#milkymist IRC log for Tuesday, 2012-05-08

Hodappwolfspraul: I just finished talking with Bill who'd made the board for that Microvision laser projector, though I had to take off before too long00:00
Hodappif you had any specific questions about the firmware I can look through what I wrote down and remember00:00
wolfspraulno specific question, just curious in general. block diagram and schematics would be nice, just learning how it's made. though I would be surprised if any of that would actually be published, I would think that someone thinks this is their secret knowledge...00:01
wolfspraulafter trying the projector for a while in real life, I am sure a lot more money needs to be invested in this first, before it can become a serious competitor to the other projector technologies00:02
wolfspraulmaybe there are some niche use cases already? don't know00:03
wpwrakwolfspraul: did you run into any troubles with schhist + m1r4 ? i think it should work, except for some false "corruption" caused by non-monotonous changes in the components library (renames or deletions), because schhist only uses the latest version of it, not the version that was around at the time of the respective revision of the schematics00:34
wolfspraulI got stuck generalizing the paths in the script00:36
wolfspraulonce the emergencies subside I'm back at it00:36
wpwrakemergencies ? did you blog against the government ? :)00:38
wolfspraulno no, just small technical things here and there00:38
wpwrakah, business as usual then :)00:39
cladamwwpwrak, tks for updating latest m1r4.pdf, i've not finished all so far.00:44
cladamwwpwrak, do you know that how to move that hierarchical sheet #19 to first sheet #1?00:46
wpwrakfor the PDF ?00:49
cladamwfor design files00:57
cladamwalso for PDF.00:57
Jiawhere can I find the Mico32-ABI manual?01:00
wpwrakhmm. you can launch eeschema on a sub-sheet, but strange things may happen then01:00
wpwrakfor the PDF, you can put the .ps files in any order you want. i just used alphabetic order (_*.ps), because that's the easiest01:01
Jialaunch eeschema?01:02
wpwrakJia: that was for cladamw01:02
Jiaoh01:02
wpwrakJia: i don't know if anyone here is very familiar with mico3201:02
Jiadoes not milkmist working on lm32?01:03
Jialm32==Mico3201:03
wpwrakcladamw: one we have schhist up, we can specify an explicit order there (partial or complete)01:03
wpwrakJia: lm32 != Mico32 :-) and yes, milkymist uses lm3201:03
wpwrakcladamw: s/one/once/01:04
Jiaall right, where can I get lm32 ABI manual?01:04
wpwrakcladamw: if you tell me your preferred order, i can add a rule to the m1r4 makefile, so that one can create a "nice" pdf also without schhist01:05
sh4rm4 http://www.milkymist.org/socdoc/lm32_archman.pdf01:06
xiangfuJia, ^01:06
wpwrakwe should get adam a roll of duct tape, for his internet connection :)01:08
Jiabbl01:12
Jiaxiangfu: if I change lm32 ABI, will break RTEMS?01:19
Jiaxiangfu: bingo01:21
wpwrakheh heh. let M1 run long enough (many days), and patches that use things like sin(time) think time has stopped :)01:26
Hodappo_O01:26
Hodappfloat precision issues?01:26
wpwrakmaybe. or just limited range of sin/cos01:27
wpwrakwell, we have bigger math issues. like modulo ...01:30
Hodappbah?01:30
wpwrakmodulo (in the PFPU) has rounding issues. the computation that's done inside is quite complex. and it shows ...01:31
wpwraksomething like  a mod b = a-b*int(a*(1/sqrt(b))^2)  if i remember correctly. 1/sqrt(x) is a PFPU operation. it's also used to construct regular division.01:33
GitHub192[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/5cb9228ec433c7bf648daef861f44e25e03c63ab04:06
GitHub192[board-m1/master] added Board revision histories. - Adam Wang04:06
wpwrakcladamw: if you tell me your preferred order, i can add a rule to the m1r4 makefile, so that one can create a "nice" pdf also without schhist04:07
stekernwpwrak: "lm32 != Mico32" <= what do you mean by that?04:09
cladamwwpwrak, the #19 i think would be good to change to #1. then power -> fpga decap. -> 4 fpga sheet -> reset -> else can be the rest. :-)04:10
cladamwwpwrak, how do like this idea ?04:11
cladamwwpwrak, yes, i like just to use 'make' to make whole pdf file. :-) You know i am not familiar with commands. :(04:12
GitHub39[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/b068ad6b8832e75f6461e3b2887f1bf105a69d0b04:18
GitHub39[board-m1/master] Ethernet.sch: dragged nETH_RESET up and put the resistor R174 (and GND) under it. - Adam Wang04:18
wpwrakcladamw: sounds good04:21
GitHub166[board-m1] adamwang pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/548392475b898a72069540527c4773da766d3b4b04:21
GitHub166[board-m1/master] modified notes text by Werner's better suggesttions. :-) - Adam Wang04:21
wpwrakstekern: argh. my brain fuzzy-matched that to a mix of microblaze (for the letters) and nios (for the length)04:22
wpwrakJia: so you were right. lm32 == "mico 32". we just never call it like that :) sorry about the confusion04:23
stekernheh, ok04:24
wpwrakstekern: thanks for the correction !04:25
GitHub107[board-m1] wpwrak pushed 1 new commit to master: https://github.com/milkymist/board-m1/commit/7bd8e418b2b3a764d85f469dee1287170d6e14ef04:35
GitHub107[board-m1/master] r4/Makefile: new targets "pdf" and "view" for overview PDF of schematics - Werner Almesberger04:35
stekernbtw, has anyone seen jpbonn lately? I noticed that the llvm-lm32 backend doesn't delete fallthrough conditional jumps (i.e. http://pastebin.com/GYN3xGbJ). I took a peek in AnalyzeBranch (where that should be handled) but he returns with a fail early and has added a comment: // The Cond parameters need to be fixed - disable this for now.04:37
stekerns/conditional/unconditional04:37
FallenouJia: by changing the lm32 ABI you mean modifying the toolchain ? gcc ?05:30
Fallenoumodifying the "function call" scheme ? the callee/caller saved registers , this kind of stuff ?05:31
JiaFallenou: yes, I'm trying fix the stack-unwinding bug.06:19
Jiayes, the callee/caller code.06:19
cladamwwpwrak, the way to generate pdf email is great, tks a lot. I'm using it. :-)07:28
cladamwwpwrak, you can update kicad-lib once try to make m1r4.pdf to get latest symbols.11:34
wpwrakcladamw in absentia: all files updated :)13:23
Hodappwheeee, I acquired my laser projector17:37
--- Wed May 9 201200:00

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