Sunday, 2016-03-20

*** tpb has joined #timvideos00:00
*** Bertl is now known as Bertl_zZ00:30
*** CarlFK has joined #timvideos02:28
*** ChanServ sets mode: +v CarlFK02:28
CarlFKmithro: did you see the video I posted showing the dotty color bars on the TV here?02:31
*** panther has joined #timvideos02:32
*** panther is now known as sab_12302:32
sab_123so i have set up everything in xilinx02:32
tpbTitle: GitHub - RacingTornado/JPEG_GSOC: A GSOC project for TimVideos (at
sab_123also looks like eveyrthing is done on 16x8 blocks. We can change this02:47
mithrosab_123: There isn't anything useful in that repository?03:03
sab_123I just set it up03:03
sab_123as a useful standalone03:03
sab_123text bench03:03
sab_123test bench03:03
sab_123I dodnt want to mix the 203:04
sab_123and I used the mkjpeg core directly03:04
sab_123It would take some time for me to look into the architecture03:04
sab_123He has a very good DOC03:04
sab_123But before I commit to working on this03:04
sab_123I would like to first make sure that there is nothing which is closer to my project which I can work on03:05
sab_123Can you think of anything?03:05
mithroYou could work on improving the USB stuff on the FX2 maybe?03:09
sab_123The firmware?03:09
sab_123I dont use fx2lib03:09
sab_123I use the cypress code03:09
sab_123the bulk loop transfer03:10
sab_123but yes03:10
sab_123I guess that would be closer03:10
sab_123I was talking to makestuff(lx9)03:10
sab_123he told me he gets 45MB/s03:10
mithroYes we get similar performance03:12
sab_123what would you like me to improve?03:12
mithroGo take a look at the issues on misoc firmware repo and see what you can dig up related to USB03:14
sab_123I did03:14
sab_123There is one for USB 3.003:15
sab_123I really dont want to get into that since testing and validating the PHY itself is going to take a lot of time03:15
sab_123But i will look an non GSOC stuff now03:15
mithroDid you actually look at the HDMI2USB misoc firmware repo?03:16
sab_123I used a filter so I only got like 5-10 issue03:16
sab_123I see quite a few interesting one now03:17
sab_123they are different repos03:19
sab_123I feel stupid now03:20
mithroI'm afraid I don't memorise issues numbers :P03:24
tpbTitle: HDMI2USB-misoc-firmware/firmware/fx2 at fx2-refactor · mithro/HDMI2USB-misoc-firmware · GitHub (at
sab_123So you want me to refactor and clean this up correct?03:25
tpbTitle: Rewrite the FX2 firmware to be understandable · Issue #149 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroThat would be a start03:29
sab_123that looks interesting03:31
mithroHave a look at how I've changed the firmware03:38
mithroMost of that is going to C development however03:38
sab_123I agree03:39
mithroYou might want to look at my fx2lib fork too03:40
tpbTitle: Issues · mithro/fx2lib · GitHub (at
*** rohitksingh has joined #timvideos04:21
*** CarlFK has quit IRC04:34
sab_123you want me to build SDCC myself?04:38
sab_123or ubuntu one is okay?04:39
sab_123the linux one is this04:40
sab_123SDCC : mcs51/gbz80/z80/z180/r2k/r3ka/ds390/pic16/pic14/TININative/ds400/hc08/s08 3.3.0 #8604 (Dec 30 2013) (Linux)04:40
mithroyou want the newest sdcc05:18
sab_123can i use a snapshot05:23
sab_123either way05:23
sab_123its okay05:23
sab_123for the GSOC proposal05:23
sab_123just want ot understand what you need05:23
sab_123Comment the code05:23
sab_123clean it up05:24
sab_123Anything else05:24
sab_123Because the date is 25th05:24
sab_123So I would like to start writing it up05:24
mithroClean up and comment code is not big enough for a GSoC project05:29
mithroWith the FX2 stuff you are going to have to be more self directed and figure out some strong goals based on the open issues05:30
mithroThe high level goal at the end of your project should be "we have a really awesome USB2.0 interface which works well, is well tested, is easy to hack on and provides the needed HDMI2USB functionality" -- how you get to that is going to require some though05:32
*** rohitksingh has quit IRC06:09
*** rohitksingh has joined #timvideos07:31
*** rohitksingh has quit IRC08:54
*** Bertl_zZ is now known as Bertl09:03
*** kflux has joined #timvideos11:13
*** CarlFK has joined #timvideos11:19
*** ChanServ sets mode: +v CarlFK11:19
CarlFKmithro: did you see the video of the TV?
