*** tpb has joined #timvideos | 00:00 | |
*** rohitksingh has quit IRC | 00:00 | |
mithro | mparuszewski: I think we should get the output of the _generate scripts sent somewhere (or at least when it changes) | 00:01 |
---|---|---|
aps-sids | planet.timvideos.us is still being redirected | 00:03 |
aps-sids | morning mithro | 00:04 |
mithro | mparuszewski: we also need that monitoring | 00:04 |
mithro | aps-sids: you have a happy flumotion component? | 00:06 |
aps-sids | mithro: yes, videotest producer (only that is ported yet) | 00:14 |
aps-sids | I'm using flumotion-admin of a working flumotion installation to view this. | 00:15 |
mithro | aps-sids: commented on your blog post. | 00:17 |
mithro | mparuszewski: I think the way we have the GitHub pages means code.timvideos.us and planet.timvideos.us are fighting? | 00:18 |
aps-sids | mithro: there is one problem with the tests (both on my machine and travis) that since some modules are ported and others are not, running all tests (using trial flumotion.test) freezes, probably due of import conflicts but individual tests run (and pass) just fine. | 00:27 |
aps-sids | https://travis-ci.org/aps-sids/flumotion-orig#L2330 | 00:27 |
tpb | Title: Travis CI - Free Hosted Continuous Integration Platform for the Open Source Community (at travis-ci.org) | 00:27 |
mithro | aps-sids: I think trial can take a list of things to run? | 00:34 |
aps-sids | mithro: I can add names of only ported modules in .travis.yml so only those tests are run? | 00:35 |
mithro | aps-sids: maybe just a "working-tests.txt" or similar file? so people can run it locally too? | 00:35 |
aps-sids | sure | 00:35 |
*** andi-m2_ has quit IRC | 00:43 | |
*** andi-m2 has joined #timvideos | 00:50 | |
aps-sids | mithro: replied on comment | 00:58 |
CarlFK | aps-sids: are you able to play a test stream in a video player? (like vlc) | 01:02 |
aps-sids | CarlFK: encoder and streamer are not ported yet, so that won't work. As I mentioned in my post, I'm going to do it in this week. | 01:03 |
CarlFK | what about the original unported code? | 01:03 |
aps-sids | test stream works for completely unported code, not partially ported. I will post an instructions post soon. | 01:05 |
*** ayush3504 has quit IRC | 02:06 | |
shenki | mithro: looks like the amazon billing thing worked this time | 02:39 |
shenki | mithro: any tips on setting up a machine? | 02:39 |
*** mparuszewski has quit IRC | 02:41 | |
mithro | shenki: we should use 14.04 as the base version I think? | 03:01 |
*** ayush3504 has joined #timvideos | 03:14 | |
shenki | mithro: yep | 03:41 |
shenki | mithro: is it more expensive to run an instance in sydney? | 03:41 |
shenki | hrm, pv vs hardware | 03:41 |
shenki | is there a 'hackers guide to aws'? | 03:46 |
shenki | it would make sense to use the free tier to install the tools, and then run it on the bigger ones when we're doing a build | 03:47 |
shenki | tija: hello | 03:48 |
mithro | Yes but I don't know by your much. The latency improvements are probably worth it. | 03:58 |
mithro | S/your/how/ | 03:58 |
shenki | ok | 04:09 |
shenki | hrm, it's a bit opaque. i we want one really fast vcpu, not many of them | 04:10 |
*** FeltonChris has joined #timvideos | 04:11 | |
mithro | shenki: I assume Xilinix stuff doesn't have an option to try and parallelize stuff? | 04:54 |
shenki | mithro: you can run the multi-threaded par algorithims, yeah | 04:56 |
shenki | mithro: they're different altorithims, so different outcomes to the single threaded ones | 04:57 |
shenki | afaik | 04:57 |
shenki | we should make sure everyone is using them if we switch to them | 04:57 |
