asciilifeform: what's it do that gnumake, cmake, etc dun do ?

asciilifeform: ugh

asciilifeform: diana_coman: what is 'jam' ?

asciilifeform: wb ave1

asciilifeform actually considered a fortran ffa, in '16

asciilifeform: ( the fortran - f77 )

asciilifeform: it , if one is to believe the docs, implements the '83 standard.

asciilifeform: phunphakt: in ye olde bolixtron, there is an ada ( and fortran! ) compiler. (however asciilifeform has not tried either item, beyond 'hello world' )

asciilifeform: 'new shithouse is built from the rubble of the old, not imported brick' or how did lenin put it.

asciilifeform: ( observe that ch1 ffa actually starts out by http://www.loper-os.org/?p=1913#selection-337.0-1150.0 , i.e. banning 90% of ada.. why didja suppose this was. )

asciilifeform: i have even moar wtf items in my collection.

asciilifeform: the world of the folx who wrote ada for moneys in the saeculum, is not , as i gather, a happy place. sorta why asciilifeform had to learn buncha things from 1st principles, rather than by reading their ugh

asciilifeform: ( observe, the only 'access pointer' in ffa is where it sucks in cmdline args from unix )

asciilifeform: mircea_popescu: sorta why i suspect 'greybeard'. prolly he also had a 200kloc (or 2e6...) that had to be dealt with as-is, pointerola & all.

asciilifeform: i suspect he was 'old guard' (ada greybeard) and the bottle took its toll.

asciilifeform: sadly vanished in '15

asciilifeform: ( pretty lively pedagogy, too, e.g. http://btcbase.org/log/2017-11-20#1741487 ) ☝︎

asciilifeform: re: ada elaboration, i found http://btcbase.org/log/2017-11-20#1741485 item useful when i was last puzzling over subj ☝︎

asciilifeform: incidentally i cant think of any reason it wouldn't work exactly same with mysql, with the exception of where i dun yet know how to do the o(n log n) indices in mysql

asciilifeform: ( phuctor connects to a local instance of postgres, so in that sense similar )

asciilifeform: so like phuctor.

asciilifeform: aa

asciilifeform: hm it actually links in ~all~ of mysql ? or just has the typical 'connect to local instance' skin ?

asciilifeform: lol prolly no bdb also

asciilifeform: aint there a multi-MB '3d' lib in there also ?

asciilifeform has not studied mircea_popescu's proggy, of yet, doesn't dare to comment in detail

asciilifeform: exactly like trb then

asciilifeform: (i.e. cut into sections that can run reentrantly)

asciilifeform: i strongly suspect it'll have to be de-threaded to work reliably with ada-cum-tasks , sadly

asciilifeform: it's a trb-like item

asciilifeform: oh i recall

asciilifeform: BingoBoingo: hey, if microshit can have weev jailed for munging URLs, why not acade-microshit and some other d00d , similarly

asciilifeform: cuz in the vice-versa variant, what you'll have is ~two~ threading systems that dun know about one another, and the shit one (cpp) doesn't have any concept of sane locks etc

asciilifeform: diana_coman: i suspect that even if you get the thing to properly link, you will discover new headaches on acct of http://btcbase.org/log/2019-01-06#1885177 . really imho proggy that has ada tasks oughta have ada main , and (if must) call static cpp turdola, not vice-versa ☝︎

asciilifeform brb,teatime

asciilifeform: last night i re-read diana_coman's piece on m-r , it is interesting just how much sweat diana_coman had to put in simply on account of koch gnarl ☟︎

asciilifeform: Mocky: ch16 will be about m-r .

asciilifeform: Mocky: ( yes stein is a pretty simple item, quite unlike e.g. barrett, but still gotta give pedantic convergence proof for it )

asciilifeform: Mocky: proggy written, nao writin' proof

asciilifeform: mircea_popescu: http://trilema.com/2019/i-think-it-might-be-the-case-im-becoming-pretentious/#comment-127374

asciilifeform: guten morgen, Mocky

asciilifeform: http://btcbase.org/log/2019-01-09#1886064 << iirc ave1 found how to expose the init knob explicitly. you may want to try his method ☝︎☟︎

asciilifeform: http://btcbase.org/log/2019-01-09#1886056 << i tried and failed to find & buy whatever it is they make for the sea ( why settle for konsoomer plastic ) but no dice, it aint gettable anywhere ☝︎

asciilifeform: ( mine has no elaboration . but i escaped using all sortsa things )

asciilifeform: http://btcbase.org/log/2019-01-09#1886065 << hmm diana_coman , what piece in your proggy actually uses elaboration ? ( is it the tasking ? ) ☝︎☟︎

asciilifeform: http://btcbase.org/log/2019-01-09#1886058 << i set mine to max when i uncrated it, but noticed that 'overshoots'. turned out, they aint acceleration modes, but pulses-per-radian, and http://btcbase.org/log/2019-01-08#1885844 actually made it feel proper (i.e. can turn the roller 2pi, and then -2pi, and come to same pixel ) ☝︎☝︎

asciilifeform: http://btcbase.org/log/2019-01-09#1886060 << ah ty phf , precisely what i wanted to learn ☝︎

asciilifeform: http://btcbase.org/log/2014-03-18#565557 << in lulzy 2014isms ☝︎

asciilifeform: neato mod6 , i look fwd to reading

asciilifeform brb,meat

asciilifeform: or do i misread, and mod6 already reground, and writing re how.

asciilifeform: trb has , what, 3x the # of patches, so it'll take you 30m at most.

asciilifeform: mod6: at the risk of sounding like mircea_popescu in earlier thread -- why is this a mega-project ? i reground ffa to keccak in about 10minute (after getting hold of a working keccak-vtron)

asciilifeform: wb mod6

asciilifeform: tho not only ( also talked to ben_vulpes and other folx who tuned in prior to asciilifeform )

asciilifeform: tru

asciilifeform: ( i find what e.g. mircea_popescu said in 2014 , interesting, but not xyzputin3333 )

asciilifeform: might be interesting ( phf? ) to have a filter toggle that only shows people who are still alive

asciilifeform finds good % of the 2014-prior l0gz to be indigestible. (tho regularly goes to read'em anyway)

asciilifeform: mircea_popescu: i not only gotta write proggy, but to make sure that it makes sense to people

asciilifeform: i dun have the necessary figure yet. hence 'i'ma come back'.

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: the ~minimal~ count

asciilifeform: then you only m-r on inputs that already passed gcd

asciilifeform: ~doing~ the gcd pre-sieve

asciilifeform: the opposite

asciilifeform: no

asciilifeform: it potentially frees up cpu for moar shots of m-r, was the idea

asciilifeform: how 'many r-m shots' is a free variable tho.

asciilifeform: right

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 )

asciilifeform: i'd be satisfied with gcd as initial sieve (given that i have one that dun leak anyffin) ~followed~ by m-r.

asciilifeform: well at least not in the initial sieve ( one might still want it for e.g. pocklington's test )

asciilifeform: right

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: )

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: mircea_popescu diana_coman et al ^

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