gnuiyl has quit [Remote host closed the connection]
fab has joined #dri-devel
gnuiyl has joined #dri-devel
coldfeet has joined #dri-devel
kts has quit [Quit: Leaving]
sghuge has quit [Remote host closed the connection]
sghuge has joined #dri-devel
fab has quit [Ping timeout: 480 seconds]
f_ is now known as funderscore
Lynne has quit [Ping timeout: 480 seconds]
Lynne has joined #dri-devel
glennk has joined #dri-devel
kts has joined #dri-devel
dsimic is now known as Guest4168
dsimic has joined #dri-devel
Guest4168 has quit [Ping timeout: 480 seconds]
kts has quit [Ping timeout: 480 seconds]
<mareko>
the instruction type of sysval loads could be changed from intrinsic to ALU, so that nir_opt_algebraic can lower and optimize them
<mareko>
alternatively, nir_opt_algebraic could support intrinsics returning an immutable value
YuGiOhJCJ has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
Duke`` has joined #dri-devel
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
Net147 has quit []
Net147 has joined #dri-devel
clfbbn has quit []
jsa1 has joined #dri-devel
<glehmann>
the second option sounds a lot saner. alu instructions should really only be alu, with the output only depending on the sources. We only just removed the ddx/ddy fake alu instructions, which caused a lot of pain everywhere.
tanbulian has joined #dri-devel
jsa1 has quit [Ping timeout: 480 seconds]
warpme has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
clfbbn has joined #dri-devel
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
guludo has joined #dri-devel
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
Net147 has quit []
Net147 has joined #dri-devel
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
haaninjo has joined #dri-devel
ity1 has quit []
ity has joined #dri-devel
kaiwenjon has joined #dri-devel
coldfeet has quit [Quit: Lost terminal]
coldfeet has joined #dri-devel
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
tanbulian has left #dri-devel [#dri-devel]
kaiwenjon has quit [Quit: WeeChat 3.8]
kaiwenjon has joined #dri-devel
clfbbn has quit [Remote host closed the connection]
clfbbn has joined #dri-devel
tanbulian has joined #dri-devel
tanbulian has left #dri-devel [#dri-devel]
tanbulian has joined #dri-devel
<tanbulian>
So the first user of specialized decoding which decodes a certain double encoding and is somewhere between super execution and decoding is division by two.
<tanbulian>
And is a carefully crafted procedure as tweaks are playing with operands of subtract and add again, but is generally meeting success here on such sequences.
warpme has quit []
fab has joined #dri-devel
fab has quit []
fab has joined #dri-devel
<tanbulian>
Time flies I should say is fascinating project but have not gotten as far to speak about real results. However real big number to small encoding and decoding is anticipated to be done on DMA ddr read-write-modify bursts. I investigated the general asm code property graph c++ code of wasmati, so java was only for delegative plugin system. I mean i prapared all my life to do it, but it's
<tanbulian>
tough mission too still, but have early success on all module specification and calculator testing.
<tanbulian>
I am jammed with time , have so many requests and things to do, so in the past i hoped to get collaboration with you, so i push the general specification and we implement the things continuously together.
coldfeet has quit [Remote host closed the connection]
sima has joined #dri-devel
<tanbulian>
I consider this as lifetime work to pimp up systems such as linux , it could take several decades yet to synthesize a real system out of linux.
<tanbulian>
but those all efforts though possible per results are not paid at, so i have to side with things to get some food too, but as of now the vision to compute heaven is clear.
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
fab has quit [Quit: fab]
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
clfbbn has quit []
clfbbn has joined #dri-devel
fab has joined #dri-devel
clfbbn has quit []
<tanbulian>
so the Cornell report is for representation of powers of two's in less lossy format. Uncle said there is no negative value logarithm and no imaginary numbers, sure if you want to operate on floating point, but my goal was set to generate indexes of powers in fixed integer format and hence get the opposite end of the scale, not under 1 but over, so one is multiply based exponention and other
<tanbulian>
division based rooting if we talk about positive vs negative logarithms, and markov chain is it not, markov chain is not something special it is a time scaler of time scale in derivatives.
clfbbn has joined #dri-devel
<tanbulian>
hence that is why it works well too, you were correct to anticipate after what i look for vaguely, but markov chain was preceded centuries back by original presentation in higher maths.
haaninjo has quit [Ping timeout: 480 seconds]
<tanbulian>
there are rules in the derivatives signs cos tan sin time scale, markov chain is derived from there imo.