shenki | tija: do you have experience using the -mt options | 04:57 |
shenki | ? | 04:57 |
mithro | tija / shenki: you guys should post about the UART stuff | 05:09 |
shenki | mithro: agreed. i think it's on the way | 05:14 |
*** mparuszewski has joined #timvideos | 05:18 | |
mparuszewski | mithro: planet and getting started are fighting, because planet folder on timvideos.github.io can behave in 2 ways: 1. It can be a folder inside timvideos | 05:20 |
mparuszewski | Timvideos.github.io repository | 05:20 |
mparuszewski | But planet folder on timvideos.github.io can also be a website from gh-pages in project planet. | 05:21 |
mparuszewski | Sorry that I answered so late, but I was in bed at the time you wrote the message. :) | 05:21 |
mparuszewski | Today I will be fixing both repos, and find a monitoring tool that will suits our needs. Maybe Google Analytics can do that. Definitely we need something to ping if website is up and check if content is loading properly. | 05:23 |
*** ayush3504 has quit IRC | 05:28 | |
mithro | There is uptimerobot for something very quick | 05:41 |
*** Niharika has joined #timvideos | 06:04 | |
*** Niharika has left #timvideos | 06:04 | |
*** tija has quit IRC | 06:06 | |
*** tija has joined #timvideos | 06:20 | |
*** slomo has joined #timvideos | 06:51 | |
*** slomo has joined #timvideos | 06:51 | |
mithro | mparuszewski: I came up with another solution for doing the run-one stuff in _generate.sh https://github.com/timvideos/planet/tree/using-flock -- from http://www.kfirlavi.com/blog/2012/11/06/elegant-locking-of-bash-program/ and looking at the run-one code at https://code.launchpad.net/~ubuntu-branches/ubuntu/utopic/run-one/utopic | 06:59 |
tpb | Title: timvideos/planet at using-flock ยท GitHub (at github.com) | 06:59 |
mithro | shenki: did you get anywhere with setting up that machine? | 07:00 |
mithro | shenki: and we should set up a way to idle timeout the machines too | 07:00 |
mparuszewski | mithro: I will be at home about 4 pm (+2 gmt), I will read everything and try to find perfect solution. :) I am happy that everything works fine now, thanks. | 07:04 |
*** ayush3504 has joined #timvideos | 07:06 | |
mithro | I figured we probably want to make it so that if you run the script manually the crontab doesn't clobber it | 07:08 |
mithro | hrm, no rohit... | 07:08 |
mithro | ayush3504: do you have your power back now? | 07:09 |
ayush3504 | mithro: yep. i'm there! | 07:09 |
mithro | just wondering where rohit is... | 07:09 |
mithro | hold on a sec | 07:09 |
mithro | ayush3504: do you have power back on, IE should we do a normal hang out? Or should I call your phone? | 07:10 |
ayush3504 | mithro: normal. i've joined | 07:10 |
mithro | okay | 07:10 |
mithro | if we need to do phone, I need to use an alternative room | 07:10 |
ayush3504 | ok | 07:11 |
shenki | mithro: no, i'll do some more on it tonight | 07:29 |
shenki | mithro: i was thinking of setting up jenkins or something on a !aws machine | 07:29 |
shenki | mithro: and that could fire up the aws machine, build the system, suck out the logs and bitstream, and shut it back down | 07:30 |
*** rohitksingh has joined #timvideos | 07:43 | |
rohitksingh | mithro: hi! | 07:44 |
mithro | rohitksingh: You get the time wrong? | 07:45 |
rohitksingh | I just woke up now! Was up all night! :( Had slept at 6AM | 07:45 |
mithro | rohitksingh: do you want to jump into the VC? | 07:50 |
rohitksingh | Yes! | 07:50 |
*** sarwarc_ has joined #timvideos | 08:04 | |
aps-sids | mithro: Any idea about this error? "GError: could not link identity to feeder:default-pay" | 08:22 |
