00:35 MaryTrampoline: You test my theorem according to simple linear algebra like this on duplicates which is enough simple. 441+114+57+114+57−75−75−57−57−57−57 is 405 and 405+405-75-75-57-57-57-57-84-270=78 and whichever sign you need, i advise plus sign dominance, since so you can use minus arithmetic with unsigned data type and fit more but on access it's just 135-78=135 or if you subtract 135 obviously
00:35 MaryTrampoline: it ends up as -78. You can test this calculation as much as you want, it always yields deterministic on fewer or higher number of duplicated members. Hence what you are is a delusion, and inappropriate nvidia pony puppets, who make zero sense in the world. Hence on plus dominance and unsigned data you do obviously 57+192+172+176−84−84−84−270=75 and 57+114−118−118−118+192+232+200=441,
00:35 MaryTrampoline: since all you ever do with the hash is single two and three and four times the value, and capture the bits by trapping them to basket. 441+57+114-75-75-57-57(aka twice the algebra needed on two 57's on twice the stored set) , it had never anything to do with FFT/DFT nor markov chains, this calculation is possible on two primary arithmetic cause of the numeral system's base repetition
00:35 MaryTrampoline: and constants like PI etc. However the flow is integrated into integral computations paragraph part and subunit called derivatives. It feels like you are sick people, you need to heal or cure your illnesses before terroring other people. I do not want your shit around me.
01:44 lobbyraindeer: I synthesized the whole mechanism for you, as to compilation and runtime works for data access, i add the specification for execution, the function of the sets/sequences are already seen from my flow, so the function is very easy, can be presented to mathematics archives, but you are fraud all the life, no one will react to this by giving any reward or favours back. So during the
01:44 lobbyraindeer: conflicts some people are already assigned death penalties to. It might seem from your attitude that more are assigned like Ryan and Dog ask one. What i got was wank spam and anal bragging by most vicious incapable leftovers possible, no one gave me anything to do something with no benefits, as injured money is earned by me in the most hardest way, after breaking RSA and declaring a war
01:44 lobbyraindeer: with my clans i just take what belongs to me by force. And those people who milked substances from my organism as well as diagnosed me mentally ill and delusion in the world, they are slaughtered to hell. We go after jack to end him, i gave him chanches, he responded nothing and is getting his death sentence soon.
03:57 MrBabySitter: I am fully prepared for a conflict and all my allies are. I understand when this mitochondrial dna change was not successful with the chip implanted onto my neck they come to lie about this anal whore to get under the skin of my allies, which seems like a small tweak with rape charges however, my lineup of allies is strong i have all the details of what has happened in life, and all the
03:57 MrBabySitter: evidence, none of those will ever be processed in court due to corruption in Estonia as well as in world. Do you think i miss any of the data, and you are so clever or what. I have fields of councilers, informators,i am one of the most strongest hacker and brilliant mind, i can tap anyone including FBI themselves.
06:54 mripard: airlied: sima: there's no patch in drm-misc-next-fixes this week, so there won't be any PR
06:54 sima: ack
07:54 mlankhorst: airlied: I see you posted a new version of the memcg patches?
08:04 airlied: mlankhorst: yeah not sure how much more churn will be needed to get them over the line
08:33 MrCooper: starting to wonder why I'm not getting any feedback on my atomic commit HW done event & deadline property series
09:07 karolherbst: anybody want to review the st_interop.c changes there? That interface is supposed to be backwards compatible, so I'd rather have proper review there. https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36249/diffs?commit_id=c825300c225bafc0429469fe453e38e423859eb6
09:22 mripard: sima: could you have a look at https://lore.kernel.org/dri-devel/20250728102809.502324-1-riana.tauro@intel.com/ ?
