07:55 diagonal6x: It was actually published on #llvm so you have say 20 items per bank UpTo bound 1024, so let's pretend all items are 524, so to access 14th item you do 13*512 minus 500 is 12, then you have 13 times 12 and 7 times 500, so you then do 524-500buffer subtract then you have 13 times 512, 7times 12, then you access one time 512 from remainder just like the above (7times 500) and subtract 6times 500 from 7 times 500, so you accessed the element,
07:55 diagonal6x: order is the original order that you used, since sum gets values towards MSB. This was having some testing done, I managed to access the needed sequences, but happy Christmas, I founded a company where I write several compilers and data compression storage methods. But it was all published , Roman Lebedev seemed to have understood it on llvm, nothing is kept secret from you.
07:55 diagonal6x: You helped me, I help you when you needed it
10:47 towo`: hello, i don't know since when, but in debian unstable with mesa 23.3.1 vainfo does not show vp9 support anymore, is it known and is there a solution? card is a Radeon RX 6600 XT
12:54 towo`: oh, i see, the problem is only with mesa git, not with mesa 23.3.1 from debian unstable
12:55 Remco: In that case a bisect would be nice. There's a bunch of vp9 issues on gitlab still open. One of which got a firmware fix which is currently being tested
12:57 towo`: can do that, but not today, will do it tomorrow
18:28 diagonal6x: Today something wrong with the logs, php script comes instead of Christoph Brill's logs for the channels.
20:20 daniels: agd5f, hwentlan__: is https://patchwork.freedesktop.org/project/amd-xorg-ddx/series/ still being used at all or can I kill it please?
20:28 agd5f: daniels, still used
20:29 agd5f: mostly kernel patches at this point, hardly any ddx stuff
20:29 daniels: k, just noticed there wasn't anything to actually update the patch status so it has to be done manually
20:31 agd5f: can it look at kernel.org or something like that?
20:32 daniels: define 'look at'?