aps-sids | This is the pipeline code is creating and using, "videotestsrc is-live=true name=source ! identity name=identity silent=TRUE ! video/x-raw, format=(string)I420, width=(int)320, height=(int)240, framerate=(int)5, pixel-aspect-ratio=(int)1 ! gdppay name=feeder:default-pay ! multifdsink sync=false name=feeder:default buffers-max=500 buffers-soft-max=450 | 08:22 |
aps-sids | recover-policy=1" | 08:22 |
mithro | aps-sids: yes - but currently in a meeting with rohit | 08:23 |
mithro | will talk about it when I finish | 08:23 |
aps-sids | ok, no problem | 08:23 |
mithro | rohitksingh: you get disconnected? | 08:37 |
*** rohitksingh has quit IRC | 08:39 | |
*** rohitksingh has joined #timvideos | 08:39 | |
*** sarwarc_ has quit IRC | 09:05 | |
*** rohitksingh has quit IRC | 09:06 | |
*** rohitksingh has joined #timvideos | 09:18 | |
*** sarwarc_ has joined #timvideos | 10:00 | |
ayush3504 | rohitksingh: are you making a 2 layer PCB or 4 ? | 10:19 |
rohitksingh | ayush3504: 2 layer! | 10:22 |
ayush3504 | just a note..if you take a look at the info about power supply rails on page 10 of AD9984A datasheet, you will see that it requires you to have ground plane. In general chips with multiple supplies, esp mixed mode chips require you to have that | 10:25 |
ayush3504 | but you can get away with that if you make sure that the supplies in your circuit are star connected. This star ground is usually at the negative terminal of the bypass capacitor.. | 10:26 |
ayush3504 | this may help : http://www.analog.com/library/analogdialogue/archives/46-06/AD46-06_FIG_02.jpg | 10:28 |
rohitksingh | yes, there would be a ground plane below and in vicinity of the AD9984A. I'll look into the star ground thing. Never knew about it. | 10:28 |
rohitksingh | many thanks for the help! :) | 10:29 |
ayush3504 | no prob! :) | 10:29 |
rohitksingh | I think i'm beginning to understand this star ground. Will see more | 10:33 |
ayush3504 | basically the idea behind a star connection is that the current flowing to ground from one circuit does not create voltage drop in other circuit. in simple words, the ground currents of different circuits get isolated from one another with star connection | 10:36 |
ayush3504 | by ground current i mean the current going from the respective power supply of a circuit to the ground | 10:37 |
ayush3504 | i got to learn this when I had trouble running some ADC that was too sensitive to noise. Even your chip looks quite sensitive to grounding issues | 10:39 |
rohitksingh | yeah, it is quite sensitive to voltages as well has noise due to its very high bandwidth | 10:47 |
ayush3504 | see if there are reference designs based on that chip, that should you out with noise problems. | 10:58 |
rohitksingh | there is one based on AD9984, and I'm using AD9984A. Mostly similar. It is there in the 'docs' folder of the "HDMI2USB-vmodvga" repository | 11:05 |
rohitksingh | although its not a complete design, just schematic... | 11:08 |
ayush3504 | Yes, I guess you can stick to that schematic except for the digital input part that you will be customizing. For the regulator part that you mentioned, I would prefer sticking to the power supplies given there in the reference design. It's generally a benefit for the manufactuers to show that their chips needs less extra parts so I would assume that those parts are necessary at least in the first few prototypes of the board. | 11:15 |
*** FeltonChris has quit IRC | 11:17 | |
*** FeltonChris has joined #timvideos | 11:19 | |
rohitksingh | yeah! great suggestions! Will not be changing the power supplies atleast until whole project is almost done | 11:25 |
