log
▁▁▁▁▁
asciilifeform finally eaten log...
asciilifeform: http://btcbase.org/log/2017-10-07#1722394 << this looks very , very painful to prove correctness of. i'ma come back to it.☝︎
a111: Logged on 2017-10-07 21:09 apeloyee: asciilifeform: turns out a simple, ffa-suitable O(N^2) algorithm exists for GCD. This is adapted from GMP docs with one extra operation in the loop: http://p.bvulpes.com/pastes/oupUJ/?raw=true . Note: the code as posted is likely wrong, but I'm sure the idea can be made to work.
asciilifeform: http://btcbase.org/log/2017-10-07#1722395 << compute and then what ? gotta multiply☝︎
a111: Logged on 2017-10-07 21:14 apeloyee: http://btcbase.org/log/2017-10-07#1722289 << and the point of doing karatsuba is? you do 2 recursive calls to Mul_Karatsuba_TopOnly and one to Mul_Karatsuba. should've simply calculated upper_part(XLo*YHi), upper_part(YLo*XHi) and XHi*YHi
asciilifeform: http://btcbase.org/log/2017-10-07#1722397 << i don't see anything that only wants ~lower~ half... whatcha talking about☝︎
a111: Logged on 2017-10-07 21:25 apeloyee: the multiply-by-approximate quotient in barrett's also needs only the lower part (plus 2 extra bits to the left), and lower part of product can be computed exactly (since rounding is not a problem)
asciilifeform: http://btcbase.org/log/2017-10-07#1722400 << bernstein's gcd method is neither here nor there, i certainly don't need anything of the kind in ffa, and quite likely it fundamentally does not ffaize☝︎
a111: Logged on 2017-10-07 21:28 apeloyee: http://btcbase.org/log/2017-10-05#1721485 << i thought bernstein's "how to find smooth parts of integers" suggests a remainder tree, not gcd?
asciilifeform: http://btcbase.org/log/2017-10-07#1722402 << this is a fundamentally wrong way to generate cryptographic primes. we had a thread about it, http://btcbase.org/log/2017-08-14#1697562☝︎☝︎
a111: Logged on 2017-10-07 21:48 apeloyee: http://btcbase.org/log/2017-10-05#1721485 << alternatively, can *construct* numbers which don't have very small factors. pick a nonzero remainder mod 2, mod 3, ... mod largest-prime-fit-in-your-primorial and find what number of primorial is congruent to it using chinese remainder theorem
a111: Logged on 2017-08-14 16:14 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: the ONLY correct method of generating cryptoprimes, is to 1) get N bits from FUCKGOATS 2) determine, in fixed spacetime every single time, whether that string of bits constitutes a usable prime.
asciilifeform: all other methods leak info via timing , amperage, rf noise.
asciilifeform: http://btcbase.org/log/2017-10-07#1722405 << in no case can the 'cheap initial primality test' primorial exceed the size of current ffa width. thinkaboutit.☝︎
a111: Logged on 2017-10-07 21:53 apeloyee: the primorial has to be, say, 2^32 times less than the ffa maxint. then you can add randomnumber*primorial, and such a number is equally likely to any prime from some interval
asciilifeform: http://btcbase.org/log/2017-10-07#1722408 << you might consider reading the code ? it has all been posted.☝︎
a111: Logged on 2017-10-07 22:39 phf: http://btcbase.org/log/2017-10-07#1722379 << this is probably true but only because ffa mutates an array of bigits, where's any language level bignum system produces a whole new one for each operation
asciilifeform: http://btcbase.org/log/2017-10-07#1722411 << 1 ) ffa is closed form. i.e. it CAN be written as a number of nand gates, with a 'funnel' at the top, to which you present a,b,c, e.g. 4096bit, numbers, and at the bottom in a little cup you get a^b mod c , and with NO UPWARDS FEEDBACK FLOW of information , i.e. answer comes after same interval of time always, and with strictly downwards signals.☝︎
a111: Logged on 2017-10-07 22:44 phf: http://btcbase.org/log/2017-10-07#1722374 << >> http://btcbase.org/log/2017-10-07#1722376 << this seems contradictory, because the python thing posted is not closed form
asciilifeform: but 2 ) the python example is of course not closed form, and it is imho meaningless to even attempt to write the closed form item in a language like python or cl
asciilifeform: ( where there is no assurance of not consing and not branching )
asciilifeform: http://btcbase.org/log/2017-10-07#1722415 << if you have a comp the size of jupiter, you could ~maybe~ have such a thing as a 128bit primorial.☝︎
a111: Logged on 2017-10-07 23:50 mircea_popescu: http://btcbase.org/log/2017-10-07#1722405 << this may actually be a better check than any miller-rabin, and at any rate a good complement. gcd with primorial.