<tanbulian>
derivative is such a thing that you know the past due to such segments of time scale, and you influence or predict the future through your functions based of the past.
fab has quit [Ping timeout: 480 seconds]
<tanbulian>
negative logarithms there are derivatives with fixed numbers planted as a clock imo, it is not yet inetgral computation by defintion but very close.
clfbbn has quit []
Net147 has quit [Quit: Quit]
<tanbulian>
time as dimension is invented as other systems, time does not exist in universe, it's measuring apparatus about events that take place in sequence is very comfortable way of describing something that took and will take place.
rgallaispou has joined #dri-devel
fab has joined #dri-devel
<tanbulian>
i asked the googles ai engine abacus ai or chat gpt, and they claim that java/scala code is not quite correct on sequences their entropy is large .
Net147 has joined #dri-devel
coldfeet has joined #dri-devel
<tanbulian>
binary entropy function is nicely documented in wikipedia, and seems to be correct, the probability theory is float based however it takes distribution that can be fixed integers too.
tanbulian has left #dri-devel [#dri-devel]
tanbulian has joined #dri-devel
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
Net147 has quit []
rasterman has joined #dri-devel
Daanct12 has quit [Quit: WeeChat 4.4.4]
Net147 has joined #dri-devel
Net147 has quit []
Net147 has joined #dri-devel
haaninjo has joined #dri-devel
haaninjo has quit []
<tanbulian>
I also speculated that those numbers repeating might be under the hood cause of PI constant, but it matters not all to go so deep.
Net147 has quit [Quit: Quit]
jannh has quit [Ping timeout: 480 seconds]
Net147 has joined #dri-devel
fab has quit [Quit: fab]
kts has joined #dri-devel
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
Net147 has quit []
heat has joined #dri-devel
Net147 has joined #dri-devel
<tanbulian>
I am a wreck so to speak, i risked too much, physical injury normally reflects to mental illness too, but not in terms of schizophrenia, the last which is some x unknown condition or root of caused unknown stuff to me in genetical form, such thing i do not have, injury causes depression though it's metaphysical condition or side in physical anatomy transmitters get pain impulses to cause
<tanbulian>
certain substances to generate.
melnary has quit [Ping timeout: 480 seconds]
rasterman has quit [Quit: Gettin' stinky!]
sima has quit [Ping timeout: 480 seconds]
Net147 has quit [Quit: Quit]
<tanbulian>
I am financially broke this year, but it's unlikely the case as such in the next, as i have gotten stable in mind, but it's pain to enter the duty and financial world still to compete without discount, but it's ruthless everywhere, first steps to higher businesses and all the work and coding is nightmare after a long off or dormant being.
Net147 has joined #dri-devel
kzd has joined #dri-devel
<tanbulian>
What kind of industry I enter? Though sky is a limit with such tech, realistic is not between fish and meat, realistic is to stay as codemonkey for years or nothing would work out if i rambled into wider eras all the time.
<tanbulian>
The critics at me has been overwhelming more like outrageous and injustified, so overall it's hard to get it going and exist when yet no results to show, i have all the calculator work done and plenty of immediate achivements and am almost certain, but everything needs more code code code and only code and then maintanence which is not matching up with my poor health already but the health
<tanbulian>
is still better than for Stephen Hawking who was considerably happier, due to his expectations being low all the time.
<tanbulian>
So the Linux world as of this year which changes into new one after few days. The system on all vendors drivers especially intel is very stable here is not yet designed for performance that hardware does not mandate deep in silicon, which can be changed but is utterly hard issue in all it's detail and communications to other programmers path.
<tanbulian>
What makes it so hard, is to go into past, i have been dealing with it more than 5years now with this goal, and could say for a loner lot easier ways probably to earn pennies, if you deal with something for a decade and have no results to show no money earned either, it can be classified as mission of patriotism.
davispuh has joined #dri-devel
<tanbulian>
But whenever i get something done the theory is presented in no harm ways, and implementation is enough complex to get anywhere so compared to self-timing which i would not accept in upstream, that kind of work once done could be accepted, self-timing has enourmous safety blocker.
Company has joined #dri-devel
<tanbulian>
The major thing i did not expect was virtually unlimited memory when i started my journey, tough to come up with the number on new addressing but same tough is to run out of memory in modern filesystems that are yet to be done in theory. So calculations show tremendous amount of memory per device tunable from sw.
<tanbulian>
And throughout most of my 35 year computer usage , i never until the very last years believed that could be possible where as i always belived things can be done fast.
