14:02 mardination: what i think, if anyone had some doubts than still the theoretical base win on some the technical dilemmas goes to me, however i am not made of steel, i also do mistakes, but i commented some of it short, from philosophy to implementation details is rough go
14:03 mardination: if someone wants to talk about implementation details, we can do that, but this requires you to do what i have wished
14:04 mardination: i.e put the nose into miaow code and if there are questions and i can answer them, i may comment or we may comment together.
14:07 mardination: having doubts on people skillset or qualification or doubts of code that it is not as good as philsophy learned from papers, that it works so...round about, but some paranoid people do not belive that hw designers are qualified not to make bugs
14:08 mardination: this all can be cured with taking the full pot on, and going throough it line by line, which i also did , i am also paranoid enough and have needs allthough not doubts
14:12 mardination: in general I did not belive designers who are paid 5k or more per month to make hardware design, that they are underqualified, in general i round about belived that all works , time to time you get this question that how, and than exactly how
14:17 mardination: i have not yet read all the TCE TTA vliws, but miaow is very good in how it does things, it is the best that i have seen so far -- i think tce tta are good as well, and synposis i have not looked, it is commercial, maybe they used cadence/synopsis generator there, my understandings are i did not break any law like always
14:17 mardination: the code was published with MIT license
14:26 mardination: it is only important to me, that i have managed to straighten up, break the barriers, and show that i understand code very well, where others that state i am hallucational get bitten hard
14:27 mardination: it is important for me to lift that trophy based of my needs to balance that situation, cause i have been terrorised and conspired against for decades now, actually those details should not be that important for others
14:34 mardination: for others it is more like today a combination that was known and comes relatively relaxed and easily, cause they did not face any bad momentum, and were croddled all the way through on ideal line
14:35 mardination: more like tuesday after coffee acheivement, but for me it is an acheivement
14:40 mardination: and for those doctor wannabes i say you are worthless criminals terrorists, whos acheivements are regarded as as very shithosotic behavior in life, fecalists during their best days of violative behavior, those can do curly poop and even triangular shit into the pipes only make smelly fart for personalities to tolerate
14:45 mardination: and to i do not want to even comment this any further, vmt kinda guys are probably rate my poo heros, who exhibit appearingly very shithosotic behavior, in response to his given diagnosis to me, that i am a psyhcotic scizophrenic
14:45 mardination: we are talking about a major clueless shithose there who is scamming terrorist like most i face in life, and that is kinda it.
14:50 mardination: and i am not even sure whome to pick, a major fraud conspiring scammer like this, or a thieve who can be reponsible after heist or snatch gone soar, caught on the act, when having planned a heist to steal scammers money.
14:51 mardination: and in general i do not like thieves at all, but this guy is a lot worse
14:56 mardination: there is no risk of any kind being a conspirant and scammer, only thing to get from such shithoses is confession that i did a mistake, without the confession that it was all intentionally done
14:57 mardination: like after a twenty years of wasted time due to this terror, one idiot comes up with i accidentally did a mistake
14:59 mardination: cause i was a lost talent a bit my own, i also looked back to history what people before me had faced in relations with foreigners, and every sportsmen before me, the biggest guns of estonia said they are violators cheaters and fraud, and using unallowed tactics
15:01 mardination: i kinda wondered, why i was erased so quickly before turning to be 18 and doing mens class sports and work
15:01 mardination: cause shithoses just violate, cause the genetics show off and are paranoid that someone can be of good
15:02 mardination: i can not hide the genetics even if i underperform, it just energetics that can not be lost
15:02 mardination: or hidden
15:06 mardination: because i was a child during those days i got injured with conspiration, that is a biggest shithoses work and is considered entirely unethical and amoral
15:11 mardination: you need to be responsible of you words, to gain estonian men or mans respect, and this shithosotic behavior that i was wrong is not enough, if someone gets killed cause of your shit, and you do not confess the firsst degree murder arrangement than imo according to your behaviro this should be forced on