asciilifeform: but certainly not 129.
asciilifeform: so no, nobody is replacing miller-rabin with gcd(primorial, x).
asciilifeform: ( certainly not even for as large a number as 64bit... much less 4096 )
asciilifeform: i proposed primorial strictly as an initial winnowing to replace the idiot trial divisions koch et al used.
asciilifeform: phf, mod6 : funnily enough i went and tried the 'fair fight' max(4096b) a^b mod c in python, http://wotpaste.cascadianhacker.com/pastes/GHATB/?raw=true , but it... bombs
asciilifeform: with eggog:
asciilifeform: OverflowError: Python int too large to convert to C long
asciilifeform: for the obvious reason.
asciilifeform: if somebody wants to make the physically possible version of this, to see what happens on max hammingweight...
asciilifeform: oh ffs, here goes,
asciilifeform: http://wotpaste.cascadianhacker.com/pastes/0A6fb/?raw=true << python 'fair fight' ver
asciilifeform: ( and skips unused hammings... )
asciilifeform: 0.018s on this box.
asciilifeform: http://wotpaste.cascadianhacker.com/pastes/saynG/?raw=true << all1s. 0.028s. tho i do suspect it shortcuts internally.
asciilifeform: ( mainly, i suspect, by recognizing masses of 0 in karatsuba and returning 0 when they get mul'd )
asciilifeform: tldr : initial py snippet i had lying around was braindamaged.
asciilifeform: http://btcbase.org/log/2017-10-07#1722367 << i gotta ask if this figure included sbcl load time !?☝︎
a111: Logged on 2017-10-07 16:42 shinohai: I get 0m1.236s using sbcl (i5)
asciilifeform: because if it did, it is meaningless
asciilifeform: ( see phf's http://btcbase.org/log/2017-07-03#1678660 , or even naggum's, re why )☝︎
a111: Logged on 2017-07-03 14:46 phf: i think ascii already made that point, that if you're profiling lisp with the vm startup, then you should also profile c machine from boot time. at the very least the vm should be warmed up by loading all the dependencies into the core, doing save-lisp on it, and then making sure that your foo.lisp has an up to date fasl. inside lisp though to achieve the optimizations you run variants of your function inside (time ...) until you bring it within the ra
asciilifeform: ( orig., iirc, thread : http://btcbase.org/log/2017-07-02#1678490 )☝︎
a111: Logged on 2017-07-02 12:50 asciilifeform: http://btcbase.org/log/2017-07-02#1678460 << how about we roll the boot time ( to shell!! ) of your cmachinekernel, how about?
asciilifeform bbl
mats: l0l an amzn frontend engineer friend has to work all through christmas week, got his vacation request denied by upper mgmt
mats: he put it in almost four months in advance and still can’t take a few days off
BingoBoingo: Well, he works in the retail industry. What should he expect?
mircea_popescu: http://btcbase.org/log/2017-10-08#1722429 << your chances of generating a random int that is also prime at that sort of length aren't so great.☝︎
a111: Logged on 2017-10-08 00:16 asciilifeform: the ONLY correct method of generating cryptoprimes, is to 1) get N bits from FUCKGOATS 2) determine, in fixed spacetime every single time, whether that string of bits constitutes a usable prime.
mircea_popescu: having a primorial at the ready to exclude a large number of common (ie, low) factors in one single gcd likely speeds this up significantly.
mircea_popescu: http://btcbase.org/log/2017-10-08#1722442 << not altogether, hold on to your horses.☝︎
a111: Logged on 2017-10-08 00:24 asciilifeform: so no, nobody is replacing miller-rabin with gcd(primorial, x).
mircea_popescu: recall diana_coman 's trick of "multiply by 6" ? pretty much the inverse of the same idea.
asciilifeform: http://btcbase.org/log/2017-10-08#1722468 << quite acceptable, 1 in few thou☝︎
a111: Logged on 2017-10-08 01:34 mircea_popescu: http://btcbase.org/log/2017-10-08#1722429 << your chances of generating a random int that is also prime at that sort of length aren't so great.