ayush3504 | thanks and btw. i found some PCB layout suggestions in pg 40 of the datasheet. | 11:27 |
mithro | rohitksingh: http://www.elmac.co.uk/pdfs/Lord_of_the_board.pdf | 11:45 |
rohitksingh | ayush3504: yeah i read those guidelines. will implement them | 11:52 |
rohitksingh | mithro: thanks for the link! Awesome information along with illustrations! <Thumbs up> | 11:52 |
rohitksingh | mithro: I've addressed the issues in the draft comments and published the weekly overview here along with pictures of the board http://dreamsxtrinsic.blogspot.in/2014/06/gsoc-weekly-overview-good-momentum.html | 11:56 |
tpb | Title: [GSoC Weekly Overview]: Good Momentum! | Dreams eXtrinsic (at dreamsxtrinsic.blogspot.in) | 11:56 |
*** Niharika has joined #timvideos | 12:03 | |
*** tija has quit IRC | 12:09 | |
*** Niharika is now known as Niharika|afk | 12:37 | |
*** tija has joined #timvideos | 12:42 | |
*** Niharika|afk has quit IRC | 12:43 | |
*** sarwarc_ has quit IRC | 12:50 | |
*** FeltonChris has quit IRC | 13:10 | |
*** ayush3504 has quit IRC | 13:15 | |
*** ayush3504 has joined #timvideos | 14:32 | |
shenki | so a xilinx ise isntall is 18GB on my machine | 14:47 |
shenki | which is annoying, as there is no way you want to dynamicly install that on a vm to do a build | 14:48 |
shenki | so i wrote a fuse filesystem that records all the files that get opened in the ISE install dir when building the firmware | 14:48 |
shenki | and copied all those files out | 14:48 |
shenki | du -sh | 14:48 |
shenki | 206M | 14:49 |
shenki | very hacky - i assume if eg. a different vhdl package is used then it will fail | 14:49 |
shenki | but for 1% of the file size, it's a bit of a win | 14:50 |
shenki | in the time that i managed to do all that, the ISE download to the ec2 VM is still only at 61% | 14:50 |
shenki | mithro: ^ | 14:51 |
aps-sids | mithro: around? | 15:13 |
CarlFK | it is 1am, lets hope he is asleep ;) | 15:15 |
aps-sids | yeah :D | 15:15 |
*** rohitksingh1 has joined #timvideos | 15:18 | |
*** rohitksingh has quit IRC | 15:21 | |
*** tija has quit IRC | 15:43 | |
*** Niharika has joined #timvideos | 16:05 | |
*** sarwarc_ has joined #timvideos | 16:05 | |
*** sarwarc has joined #timvideos | 16:06 | |
*** CarlFK has quit IRC | 16:47 | |
*** Niharika has left #timvideos | 16:48 | |
ayush3504 | shenki: That's interesting. Did you try running the implement for HDMI2USB? I'm wondering how fast it goes on the EC2, takes ~15 min on my machine. | 17:02 |
*** CarlFK has joined #timvideos | 17:11 | |
*** ChanServ sets mode: +v CarlFK | 17:11 | |
*** slomo has quit IRC | 17:30 | |
* sarwarc is away for some time | 18:06 | |
*** slomo has joined #timvideos | 18:10 | |
*** slomo has joined #timvideos | 18:10 | |
*** sarwarc has quit IRC | 18:52 | |
*** sarwarc has joined #timvideos | 18:53 | |
*** ayush3504 has quit IRC | 19:14 | |
*** sarwarc has quit IRC | 19:25 | |
*** ayush3504 has joined #timvideos | 20:10 | |
*** ayush3504 has quit IRC | 20:16 | |
*** slomo has quit IRC | 20:33 | |
*** tija has joined #timvideos | 20:59 | |
*** tija_ has joined #timvideos | 21:07 | |
*** tija has quit IRC | 21:10 | |
*** rohitksingh has joined #timvideos | 21:20 | |
*** rohitksingh1 has quit IRC | 21:22 | |
*** CarlFK has quit IRC | 22:45 | |
*** tija_ has quit IRC | 22:57 | |
*** CarlFK has joined #timvideos | 22:57 | |
*** ChanServ sets mode: +v CarlFK | 22:57 | |
*** rohitksingh has quit IRC | 23:27 |
Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!