(overload) hmm. so this also means ... november news then ? :)02:21
wolfspraulha ha02:26
wolfspraul:-)02:26
wolfspraulyou do want to push me into the cemetery faster...02:26
kristianpaulbien, igual .... -_-02:26
kristianpauloops02:26
wolfspraulyes I am late on that too02:26
wolfspraulI am sitting here explaining m1 to our new China sales person :-)02:26
wpwrakheh, great ! :)02:27
wpwrak(the sales person, not the cemetery :)02:27
wolfspraulthe plan is to get it into one reference club, then work with that club to polish the product, in parallel sell that club's success story to other clubs02:27
wolfspraulthe sale into the first club is 'almost' a done deal02:27
wolfspraulbut I want to see cold hard cash first...02:27
wpwrakreference customers sound like a good idea02:28
wpwrak;-)02:28
wpwrakthey should consider the value of all the free support they're getting. priority access to the product's development team of a globally acting enterprise. where do you find that without putting a six-figure sum on the table ? ;-)02:29
wolfspraulthat's the club http://www.acupuncture-records.com/club/lantern-club/?lang=en02:30
wpwrakno pictures ?02:36
wpwrakdescription sounds good, though. avantgarde, just the right fit02:37
wolfsprauljust opened 2 weeks ago03:08
wpwrakhmmm. i locked the usual suspect yet i see this  "Checking : flickernoise.fbi(rescue)(CRC)      CRC failed". looking forward to a deeper analysis of what happened to my lock bits ...05:14
wpwraks/suspect/suspects/05:14
kristianpaulhttp://cdn.opencores.org/downloads/opencores_coding_guidelines.pdf not bad05:19
wolfspraulkristianpaul: very nice! (downloading, haven't read yet)05:33
wolfspraulkristianpaul: oh my :-)05:34
wolfsprauljust look at the usage restrictions on page 2 :-)05:34
wolfspraulopencores, ha ha05:34
wolfspraulsome people really always have it backwards05:34
wpwrakand where are the remaining 200 pages of fine print ? there are LOTS of cases they don't cover :)05:36
stekernhttp://opencores.org/websvn,filedetails?repname=common&path=%2Fcommon%2Ftrunk%2Fopencores_coding_guidelines.pdf05:45
stekernthat is the version before the restrictions were added...05:46
wolfspraulseriously? they were *added* at some point?05:46
wolfspraulmy god05:46
wolfspraulI mean we all learn, adapt, etc. we see what works and what doesn't work. but the motivation of someone adding this to an existing document must be the weirdest thing I can imagine.05:47
wolfspraulI am 100% sure they did not have cases where those restrictions actually made a difference. So they must have added them just for the heck of it :-)05:47
wolfspraulstekern: ok but that version has no copyright information at all, and with the very specific information on later version I'd hesitate to reuse the older one as a starting point05:50
wolfspraulit seems they don't want to be built upon...05:50
stekernwell, "they" are different "theys" in the different versions05:51
stekernalso, compare: http://cdn.opencores.org/downloads/wbspec_b3.pdf vs http://cdn.opencores.org/downloads/wbspec_b4.pdf05:52
sb0do not read any opencores guidelines. they lead to the results we know too well.07:11
sb0http://www.ohwr.org/documents/2407:19
stekernI think the opencores guidelines are fairly good, a bit outdated in parts perhaps though07:25
stekernthe ohwr ones are good too07:26
sb0cheap fpga board with sdram and usb: http://www.xess.com/prods/prod048.php07:35
kristianpaulwolfspraul: yeah license is a shame :-/11:15
kristianpaulstekern: nice memory ;)11:17
kristianpauli agree with stekern11:18
kristianpaulalso opencores is the nearest thing we have to an upstream (for some glue stuff like wishbone, ethernet cores?) :) and yes there is ohwr :-)11:20
kristianpaulhttp://www.ohwr.org/companies/milky-mist11:34
wpwrakhmm, no hw development, no hw commercialization, no sw development ?12:07
wpwrakhmm, a question about the bitstream format: our standby bitstream ends in ... 00 b0  04 00 ...  ff ff ...12:43
wpwraki understand the 00 b0 and the ff ff but not the 04 00. ff ff is simply uninitialized NOR.12:43
wpwrak00 b0 must be the DESYNC WORD from table 5-50 page 102 of ug380. i.e., 0x000d, bit-swapped12:44
wpwrakbut where does the 04 00 come from ? (or 0x2000, bit-swapped)12:45
xiangfuwpwrak,  0400 bit-swapped is 2000 which is is NOOP, seems there are two many NOOP at the end of bitstream13:03
wpwrakaah, they never introduce it in the document but they use the NOOP at a few places. thanks !13:14
wpwraki think the trailing NOOPs in standby.fpd that don't get flashed into the NOR are a mismatch between the file and what urtag expects13:14
wpwrakurtag seems to expect file data to be a multiple of a certain size, maybe N*16 bytes. anything that's not such a multiple is ignored13:15
wpwrakand the file doesn't have such padding13:15
wpwraki guess urtag should either do the padding itself or complain if it's given a file with an unaligned size13:16
xiangfuwpwrak, yes.13:17
xiangfuneeds to read more document and the urjtag code :)13:18
wpwrakhmm, http://www.milkymist.org seems to have troubles15:35
kristianpaulah still15:35
wpwrakapparently some database backend not responding15:35
mwallewpwrak: the bitstream loading code?20:14
mwalle(padding, multiple of N)20:14
wpwrakyes20:14
wpwrakwell, the bitstream flashing code20:15
mwallemh?20:15
mwallefjmem ?20:15
wpwrakfile -> NOR. somewhere along the path, something seems to go wrong with unpadded files20:15
mwallemh, ok no code from me expect the fjmem core ;)20:16
wpwrakhehe :)20:17
--- Wed Oct 12 201100:00

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