10:24 pastedtext1: It's very professional what this code does by the looks, Armin Biere seems austrian, but the author of gini is this man https://github.com/scott-cotton/
10:25 pastedtext1: He is likely some british or american
11:22 meteorites2: I will roll my own solution too, but they have nearly the same logic in fact already all implemented. All one has to do is implement the compression routines and and inverter graph in packed format. Gini loops are good, it has good allocation strategy, can squeeze to machine words can arbitrate from buffers, and takes care of the branches, so this looks like it generates dimacs then aiger, and even this it does very well, but one needs to add
11:22 meteorites2: compression like outlined by me, and you need to find a hook where to add decompressed writes, that is only sat solver it does not execute, it verifies but it has those routines to ask for the value, little bit needs to be written as of yet, but this is fun, it really is fun stuff.
13:26 onlymarshal: No, i am not mentally Ill from birth, i was trained for performance, not for fraud doctors or sexual intrigues or terror, which means after getting injured i had no era nor job, cause i do not respect terrorists , computer technologies is my hobby and that is why i still live, however i do not know for how long, since i have had incredible amount of luck many times to escape death that they arranged. Hence yes, i put my findings
13:26 onlymarshal: out for free.
14:00 onlymarshal: but yes, i was in many files in the solver, and it is functionally equivalent to my ideas, but they had it for more than ten years ago already, you can look at logic/roll_test.go and logic/roll.go and they claim that variable mapping is generic for the solvers, in other words the unroller is just one way to write as an example, you can lift it to the dimacs as well, dmap[v], res append puts it into one variable in roll.go cause it
14:00 onlymarshal: appends and writes res out inside the examples file loop
14:01 onlymarshal: the thing is my first idea came 5years back, they had it a lot earlier
14:18 onlymarshal: the calculation is two sided, there is math side, which is basing on roman number theory, but they do the exact same thing in computer boolean logic
14:21 onlymarshal: i can write it in roman number theory based too and rewrite the code and in any ways possible, but it should end up as very performant in execution of whatever code
14:31 onlymarshal: I think you want to say it is technically better than mine, which is actually yeah true, cause it makes the masks in lower granularity too, at a boolean level
14:34 onlymarshal: so you win, but whatever, i agree, yes, maybe the compressing and decompress code additions is not relevant here, it already implements everything needed, but the unroller logic for other bit lengths then has to be manually adjusted but only the bounds
14:42 onlymarshal: not only the bounds but based of that file, but you can see as to what they do, this solver has all the docs too available, so a counter of 16bit is used there, which is a latch, to eliminate this as sequential irrelevant code it is removed from the buffers and it seeks forward, so you have a 7bit not logic to remove that, you see 16bit and 7 bit+carry, its the same as 512 from 1024 max
14:42 onlymarshal: but their version is in boolean logic
14:44 onlymarshal: but i have to go to catch a bus, to go to shop buying food for my granny and me
15:45 Caterpillar: amd radeon drivers quality is getting worse and worse
15:45 Caterpillar: damn
15:46 Caterpillar: 100% times my X13 gen3 laptop crashes is because of the GPU drivers
15:47 Caterpillar: https://gitlab.freedesktop.org/users/germano/activity
16:07 f_: Caterpillar: Might want to properly state your issue with as much info as possible instead of complaining.
16:16 Caterpillar: f_: if you bothered to open the URL you could have seen the amount of times I provide infos
16:17 f_: :c
19:16 mindyourthings: I mean yeah all the code in the networks is bullshit ideas, only real thing is to write things your own, but i am no longer present on channels, since i described how real code would work, i am tired of seeing no real solutions, i have my own in months, and no longer continue to write on irc neither do any browsing, i did describe what i am about to code, so i was still fair. Just i do not have time left anymore. Cheers.