*** CarlFK has quit IRC13:54
*** sab_123 has quit IRC14:01
*** rohitksingh has joined #timvideos14:31
*** GChriss has joined #timvideos14:32
*** kflux has quit IRC16:19
*** se6astian|away is now known as se6astian16:32
*** GChriss has quit IRC16:33
*** CarlFK has joined #timvideos17:22
*** ChanServ sets mode: +v CarlFK17:22
CarlFKtumbleweed: lightning talks in 144!17:30
*** rohitksingh has quit IRC17:30
*** GChriss has joined #timvideos17:50
*** kflux has joined #timvideos19:15
*** kflux has quit IRC19:32
mithroCarlFK: yes19:36
mithroGChriss: ping?19:36
CarlFKmithro: I have checked out of my room with that TV, so can't do anything more there.19:41
CarlFKmithro: I have my Opsis board and found a teacher that is excited about our GSOC fpga stuff19:42
mithroCarlFK: cool19:42
mithrohave you been pimping GSoC?19:42
CarlFKmithro: oh yeah, good morning.  hope you got some sleep19:42
mithroCarlFK: not really, I kept worrying that I would miss waking up19:43
CarlFKI did a lightning talk: "this is the board, this is what it does, GSOC is $5k for students to work on OS, who's interested talk to me!"19:43
CarlFKmithro: quite a bit has been added to
tpbTitle: FSFE Etherpad: public instance (at
mithroGChriss: is there anything in the room now?19:53
mithroGChriss: IE when can we start testing the set up?19:53
mithroGChriss: I'm just getting things going on this side, will probably take me 20-30 minutes19:53
mithroSilvia is on her way here too19:53
CarlFKmithro: there should be a talk in there.  /j #libreplanet_room15519:57
GChrissmithro: pong19:58
GChrissthere's breaks between sessions19:59
GChrissI can plug in my laptop (videoconference laptop) and sit there now if that helps20:00
GChrissthen try calls20:00
GChrissok going to the room now20:07
*** GChriss has left #timvideos20:07
*** GChriss1 has joined #timvideos20:12
GChriss1ok in 52--155 now20:13
CarlFKGChriss1: I'll be there as soon as the Blender talk breaks for Q&A20:13
*** GChriss1 is now known as GChriss20:14
*** ssk1328 has joined #timvideos20:18
mithroGChriss: ping?20:29
mithroI'm now in the room we will be using here20:29
GChrissmithro: pong20:30
GChrissdo you have a ringID?20:30
GChrisssilvia does, can help guide/setup if needed20:31
tpbTitle: Ring | Ring gives you a full control over your communications and an unmatched level of privacy. (at
mithroOkay, she just arrived20:31
GChrissit looks like the panel being held now is wrapping early20:32
* GChriss started the clap20:33
GChrissbrb again, changing internet cable to front table20:34
*** GChriss has left #timvideos20:34
CarlFKon way to room20:38
*** CarlFK has quit IRC20:39
*** skay has joined #timvideos20:47
*** ChanServ sets mode: +v skay20:47
skaymithro: they're still working on it20:47
skayck here.  I think there are at most 5 people in the room that are here for the talk20:50
skayand 10 or so that haven't left fromt he previos talk and need a place to sit.20:50
skayskay here20:50
skayCarl has self deprecating humor20:50
*** GChriss has joined #timvideos20:51
tumbleweedit's a tough slot to be scheduled in20:51
*** GChriss has quit IRC21:02
*** ssk1328 has quit IRC21:06
*** ginger has joined #timvideos21:07
gingersilvia here - we're trying to set up audio output21:07
skaytumbleweed: are you doing pypy stuff?21:11
*** ginger has quit IRC21:16
skayyay tim21:16
*** ginger has joined #timvideos21:16
tumbleweedskay: yep21:20
tumbleweedspot the mandelbrots? :)21:21
skaytumbleweed: yep21:23
* tumbleweed broke pypy in Debian, now I'm trying to unbreak it21:24
tumbleweedactually, upstream broke it, by making it bigger, so it doesn't build on 32bit hardware any more21:24
*** ginger has quit IRC21:35
*** ginger has joined #timvideos21:35
*** ginger has quit IRC21:42
*** ginger has joined #timvideos21:43
*** gb_mks has joined #timvideos21:49
*** se6astian is now known as se6astian|away21:51
*** ginger has quit IRC21:55
*** gb_mks has quit IRC22:23
xfxftumbleweed: nobody should be using 32-bit arch's anymore anyway, right? ;-)22:39
tumbleweedthis is only an issue in Debian, on armhf22:40
tumbleweedevery other 32bit arch is actually built on 64bit kernels22:40
tumbleweed(or has no pypy JIT, and still fits...)22:40
xfxfwierd, the bug is kernel dependent?22:40
tumbleweedno, it just needs lots of address space22:41
tumbleweedif you have to share your 4GiB with the kernel, you don't have enough22:41
xfxfkinda mindblowing it needs that much space to build, but makes sense, i guess22:42
mithroThere are plenty of reasons to use 32bit hardware22:49
tumbleweedeven 16 or 8 bit22:49
xfxfmithro: i hope it was obvious i was being facetious22:49
tumbleweedany the good news for pypy is that pypy can build pypy inside 3GiB of address space, with some tweaking22:50
mithro16bit is a bit of a strange one, 8bit still makes sense for the *very* small things22:50
mithroAnywhere you could use 16bit is almost always better to go with 32 or 64bit22:50
tumbleweedprobably true22:51
xfxfwhat modern archs are 16 bit?22:57
xfxfi can't recall the last time i've seen one - plenty of 8/32/64 bit22:57
mithroxfxf: exactly :P22:58
mithroxfxf: Can you take a look at deeprave's pull request? ?22:59
tpbTitle: Make the debug command be more deterministic about debug state. by deeprave · Pull Request #214 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
xfxfmithro: he's been sick, which is why he hasn't responded23:00
xfxfi shall poke IRL23:00
mithroxfxf: No, I don't need you to poke him - I need you to test his change on hardware23:00
xfxfoh, right.  yeah, can do in the next few days23:01
mithroxfxf: If it works as he describes, then I can merge it - it looks good from a visual inspection point of view23:02
mithrosab: How is your GSoC application going?23:02
*** Bertl is now known as Bertl_zZ23:40

Generated by 2.13.1 by Marius Gedminas - find it at!