asciilifeform: ( try it )
mircea_popescu: yes, but then would you rather 999 r-m or 995 primorial gcd and 4 r-m ?
asciilifeform: http://btcbase.org/log/2017-10-08#1722470 << is why i suggested it to begin with, zaps items with factors up to 16bit or so quickly☝︎
a111: Logged on 2017-10-08 01:35 mircea_popescu: having a primorial at the ready to exclude a large number of common (ie, low) factors in one single gcd likely speeds this up significantly.
mircea_popescu: so then what exactly is the argument about.
asciilifeform: http://btcbase.org/log/2017-10-07#1722415 looked like a 'who needs miller-rabin'☝︎
a111: Logged on 2017-10-07 23:50 mircea_popescu: http://btcbase.org/log/2017-10-07#1722405 << this may actually be a better check than any miller-rabin, and at any rate a good complement. gcd with primorial.
mircea_popescu: yeah well.
asciilifeform: ( where first suggested , ftr : http://btcbase.org/log/2017-08-14#1697598 )☝︎
a111: Logged on 2017-08-14 17:15 asciilifeform: idea is, for pre-millerrabin litmus, take gcd(candidate, Qw) where Qw is largest primorial that fits in the ffawidth
asciilifeform: worxgreat
mircea_popescu: incidentally, if looking for 4096 bit prime wouldn't the correct approach be to take 4094 bits of rng and glue 1 on either end ?
mircea_popescu: as no 0 led or 0 terminated string will ever pass anyway
asciilifeform: primes >2 are odd, noose at 11
mircea_popescu: aha. get some free bits that way, fwiw.
asciilifeform: ( yes you set the low bit to 1 )
asciilifeform: http://www.loper-os.org/?p=568&cpage=4#comment-18272 << in other strange.
shinohai: Lol asciilifeform got a brony
shinohai: Oh wait, name is "Tina"
asciilifeform: i have deeply nfi
asciilifeform: in other lullies, http://www.loper-os.org/pub/nsawagenhoneypot.jpg << found on washington metro train today
shinohai: TOP KEK "Anonymous access through tor browser"!!!!
shinohai: They even bothered to vanitygen a custom tor addy
asciilifeform: ads on these trains, ftr, not cheap. ( and certainly not 'to allcomers' )
mats: saudis join turks in s-400 purchases
mats: washington must be glowering
lobbes: !Qhelp
lobbesbot: lobbes: http://lobbesblog.com/lobbesbotcommands
lobbes: !~later tell mircea_popescu ^^ 'help sexpr' and 'help json' also working. lobbesbot has been brought up to spec
jhvh1: lobbes: The operation succeeded.
mircea_popescu: http://btcbase.org/log/2017-10-08#1722492 maybe tina's looking for a new home.☝︎
a111: Logged on 2017-10-08 01:58 asciilifeform: http://www.loper-os.org/?p=568&cpage=4#comment-18272 << in other strange.
mircea_popescu: http://btcbase.org/log/2017-10-08#1722496 << bwahaha, http://btcbase.org/log-search?q=private+internet+access is quickly becoming a portion of the gosplan "gdp" innit.☝︎
a111: Logged on 2017-10-08 02:02 asciilifeform: in other lullies, http://www.loper-os.org/pub/nsawagenhoneypot.jpg << found on washington metro train today
mircea_popescu: is that up left item supposed to be the pennsylvania bell ? or rather some ad-hoc, tesla times large inductor ? perhaps some nuclear sikrit ?
mircea_popescu: http://btcbase.org/log/2017-10-08#1722501 << pretty weird, middle east tensions apparently resolved through tribute ? i can make no sense of it whatsoever.☝︎
a111: Logged on 2017-10-08 04:24 mats: washington must be glowering
shinohai: The sound of that bell instantly alerts patriotfags and sends Cuban diplomats running, complaining of sonic attacks.
mircea_popescu: oh i see
mircea_popescu: is it patriotic to leak the dnc's self-important bullshit leading to the republic sinking clinton ?
mircea_popescu: !Qhelp sexpr
lobbesbot: mircea_popescu: http://lobbesblog.com/lobbesbotcommands.sexpr
mircea_popescu: lobbes cool deal, grats on being FIRST!\
spyked finally processed ~a week's worth of logs. hi all!
spyked: http://btcbase.org/log/2017-09-29#1718287 <-- would buy, esp. if custom pcb cannot be avoided (and I suspect this is the case).☝︎
a111: Logged on 2017-09-29 16:39 asciilifeform wonders whether anybody would actually buy a generic fpgatronic packet eater-shitter
mircea_popescu: lol wd.
