log☇︎
52700+ entries in 0.031s
mircea_popescu: IF 4 takes too little to bother, 2-3 aren't worth doing ; otherwise, otherwise.
asciilifeform: steps 3, 5, 6 take ~0. step 4 takes a certain yet-unknown time, which is a function of width(reg) and shots.
asciilifeform: step 2 takes ~1msec (on asciilifeform's aged box, and less elsewhere.)
asciilifeform: nao, on a machine with 1 FG, step 1 takes ~36msec. every time.
asciilifeform: this is the 'pre-sieve' variant contemplated.
asciilifeform: let's spell out the algo. 1. reg := read_from_fg(2048bits). 2. l = gcd(reg, primorial) 3. if l != 1, goto 1 4. mr = m-r(reg, shots) 5. if mr != true, goto 1. 6. reg is probabilistically-prime .
asciilifeform: then you only m-r on inputs that already passed gcd
asciilifeform: ~doing~ the gcd pre-sieve
mircea_popescu: what does ? not bothering to gcd ?
asciilifeform: it potentially frees up cpu for moar shots of m-r, was the idea
mircea_popescu: gcding small primes doesn't reduce the m-r constant.
asciilifeform: how 'many r-m shots' is a free variable tho.
mircea_popescu: why overload the machinery with unreliable testing.
mircea_popescu: asciilifeform IF rm doesn't take much time, THEN i don't see the point of doing any gcd. because r-m is a complete test for what we need, and gcd is not.
asciilifeform: i'ma come back to this binomial once i have the m-r.
asciilifeform: grr, moar complicated than i initially pictured, because m-r ~also~ demands rng
asciilifeform: so depending on how many m-r shots you want, it can make sense to pre-sieve, to give m-r moar shots.
asciilifeform: the interesting thing about m-r however, is that it can make use of any available cpu given to it, to produce smaller probability of death
asciilifeform: i dun have the m-r yet, so cannot give the equation quite yet.
asciilifeform: ( and that on iron where m-r ends up taking substantially less time than FG takes to fill up candidate reg, there may not be a point to pre-sieving )
asciilifeform: my observation was that on a box with 1 FG, the latter will almost certainly be the limiting reactant in prime-baking.
asciilifeform: i.e. to save on r-m time.
asciilifeform: of what ? that's the obv. version ( get from rng, until gcd shows that no small factors, and then m-r )
mircea_popescu: i don't see the point.
asciilifeform: i'd be satisfied with gcd as initial sieve (given that i have one that dun leak anyffin) ~followed~ by m-r.
mircea_popescu: something like that.
asciilifeform: well at least not in the initial sieve ( one might still want it for e.g. pocklington's test )
mircea_popescu: if it doesn't save any time it's not worth having
mircea_popescu: the whole idea of prior gcd was to save on rm time.
asciilifeform: http://p.bvulpes.com/pastes/TkmoM/?raw=true << sneak peek of constant-time stein. ( afaik there isn't one anywhere else on the net, funnily enuff. )
asciilifeform: ( afaik there are no integers that are divisible by small prime (such as will fit in the primorial) and fail ~any~ number of m-r shots.
asciilifeform: cuz in the former case, there aint actually any point to gcd.
asciilifeform: nao, philosophical q : does one actually want to gcd + m-r always ? or is it acceptable to reject input after failed gcd litmus, and only ~then~ m-r .
asciilifeform: ( all of this assumes that nothing is parallelized. asciilifeform in particular does not like parallelized subcomponents in rsatron, if it can be avoided , tho there aint anyffin wrong with running ~multiple~ rsatrons , on diff inputs, in parallel , if iron is available )
asciilifeform: on a machine with multiple FG harnessed together, divide the figure by the # of FG in use.
asciilifeform: if primality test ( which consists of GCD ~and~ m-r, in order to constant-time ) does not exceed 0.0356sec, then on machine with 1 FG it can be considered that the FG is the limiting reactant.
asciilifeform: let's say yer baking one of the p, q of a 4096bit rsa mod. it needs 2048bit , i.e. 256byte of FG. a standard FG at room temp shits out 7kByte/s. therefore 256 / (7 x 1024) ~= 0.0357 sec., for a fillup of candidate register.
asciilifeform: on a box with 1 FG, the wait for a random fillup of a e.g. 2048bit reg, i suspect dwarfs the runtime of stein (and possibly even of m-r, dunno yet)
asciilifeform: the catch is that ~all != all.