<tanbulian>
terabyte multiplied by terabyte is such storage that is a number which i do not know either. but it is somewhere times ten in power of more than 20
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
edolnx has joined #dri-devel
Duke`` has joined #dri-devel
edolnx_ has quit [Ping timeout: 480 seconds]
kts_ has joined #dri-devel
kts has quit [Ping timeout: 480 seconds]
kts_ has quit [Ping timeout: 480 seconds]
fab has joined #dri-devel
jannh has joined #dri-devel
melnary has joined #dri-devel
dviola has joined #dri-devel
iive has joined #dri-devel
melnary has quit [Remote host closed the connection]
melnary has joined #dri-devel
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
tanbulian has quit [Remote host closed the connection]
tanbulian has joined #dri-devel
epoch101 has joined #dri-devel
tanbulian has quit [Remote host closed the connection]
hawer_ has quit [Read error: Connection reset by peer]
flowerworks has joined #dri-devel
flowerworks has quit [Remote host closed the connection]
epoch101 has quit []
sanepromer has joined #dri-devel
<sanepromer>
on 64bit that is just quite single round packing/compression that is tera*tera, you can synthesize data in several rounds of packing too, technically there is no upper limit.
<sanepromer>
So those hyper hypothetic so far but prolly possible systems i described , cause people are not optimistic and are cranky and sceptical and not impressed by the work of computer engineers, or they just get lost to the sea of obscure concepts or wrong concepts.
parthiban has quit [Quit: Leaving]
<sanepromer>
psychotic periods i have not faced in a decade, it's just that the design i do requires thinking and testing.
<sanepromer>
you are thinking that i undergo some periods of psychosis and mania consistently, but it isn't so
<sanepromer>
that my life was in danger did not cause me any additional manic episodes, in jurisdiction i do not know what hypersonic weapons are paragraphed under is it equal to cold or hard weapons firearms , but i have experience with them, they beamed the bacteria cultures in me. And that my so called girfriend was cheating on me in public everywhere was nasty i saw no additional delusions there.
<sanepromer>
for very long time before i quit dealing with her, she just gets very raw at psychological world goes straight to the most painful area and enjoys hurting others.
riteo_ has quit [Ping timeout: 480 seconds]
<sanepromer>
in other words if you carry Kali linux based modded ultrasonic or hypersonic weapon or any other type of dangerous out of spec beamer with you according to law, law enforcement agent such as police may and very very likely will sanction you in some way cause you are potentional terrorist there and you have to explain why you need such equipment and who are you.
crabbedhaloablut has quit []
xmas has joined #dri-devel
crabbedhaloablut has joined #dri-devel
<sanepromer>
let alone that such beamers should be made easy to access from programming standpoint, it's safety issue and i was wrong to bring it up, hardware can do also millions of fancy things what i mean, so but such systems are monitored and controlled over/after.
<sanepromer>
otherwise technically we could release it as there is no technical issue , but that gets the user subject into problems with authorities.
<sanepromer>
self-timing as in baseband chipset such as wifi,mobile broad,bluetooth is in hw, can be pulled to rest of the interconnect in sw, but that is never gonna happen, in theory yes if we were good people not to threatten each other and stick to our lives without ruining others , radiators and coolers are never needed and the implementation is easier than my compiler ideas.
<sanepromer>
there were some failures in finishing me, but for some time aggression has not been now showed after cubious amounts of failures, do not question my judgement on those, the neck attack which is not the last assault, got me totally knocked down and the nights memory was not there anymore from certain point, earlier memory was there, what happened is not known to me, someone said something
<sanepromer>
but pictures and recordings have not been analysed, if it was not from behind the back carried out assault, then i would have had the memory as it never fails without a reason or force before the event.
<sanepromer>
allegedly this was another english maniac among many there assaulting and humiliating me, talking about some joss, so it routes to this channels morons.
<sanepromer>
his soccer nor his public toilet in phom penh or islands or his downy face i am not interested to see.
funderscore is now known as f_
tobiasjakobi has joined #dri-devel
tobiasjakobi has quit []
<sanepromer>
neck has some bump and has obvious movement restrictions as of now, but as to what happened can be so many possible things that i do not speculate over them, some people talked to me, about amniotic fluid donors through stem forwarding, some about neuralink, those are the most scifi-est possibilities.
<sanepromer>
what i know is that i got several warnings and indications to this by total criminals that someone is after my head, and we declare a war to them and kill them off.
sanepromer has quit [Ping timeout: 480 seconds]
sanepromer has joined #dri-devel
riteo has joined #dri-devel
nerdopolis has joined #dri-devel
<sanepromer>
I did the encoded 144 to 72 transition in the calculator in small sum like for sequence of elements, and it appears to work, i have done all the numbers testing and everithing is fine there, but surely i need another year to come out with first architecture on sw compilers. But there is no rush, my computers work satisfyingly.
