| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.172.16> has quit IRC | 05:24 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.55.40> has joined #libre-soc | 05:25 | |
| *** Ritish <Ritish!~Ritish@27.7.57.151> has joined #libre-soc | 06:18 | |
| *** Ritish <Ritish!~Ritish@27.7.57.151> has quit IRC | 06:19 | |
| *** Ritish <Ritish!~Ritish@27.7.57.151> has joined #libre-soc | 06:20 | |
| *** Ritish <Ritish!~Ritish@27.7.57.151> has joined #libre-soc | 06:43 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:5533:3ca0:acee:7057> has joined #libre-soc | 06:45 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.55.40> has quit IRC | 06:47 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.40.26> has joined #libre-soc | 06:47 | |
| *** Ritish <Ritish!~Ritish@27.7.57.151> has quit IRC | 06:48 | |
| *** ritty <ritty!~Ritish@27.7.57.151> has joined #libre-soc | 06:50 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:5533:3ca0:acee:7057> has quit IRC | 06:52 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:afc2:4db:7336:3b7a> has joined #libre-soc | 06:53 | |
| *** ritty <ritty!~Ritish@27.7.57.151> has quit IRC | 06:54 | |
| *** psydroid <psydroid!~psydroid@user/psydroid> has quit IRC | 07:00 | |
| *** yambo <yambo!~yambo@069-145-120-113.biz.spectrum.com> has quit IRC | 07:00 | |
| *** sauce <sauce!~sauce@sauce.icu> has quit IRC | 07:00 | |
| *** sauce <sauce!~sauce@sauce.icu> has joined #libre-soc | 07:00 | |
| *** yambo <yambo!~yambo@069-145-120-113.biz.spectrum.com> has joined #libre-soc | 07:01 | |
| *** psydroid <psydroid!~psydroid@user/psydroid> has joined #libre-soc | 07:05 | |
| *** ritty <ritty!~Ritish@27.7.57.151> has joined #libre-soc | 08:28 | |
| *** ritty <ritty!~Ritish@27.7.57.151> has joined #libre-soc | 08:31 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:afc2:4db:7336:3b7a> has quit IRC | 08:31 | |
| *** ritty <ritty!~Ritish@27.7.57.151> has quit IRC | 08:32 | |
| *** ritty <ritty!~Ritish@27.7.57.151> has joined #libre-soc | 08:32 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:ab6c:530f:9209:3dcc> has joined #libre-soc | 08:51 | |
| *** ritty <ritty!~Ritish@27.7.57.151> has quit IRC | 08:54 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:ab6c:530f:9209:3dcc> has quit IRC | 10:08 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:39f7:3fb0:87d8:b07d> has joined #libre-soc | 10:08 | |
| *** ritish_ <ritish_!~Ritish@2401:4900:1f2d:bbee:39f7:3fb0:87d8:b07d> has quit IRC | 10:11 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.40.26> has quit IRC | 11:26 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.135> has joined #libre-soc | 11:26 | |
| *** alethkit is now known as nobody | 11:58 | |
| *** nobody is now known as alethkit | 11:59 | |
| lkcl | yes there exists (surpriiise) a debian package for that | 12:18 |
|---|---|---|
| lkcl | apt-mirror or something | 12:18 |
| lkcl | where you can then set the entries in /etc/apt/sources.list to point at the apt-mirror instead | 12:19 |
| lkcl | and yes, you reaallly don't want to be slowing down things by not using apt-mirror | 12:19 |
| lkcl | there exist several such packages btw, i don't know their differences, i used them over 10 years ago. i just remember they exist | 12:20 |
| markos | apt-mirror is great and very easy to setup indeed | 12:27 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.135> has quit IRC | 12:28 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 12:28 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 13:41 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.230> has joined #libre-soc | 13:42 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.230> has quit IRC | 16:16 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 16:16 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 18:51 | |
| *** gnucode2 <gnucode2!~gnucode2@user/jab> has joined #libre-soc | 19:17 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 19:21 | |
| sadoon[m] | <lkcl> "there exist several such..." <- I've used apt mirror before but got throttled / semi banned because I did parallel downloads lol | 21:05 |
| sadoon[m] | Using debmirror currently which seems to do the same thing | 21:05 |
| programmerjake | lkcl, can you move the budget from #1041 to #972? thx | 22:09 |
| markos | lkcl, I thoght libvpx was bad | 22:23 |
| markos | but libaom is worse | 22:23 |
| markos | in the sense that each function can have multiple versions due to particular coefficients that require totally different SIMD algorithms and unrolling | 22:24 |
| markos | well not each function | 22:24 |
| markos | but some quite used ones | 22:24 |
| markos | some convolution functions in particualr | 22:25 |