asciilifeform: ( afaik nobody ever proved that you couldn't, and the nonconstanttime lehmer and variations run in ~O(n) for ~all inputs )
asciilifeform: incidentally, whether it is possible to gcd arbitrary ints in non-quadratic time, is an open problem in numtheory
asciilifeform: next can complain that newton didn't put out 7bit ascii.
mircea_popescu: cuz erryone gotta be spashul, come up with diagrams. text not good anymore etc.
asciilifeform: it aint edible this way.
asciilifeform: mircea_popescu: lol, ~whole thing is diagrams
mircea_popescu: let's then put the more useful part in teh archive hopper : https://archive.org/stream/bitsavers_burroughsBkComputerSystemOrganizationTheB5700B6700_10821314/Organick_Computer_System_Organization_The_B5700_B6700_Series_1973_djvu.txt
feedbot: http://bimbo.club/2019/01/philosophical-transactions-for-the-months-of-january-february-and-march-1714-part-iii/ << Bimbo.Club -- Philosophical Transactions. For the months of January, February and March, 1714 - Part III.
asciilifeform: ( and it aint a lie, i studied the burroughs arch, indeed was non-overflowistic iron )
asciilifeform: ( nfi what happened to it after '15 , but sorta terrydavis-flavour )
asciilifeform: meanwhile, in the 'man alone' dept, http://interim.mntmn.com
asciilifeform: and hey, iirc google (or was it crapple?) occupies the old sgi campus
mircea_popescu: in other sad lulz : etsy moved into the new york building where a scottish immigrant invented the prefab packaging box during the gilded age. who knows, maybe it rubs off ? maybe perhaps http://trilema.com/2016/the-%d0%ba%d0%b2%d0%b0%d1%81-bdsm-party/#comment-118604 ?
a111: Logged on 2019-01-08 17:26 mircea_popescu: every night spent in the cold to "save fuel" by whole neighbourhoods, the inaccessible meat or bananas or whatever -- ALL OF IT!!! went into one "humongous" pile of "the people's wealth" that was, looking in own pockets, not retrospectively that worth the mention ?
asciilifeform: http://btcbase.org/log/2019-01-08#1885903 << 'pile' wento to washington/eurpoistan coffers, kept the reich going 30yrs past its expiration date ☝︎
asciilifeform: 'There were people working behind the scenes trying to get the Symbolics VLM stuff legitimately released in some way to the hobbyist community - and this is likely to throw a wrench into that effort'
asciilifeform: it's exactly scale-model ver of the ru 'privatization' afaik.
mircea_popescu: what the fuck more is there, if this dun suffice re vanitas vanitatum what does.
mircea_popescu: every night spent in the cold to "save fuel" by whole neighbourhoods, the inaccessible meat or bananas or whatever -- ALL OF IT!!! went into one "humongous" pile of "the people's wealth" that was, looking in own pockets, not retrospectively that worth the mention ? ☟︎
mircea_popescu: it's one hell of an object to contemplate, this evolution, for someone who was there in the 80s. "you mean every penny saved off every old schmuck that died in soviet-style economical health care system / every bled out 1950s style (in the us, heh, http://trilema.com/2009/comunismul-capitalismul-si-oprimarea-femeii/#selection-71.0-71.8 ) woman because contraceptives cost money whereas her slavespawn is worth money, ALL THAT,
mircea_popescu: now it's like... hey, remember when bitcoin had JUST breached 1k and "mp is a billionaire" still "controversial" / difficult to swallow for the tards talking ?
mircea_popescu: anyway, twas a mere coupla billion. which, at the time, was an IMMENSE sum, that ceausescu had gathered together over decades of hard fucking work for his 20mn slavecrop.
mircea_popescu: that was something else, that january.
mircea_popescu: just about yeah. dood ran off with the commies accounts receivable in 1989, so that isarescu was stuck inventing acreditives for fuel ships to unload rather than leave that winter.
asciilifeform: mircea_popescu: was that the ro berezovsky or who
asciilifeform: ( observe, goering -- morphinist from ww1 to '45, and never od )
asciilifeform: it's how they od, also
mircea_popescu: (that article, btw, remarkable because i told ro bezos "Si, draga Varanule, nu-i clar ca mai ai cinci ani" ie, "dear reptile, it's not clear you have five years left" in 2012 -- by 2014 he was in jail. because hey, mp is right now and again accidentally which is to say 1024/1024 occurences on his blog etc etc) ☟︎
mircea_popescu: the unsteady approach being in itself a great catalyst of both the paroxistic problem encountered and the incapacity for management displayed.
mircea_popescu: yeah, they ~party~ ie, don't keep a ready supply of the stuff (chemicals, biologicals) for whenever feel like, but rather whenever the "pipe" leaks something run out and buy some of each, twice a eyar or w/e.