<sanepromer>
i had small issues in the past in memory leaks for amd hw and intel had them, but as of 2024 on a snb, well all that has changed, this laptop i type i exchangd a battery and is running for a month in a row, now this has been brilliant from graphics vision experts, we can not forget that to see those graphics they need to trim to human sensible wave frequencies.
<sanepromer>
and it looks like the official emplyees knew that and stability matters pretty well.
<sanepromer>
it's the same laptop that linus shouted at because it under debian did not install for years nor ubuntu or any known linux distros, until something got fixed likely in the kernel drivers.
<sanepromer>
but also various same models today have landed coreboot support for this, it was hacked on, due to it's classy and user loved form factor and user experience.
rasterman has joined #dri-devel
<sanepromer>
i am broke at the moment, but get contnuous salaries, however i have enough of functional hardware where as complete laptop park including one ice lake low power version too, is also functional in the next year soon, i collected a lot of hw.
<sanepromer>
i do not feel like i ever was a completely mentally ill person such as an idiot, i think that is your take on things, due to my behaviour to disturb you, actually things do not look as bad.
<sanepromer>
i consider the work done so far being just a base , where rich possibilities are though offered with all hw supported by drivers and have understanable APIs to write new ones, more importantly picture is well offered and functionality for free.
<sanepromer>
however i do not bet on the base systems my own, even though to get there a lot of work was needed to be done, i bet on more modern ways which plumbs into that base.
coldfeet has quit [Remote host closed the connection]
<sanepromer>
my grandfather always told, there is so thin line that seperates an idiot from genius, and i risk heavy for not caring what others think, aries stubborn person i am, the line is really very thin, since we can not tell which end it goes, the cited or talked TempleOS author had some ideas too that were not so bad, but he indeed went to cryptically offshore with his take on being godly
<sanepromer>
engineer.
<sanepromer>
to look at this schematic what i plan on encoding the powers with smaller weights, in the beginning it seems like 50 50 chanche for longer period, i could not tell my own if that is going to work, after dealing it for years in a row, i could tell that is worth a shot, and likely works as the odds moved to 10 90 or so for possible success, and travises talks also did not give away so much
<sanepromer>
technical retarding, i could not tell if he was ill or not at all, so my grandfathers sentence kind a holds, there is very thin line between a genius and an idiot.
<sanepromer>
grandfather passed away couple years ago, and i have been living with my last grandparent , and have tested quite a lot here in big silence, as it is still close to town but a rural district though shop is close, but very nicely silent to conduct many researches at.
<sanepromer>
well the figure i am in the world is notorious where as i do not feel like deserving this, the false imagery has been painted by et on nokuke nii pikuke sull scammers in their industry of intrugues in club HALL where they bone eachother after plumbing in all the phone cameras at the entrance, and play wank tunes about the person they never knew , and hence scam a violative false image at
<sanepromer>
me, the abuse is disturbing and hippocratic betrayal felt though i do not know most of the people behind that absurd behaviour though.
<sanepromer>
the whole story i won't ever cover in that world, but it is very cruel and entirely delusionally absurd.
sanepromer has quit [Remote host closed the connection]
bcheng has quit [Remote host closed the connection]
bcheng has joined #dri-devel
pcercuei has joined #dri-devel
rasterman has quit [Quit: Gettin' stinky!]
coldfeet has joined #dri-devel
fab has quit [Ping timeout: 480 seconds]
swissboyant has joined #dri-devel
<swissboyant>
What i have is so narrow but still very wide area where i expertise and i chose to be good in there. So that is why i engineered nothing from the base which was crucial, other people did and are in charge of those codepaths. I chose to deal there where's need to really think very deep and careful, but it's not more important than other work.
sima has joined #dri-devel
coldfeet has quit [Quit: Lost terminal]
<Ermine>
Are there GPUs that can't even display anything before their firmware is loaded?
<DemiMarie>
Ermine: On x86 or elsewhere?
swissboyant has quit [Read error: Connection reset by peer]
<Ermine>
DemiMarie: wherever
swissboyant has joined #dri-devel
swissboyant has quit [Remote host closed the connection]
<DemiMarie>
Ermine: On PC, VGA output in early boot is expected to work so the GPU firmware would need to be loaded by CPU firmware in this case. Otherwise, I have no idea.
Duke`` has quit [Ping timeout: 480 seconds]
sima has quit [Ping timeout: 480 seconds]
yogesh_mohan has joined #dri-devel
yogesh_m1 has quit [Ping timeout: 480 seconds]
pcercuei has quit [Quit: leaving]
<mareko>
firmware is an optional component of hw and can be fully replaced by fixed logic
glennk has quit [Read error: Connection reset by peer]
glennk has joined #dri-devel
Mangix has quit [Read error: Connection reset by peer]