| programmerjake | well, it's designed for experimentation, not as much for high performance | 22:25 |
| markos | original function 15 lines, SSE/NEON ~450 | 22:26 |
| markos | true, but both Intel and Arm are keen on getting it as optimized as possible | 22:26 |
| markos | it's like a contest | 22:26 |
| markos | I don't have final benchmarks for libvpx but with latest patches we managed to get higher speed in 4k highbd encoding on graviton3 than the other intel graviton offerings | 22:27 |
| markos | I guess Arm wants to repeat that with libaom | 22:28 |
| programmerjake | imho if you have the choice, work on a av1 decoder/encoder actually designed from the beginning for high performance | 22:28 |
| markos | anyway, the point is that it would take a fraction of the time to finish the port to SVP64 | 22:28 |
| programmerjake | wait, intel has graviton offerings? | 22:28 |
| markos | sorry s/intel graviton offerings/aws x86 offerings/ | 22:29 |
| programmerjake | i thought that was purely amazon arm | 22:29 |
| programmerjake | ah | 22:29 |
| markos | but I'm actually getting paid by Arm to do NEON optimizations on libvpx/libaom :) | 22:29 |
| programmerjake | oh, fun... | 22:30 |
| programmerjake | well, good luck :) | 22:30 |
| markos | well, it's not that bad, it's quite rewarding to see your code upstreamed in a project like that | 22:30 |
| markos | and tbh I'm glad it's so hard because they hire me to do it :) | 22:31 |
| programmerjake | yeah, i was pretty happy to get my code into wine and libsdl | 22:31 |
| markos | but having worked on most major SIMD platforms out there -except MIPS MSA- I'm just amazed at how much simpler it is to stuff on SVP64, even on assembly | 22:32 |
| programmerjake | though it wasn't nearly as difficult | 22:32 |
| markos | s/to stuff/to do stuff/ | 22:32 |
| markos | btw regarding the av1 decoder part, they're already doing the dav1d decoder themselves | 22:33 |
| markos | when av1 becomes completely stable, they're going to fork that to av2 and want to have most optimizations for NEON already in there before that happens | 22:34 |
| markos | libaom av1 that is | 22:34 |
| markos | anyway | 22:34 |
| programmerjake | they haven't already started on av2? | 22:35 |
| markos | in specs yes, code not afaik | 22:35 |
| programmerjake | i'd expect them to start right after releasing av1 | 22:35 |
| markos | well, perhaps they have some older fork, but the goal is to sync and propagate all the optimizations to the av2 fork | 22:36 |
| markos | I'm not following av2 tbh | 22:36 |
| programmerjake | ah | 22:36 |
| markos | so there is some pressure to finish everything before the summer | 22:36 |
| programmerjake | maybe av3 will use machine learning for the decoder/encoder instead of dcts/arithmetic encoding | 22:37 |
| markos | if we had svp64 hardware to play with, I would probably finish a complete port in much less time | 22:37 |
| markos | :D | 22:37 |
| markos | yes, and require a 4090 Ti and 900Watts to do the decoding | 22:38 |
| programmerjake | i tried to build a machine learning lossless audio codec but never finished. | 22:38 |
| programmerjake | not necessarily, some simpler algorithms can run on much weaker hw, e.g. a cellphone | 22:39 |
| markos | you just do different kinds of LA calculations, ML is basically tons of LA and graph theory | 22:40 |
| markos | perhaps it would be quicker in finding common patterns in most videos and encode them faster | 22:41 |
| markos | ah sorry you said audio | 22:41 |
| markos | still | 22:41 |
| programmerjake | in case you're curious: https://github.com/libsdl-org/SDL/commit/25e3a1ec90cbc08acbb1d33668ad71e6ca241e05 | 22:41 |
| markos | this is a huge patch :) | 22:42 |
| programmerjake | yeah, the sdl maintainer squashed a few weeks of work into 1 commit | 22:43 |
| markos | still undecided whether squashing huge commits is a good thing in the end, sure it makes things easier to rollback in a single move | 22:44 |
| markos | I've worked in some projects with squashing and it did make some things easier | 22:45 |
| programmerjake | here's the full unsquashed history: https://github.com/programmerjake/SDL | 22:45 |
| programmerjake | my basic idea for a ml audio codec is to build a 100% reproducible neural net that predicts the next audio sample and store the difference between the prediction and the actual value | 22:48 |
| programmerjake | kinda like FLAC except with a different predictor | 22:49 |
| programmerjake | we specifically rebased all our sdl commits so they would properly separate who did what so i was a little disappointed they all got squashed without even a Coauthored-By tag | 22:52 |
| markos | yeah that sucks | 22:56 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!