BingoBoingo: Anyways, it seems the reason he didn't pile up 2 dozen stiffs over the years is that same poverty
mircea_popescu: somehow this doesn't bother them, "hey, our entire understanding of money is, let's get some schmuck a bunch of taxchest so he can http://trilema.com/2012/dan-voiculescu-acest-gunoi-abject/ all over."
BingoBoingo: Except cut a couple figures off for the US version
mircea_popescu: precisely the "russian oligarchs" pseudo-economic model, yes.
asciilifeform: mircea_popescu: pretty sure most of their dough comes from bezos
mircea_popescu: asciilifeform my point is rather that the democrats ~are very fucking poor~.
BingoBoingo: Other sources report ~500k still small. More interesting is the information brokerage he sold in his 30's.
asciilifeform: mircea_popescu: they have this custom, iirc you can donate 1 orcbuck and still show up on 'list of donor'
mircea_popescu: since when the fuck is 30k a sum of money.
mircea_popescu: He formerly served on the steering committee of the Stonewall Democratic Club. He has donated "nearly $30,000" to Democratic candidates like Hillary Clinton, Ted Lieu, Pete Aguilar and Raja Krishnamoorthi as well as the Getting Things Done PAC. << dude give me a fucking break ?
asciilifeform: it's a lul tho, the incas actually think they have 'literacy'
mircea_popescu: the odds a 20something ~black~ ~male~ actually keeping a journal are 10ppm or so ; having the capacity to even in principle do so 0.1% sorta deal
mircea_popescu: asciilifeform you may think what you wish ; i've seen the fetards and i know better.
BingoBoingo: Anyways, first incident was dismissed after pantsuit outrage over the investigation as an attack on the old dude's lifestyle
mircea_popescu: oh, you mean the ipad ? i think they cut off what he was trying to say.
mircea_popescu: BingoBoingo the ridiculous notion " Gemmel Moore, a 26 year old male twink" was ~keepin a journal~. srsly ? with what ?
mircea_popescu: in short : it's a wonder there weren't two dozen.
mircea_popescu: that one in x thousand black gay-for-pay tards (who are also drug users, nqa) will od a year... gimme a break.
mircea_popescu: well, in all fairness to the http://trilema.com/2016/welcome-to-baluba-island/ cuck in question -- if your wife fucks a dozen or so "bulls" before they do you in the ass per week say, and it's been a hundred weeks, you're looking at five or so thousand of these having spent a day with you within the past half week, which is the "relation" standard contemplated.
BingoBoingo: White pantsuit hiring black men for sex and hooking them on meth... like that Atlanta dude but with different taste in holes
BingoBoingo: From the text of the Qntra piece: "In the previous incident Gemmel Moore, a 26 year old male sex worker's death was ruled a drug overdose following an investigation which found numerous syringes and indicators of habitual methamphetamine use in Buck's home. According to Moore's journal, Buck introduced him to methamphetamine."
BingoBoingo: <mircea_popescu> BingoBoingo are these tynisha ysais relatives ? any bits missing ? << Apparently they are just dead sex workers
asciilifeform: but i got the idea from him
asciilifeform: i dunno how phf ended up into it tho, iirc he dun cad ☟︎
asciilifeform: the win is that when you e.g. cad, less muscle aches. ( asciilifeform aint cad'ing currently but expects to again later this yr.. )
asciilifeform: it takes ~same space as mouse, if you have a decent two-bricks-sized box
mircea_popescu: not proposing this as problems ~with the item~.
mircea_popescu: i get the "takes less space than mouse" etc. it just seems bulky and inconvenient to me. and too high. and so on.
mircea_popescu never got into that.
asciilifeform: ( 1:1 mm/rolled to pixels )
asciilifeform: well if it's a trad mouse you generally wouldn't