asciilifeform: it can go to hell.

asciilifeform: the hypothetical future 128ism is not , in that light, needed for anything.

asciilifeform: goal is ADEQUATE performance on iron of TODAY

asciilifeform: can.

asciilifeform: it is idiocy and i won't countenance it.

asciilifeform: because you lose >2x performance for no gain.

asciilifeform: just like ffa worx just fine on opteron today dialed down to 32

asciilifeform: m

asciilifeform: mno. on 128bit machine, we still can use 64bit arith,

asciilifeform: 'W is constrained, such that any permissible value of W must be representable in a whole number of machine words on 8, 16, 32, 64-bit ALU.'

asciilifeform: 'ffa represents a W-bit integer as a contiguous array of N machine words of bitness B, W = N*B.'

asciilifeform: i have nfi where to even begin

asciilifeform: you want 8x slowdown? for what?

asciilifeform: why?!

asciilifeform: nor 192bit cpu

asciilifeform: no!!

asciilifeform: understand, it is not possible to use partial machinewords in the simplest possible (i.e. the one in ffa) arithmetic method.

asciilifeform: 64 divides 128

asciilifeform: so it gotta work on 16, 32, and 64 bitness machine.

asciilifeform: which requres a 64b quantum

asciilifeform: it has to do SAME THING on all iron

asciilifeform: btw does it makes sense why i put in that req ?

asciilifeform: ( at which point you oughta use THAT as the cap, not it )

asciilifeform: unless you push it up to next multiple.

asciilifeform: and won't work in ffa.

asciilifeform: that ain't divisible by 64

asciilifeform: > 4096

asciilifeform: that's 4104 bits...

asciilifeform: lool

asciilifeform: breakable even today

asciilifeform: 1024b?!

asciilifeform: waat

asciilifeform: wai 515

asciilifeform: for given w.

asciilifeform: this is still difficult for gcd because gotta prove max number of shots needed

asciilifeform: 2) replace variable iteration params with fixed, conditional terminations with same

asciilifeform: 1) replace termination condition with a mux that starts discarding new results of iteration in favour of old, at iteration T and after

asciilifeform: the basic, naive method for magicking a conventional algo into a constanttime algo, is to

asciilifeform still devising a constant time gcd

asciilifeform: idea is, for pre-millerrabin litmus, take gcd(candidate, Qw) where Qw is largest primorial that fits in the ffawidth ☟︎

asciilifeform: approx!yes

asciilifeform: you wouldn't want a mb of anything in ffa tho

asciilifeform: ah hm will have to see re largest prime

asciilifeform: it's a prelude for avoiding expensive miller-rabin when the latter is doomed to fail

asciilifeform: you wouldn't use it alone , lol!!

asciilifeform: it's largely useless tho

asciilifeform: 1.2G at last count

asciilifeform: 'rng bits are expensive' spawned quite a few idiocies , by itself

asciilifeform: the important bit : if fails, make a ~new~ one, rather than n+1

asciilifeform: as usual asciilifeform has deeply nfi why NONE of the published rsatrons, to date, do this.

asciilifeform: ( then, elementarily, you gcd against '8ball', primorial of $largeint ) ☟︎

asciilifeform: the other thing, you don't need ANY trial-divisions in the prelude to miller-rabin, IF you have a constant-time gcd

asciilifeform: ( tldr : superiority of the FUCKGOATS-enabled approach, of get-new-N-bits-from-rng-then-primalitytest-until-done, vs the kochian get-N-bits-then-increment-until-passes-millerrabin ) ☟︎☟︎

asciilifeform: in other olds , https://www.iacr.org/archive/ches2009/57470141/57470141.pdf << traditional rsa prime generation is quite 'loud'. this is not a seekrit. subj demonstrates algo for actually recovering the prime.