09:23 mripard: I'm triggered by any "vendor specific" API I think, I'd like your opinion on it
09:23 sima: mripard, I did comment on earlier patches and looked confused
09:23 sima: and yeah same
09:23 sima: but it looks like it's not the worst of the options
09:24 mripard: I'm not really opposed to it on principle, but then the documentation should be better than "run that script that we don't provide"
09:25 sima: yeah the doc patch doesn't look great tbh
09:25 mripard: I'll reply then
09:25 mripard: thanks :)
09:27 sima: the survivability_mode sysfs for the runtime case also doesn't look super documented
09:27 sima: but yeah personally I'd have leaned towards just enumerating them with vendor prefix, but didn't feel strongly
09:27 sima: so intel-fw-cs-error or something like that for this
09:27 sima: but yeah the docs should be a bit better, I agree
09:28 mripard: that's a good idea indeed
09:28 sima: yeah, people didn't like it much, but did consider it
09:28 sima: which I think for this one is good enough
09:37 sima: airlied_, oh dear the cherry-pick discussion
09:38 sima: I'm not sure we can explain it, it boils down to "not enough maintainers"
09:38 sima: I guess we should try to figure out how to be more careful with the merge resolutions, but dunno how
09:38 sima: maybe drag agd5f into drm-tip?
09:42 sima: also not sure why we recently seem to have screwed this up a few times
10:00 realsessioneight: You alternate your stirred saliva in terror and joke mode like bragging to hurt mentally over past injuring missioned, never having any self-worth moral nor ethics so that way no work will ever get done. So i explained the execution paradigm with an example of 371 in the logs, so 512 removed and all the reverse indexes, gets a negative -141 and 32-1 times 256 on two dependency banks
10:00 realsessioneight: offsets. -141-138 to combine and subtract from transition bank i.e answer sets, will result into reverse index base+transition +index sums be unchanged, and focal element in positive transition transvalue+base, so you want to add a new base to the positive while removing the no interest in sums. and subtract two of this base value, so that is opportunity to just a smaller add in net
10:00 realsessioneight: total, where the transition values of not needed combinations are removed just with delta , all works in two barrells. one inputing index+base and one inputing value to the needed hashes, hence the removal of leftovers becomes possible. where instruction nr2 banks non-focal sets are removed just by removing the twice of the bases again or three times base by indexandbase barrel. and
10:00 realsessioneight: focal value gets or azimuth value is pulled again to negative. where as triple base removed gets to -397 aka negative inverse index. So the proof system is similar to the data for execution, and i have had it for some time already in many forms, i will just finalize the code a bit this year ontop of wasmati. I no longer have seen how this could cause problems to real programmers, but
10:00 realsessioneight: whatever you seem to want to pick up a fight instead.
10:31 realsessioneight: I actually offered a set of methods to find the index from values and however needed based of such trick if i remember correctly https://www.youtube.com/watch?v=SFpcglY_P5Q , however i later refined it to fixed indexes without any of dividing, so to me this stuff does not take long to land this year.
16:43 anarsoul: are there any known issues with CI? jobs in https://gitlab.freedesktop.org/sarbes/mesa/-/pipelines/1481461 fail with HTTP 500
17:10 anonymix007[m]: Is data which is written to a GBM BO automatically converted on AMD or Intel? Fourcc is the same, so the only difference should be the modifier
17:57 mareko: controversial idea: renaming src/mesa to src/gl
17:59 emersion: that actually sounds like a good idea
18:00 emersion: always bugged me that a mesa folder existed in the mesa repo
18:14 karolherbst: we could go one step further and just make it a real frontend :P
18:16 mareko: also moving src/compiler/glsl to src/gl/glsl
18:25 mareko: tarceri: ^^
20:08 agd5f: airlied_, sima sent a PR with the DSC fix for Linus and a few other fixes
20:10 jenatali: Yeah if anything it seems like it should be src/gallium/frontends/gl
20:17 airlied: agd5f: thanks, will pass it on today
20:29 zmike: mareko: do you have further comments on https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36296 or ?
20:30 zmike: I would like to merge before I start getting conflicts
20:45 HdkR: Oh no, refcounting, my beloved.