15:11 mardination: extorted uunder pressure sucked out info
15:13 mardination: as a group of conspirants intentionally on a road of first degree murder without a reason should be sent behind bars cleverly
15:15 mardination: and there is a proper code for you from the real thieves, who hate such scammers, and this time no pipes will be filled your shit, but a dick is in your ass all and all the time, as conspirants who conspire against children are treated alike there
15:28 mardination: criminal investigators are very interested how sorry i did a mistake happened against the same individual more than hundred times, wwitthout hard evidence whatsoever me being a danger as a kid to anyone, or hard evidence about the presence of scizchophrenia
16:10 mardination: generally or generically summing this up, is i picked a central most important file, it is all subjective, it is fetch/round_robin.v this file feeds f_decode_wfids to the chip, it also centrally fetches opcodes from pcblocks
16:12 mardination: it based of the vacant wfid entries, which is...vacant_wr_reg=instr_done_en | wr
16:13 mardination: that means if the dispatcher does not register a dispatch , cause wr=wf_dispatch_i
16:13 mardination: the whole chip can not feed valid bits to the decoder and issuee module via wavepool
16:14 mardination: so in other words, decode_wr_data is in two states the wr port for the read muxes, in instr_info_table.v
16:14 mardination: depending if the chip fed some valid entries f_decode_valid or not
16:15 mardination: if dispatcher is clamped to fail , no valid entries will be fed, and chip takes previous issued_rd_data from the flops
16:16 mardination: however when valid entries are feed, it takes the current fetched and decoded instruction using the full pipeline length, where the earlier can be managed with lsu
16:16 mardination: now, the detail that i allready also mentioned before was.
16:17 mardination: if the stream meets a dependency
16:17 mardination: meaning that for whatever reason instruction can not be issued
16:17 mardination: than the instructions decode instance will be leaving valid bits on
16:19 mardination: during such a modent, it feeds the alus directly from the decode module issue_opcode, without even going through the
16:19 mardination: *moment, going through the issue flops, but only scoreboard flops
16:20 mardination: in other words, issue_opcode and it's brothers have two drivers
16:20 mardination: but one of them is not driven conditionally
16:20 mardination: this in verilog means a tri-state buffer technique
16:21 mardination: to avoid transistor damage whatever because of multiple drivers for the same wire/net or port
16:27 mardination: anyways this method is very clever, and definitely one good and right way to do it
16:36 mardination: why it carries some importance, because it is the fast fallback path for one method, to get additional slots
16:37 mardination: cause now fetch queue content passes through the issue module basically, so it is very fast
16:38 mardination: theoretically there are two methods, one is pinning the opcodes to different flops and lsuing there by parsing the opcode from registers and data cache
16:45 mardination: I implement all that code if needed, since I have talked about this several time, and for some reason you yet have not understood
16:54 mardination: for me some of the issues are with remembering the tessellation and geometry shader pipeline at the moment
16:55 mardination: shader stages only, but this method can be also done in generic ways which is suitable for every shading stage
17:00 mardination: very long book can be cooked together based of miaow, to help people to understand the code, but heck i do not have resources or momentum to fill this out
17:00 mardination: too many issues i should deal with urgently
17:01 mardination: i had c++ book when i started excersizing
17:02 mardination: it was scanned in and i still have it, public library in tallinn had it
17:02 mardination: turbo c++, it was one of the fewest programming books, and i generally say noone was interested in this, except me
17:02 mardination: and my uncle
17:03 mardination: the author did not have psychotic look or scizho look, but was a serious programming guy
17:04 mardination: you could tell from his eye stare that this one had worked with code
18:16 mardination: i have some more tests to do, testbenches, to yet reconfirm things, I'm not always entirely sure what i read :) lucky guesses
18:19 mardination: there is some percentage of accuracy in my views, not always entirely sure, even though i think, if we were to discuss about miaow, i do not promise anything to be accurate , books would also probably neeed to undergo some reviews and mistakes fixed etc.
18:39 mardination: i am pretty sure i know how that hw works but if i explain it right is another matter.