asciilifeform: ^ ar propaganda cartoon re 'malvinas', with 'fixed' sub

asciilifeform: in other hilarities almost worthy of mircea_popescu's robotzi, https://www.youtube.com/watch?v=VKkcTpCur7g ☟︎

asciilifeform: 'the police abandoned us!' etc

asciilifeform: in other lulz, https://www.youtube.com/watch?v=IzZb6kfctEU << 'my father would spin like a dreidel in his grave if he saw' ☟︎☟︎

asciilifeform: heiya mod6

asciilifeform: ( it's in the gcc lang set, right next to f77 )

asciilifeform: in all likelihood you already have a copy.

asciilifeform: standard gnat worx

asciilifeform: nah

asciilifeform: and 'to ada' also wrong word, (ultra fascist subset of) ada is simply what remained on the table after we went methodically through all reqs for a sane programming (for existing iron) system

asciilifeform: port may be wrong word

asciilifeform: also interesting to mike_c will be the 'trbi' threads..

asciilifeform: mike_c: i also have adatronic bitcoin formats ( block, tx, message, indices ) encoder/decoder; and mmaptronic db . but this is not yet primetime product

asciilifeform: this is a somewhat obsolete version but illustrative

asciilifeform: ( http://btcbase.org/log/2017-07-16#1685034 moar thread ) ☝︎

asciilifeform: http://btcbase.org/log/2017-07-15#1684808 ☝︎

asciilifeform: comes with makefile etc

asciilifeform: mod6: though there is the factorial demo

asciilifeform: mike_c: not as such. no vrelease yet - pastes in logs

asciilifeform: about 1kloc

asciilifeform: mike_c: currently we have ~90% of a nonbranching ( no timing leak ) , provably correct arithmetron

asciilifeform: *dijkstraist

asciilifeform: i've been pushing folx to return to a 'dijkstaraist' concept of program, as something primarily to be ~read~, and only secondarily as something to execute on comp

asciilifeform: so in the end it will not be strictly necessary to eat every gram of log that added up to it, to properly grasp

asciilifeform: the saving grace is that almost whole point of p is for it to be readable/self-contained

asciilifeform: lol

asciilifeform: 'p' is very different flavour from the old heathenry , however. think 'forth with bignums' approx.

asciilifeform: http://btcbase.org/log/2017-08-13#1697465 << correct. is how i ended up working on a replacement for pgp ☝︎

asciilifeform: ty mike_c

asciilifeform: back to the living!

asciilifeform: wb mike_c !!

asciilifeform: woah

asciilifeform: meanwhile, in the circus, https://archive.is/t14Cy << 'extremism detector ai' ☟︎

asciilifeform: phf: all of open sores is crawling with cockroaches, i'd be quite surprised if sbcl turned out to be inexplicably clean even now

asciilifeform: http://btcbase.org/log/2017-08-13#1697308 << simple pill : run no sbcl made post-2011... ☝︎

asciilifeform: http://btcbase.org/log/2017-08-13#1697295 << our top quisling was also 'physicist' -- saharov ☝︎

asciilifeform: http://btcbase.org/log/2017-08-13#1697299 << if i were keeping seekrits for a mega-quisling, i at least wouldn't advertise, lol ☝︎

asciilifeform: http://btcbase.org/log/2017-08-13#1697290 << interesting... friends... phf has ☝︎

asciilifeform: mircea_popescu: police choppers in particular -- not merely 'no armour', but thinner skin than toyota

asciilifeform: lulzily, it's a treatise on prng and hashing

asciilifeform: 'Adding to the turmoil, the Federal Aviation Administration said late Saturday that a Virginia State Police helicopter had crashed about seven miles southwest of Charlottesville. State Police officials said two people died in the crash, the cause of which was not known.'

asciilifeform: where d00d's personal collection was.

asciilifeform walked into b00ksh0p being sold off 'by the pound' after death of owner, found there a b00k by luby (of luby code) that had nfi existed at all

asciilifeform: ( i.e. undergrad 1st term homework )

asciilifeform: https://curl.haxx.se/changes.html#7_55_0 << holy SHIT that's a lengthy buglist for a proggy that... loads http and puts to stdout ☟︎

asciilifeform: ( not 1st time, either, see http://btcbase.org/log/2016-02-17#1408732 ) ☝︎