mircea_popescu: as it happens i was JUST reading http://trilema.com/2017/the-practical-costs-of-hallucinated-freedom/#selection-549.0-549.226 which is eerily apropos.
spyked: http://btcbase.org/log/2017-09-30#1718769 <-- afaik ubb ran a "digitalization" program for library. but they prolly won't make those public, eh?☝︎
a111: Logged on 2017-09-30 19:14 mircea_popescu: sorry asciilifeform . all i have are my own notes, which are as all hand notes useless without hte backing of the library of origin (in this case, the universitary library of cluj). teh interwebs dun seem to have a "here's the list of trotsky letters".
mircea_popescu: here's the problem of ourdemocracy : if people are allowed to misperceive they "have choices" and feed this hallucination out of the field of choices they do not have, you end up with the sexy problem, whereby linear cost encounters unreasonably power-law distributed benefit.
mircea_popescu: so yes, they digitized the "sexy" parts.
mircea_popescu: you take my meaning ?
spyked: yeah, thought this is probably the case
spyked: (read through the old newspapers at some point. I think it was through Trilema piece?)
spyked: http://btcbase.org/log/2017-10-01#1719047 <-- lol! this reminds me of a horatiu malaele piece (Romanian actor/comedian), "doua vaci". wait, it was on the interwebz (I should translate it anyway at some point)☝︎
a111: Logged on 2017-10-01 04:06 mircea_popescu: "If pet food companies used the same business model as startups: Jim creates a dog food factory and gives away dog food for free. 450 million dogs line up for free dog food. Purina Dog Chow understands that non-paying dog food consumers are currency, and buys Jim’s factory for $42 per dog." << in other historical elaineo lulz.
mircea_popescu: spyked such as in eg http://trilema.com/2011/sa-ne-agitam-putin-pe-tema-notatiei-limbii-romane and etc yes.
mircea_popescu: spyked probably the peak of his work huh ? see if you get that gherman fellow to read it once you translate it!
spyked: http://www.220.ro/umor-romanesc/Horatiu-Malaele-2-Vaci/nUSRHOCJEP/ <-- ro. only, unfortunately; also buried under piles of shitads, but I salvaged the videofile.
spyked: good idea. will give it a shot
spyked: http://btcbase.org/log/2017-10-03#1720147 <-- whatabout ol' zamyatin? We was not bad imo.☝︎
a111: Logged on 2017-10-03 13:34 mircea_popescu: even exists in early anglo stuff, christ resurrector, christ almighty etc. though the vein exhausted itself readily and apparently without leaving much trace. i guess in the same way "everyone" knows of bedwetter's 1984 but nobody read point counterpoint, notwithstanding that huxley is the important kid in that class, not fucking blair.
mircea_popescu: how is zamytatin a british writer ?
mircea_popescu: and for that matter how is pulp fiction part of literature now.
spyked: wasn't, but wrote about "big brother" before it was cool, neh?
mircea_popescu: i guess the argument could be made, but what's it pay ?
mircea_popescu: there's a petulent dork going about on twitter about he "wrote about bitcoin before it was cool". i suspect the whole "cypherpunk" group of kanzure s secretly hold the same belief, that they're relevant through their failure.
mircea_popescu: you see some value in this ?
spyked: same value I see in "propuneri de masuri pentru imbunatatirea [...]". historical.
mircea_popescu: sure.
mircea_popescu: i would agree that zamyatin is a brilliant pamphleteer and an interesting ethnological/"historical" source.
phf: http://btcbase.org/log/2017-10-08#1722496 << heh i was marveling at the same ad couple of days ago, which i assume means they are all over the trains. i wonder if it's sunlight foundation project☝︎
a111: Logged on 2017-10-08 02:02 asciilifeform: in other lullies, http://www.loper-os.org/pub/nsawagenhoneypot.jpg << found on washington metro train today
BingoBoingo: From the "Why not make the trilema re-read of the now a surprise files" https://archive.is/pPEfc
phf: nope, it was started by john tye, https://arstechnica.com/tech-policy/2014/08/meet-john-tye-the-kinder-gentler-and-by-the-book-whistleblower/. i suppose he should be classified as a chair shuffler
spyked: mircea_popescu, I'm not sure how I would evaluate it other than by looking at the "boy has no aspirations of his own; boy meets girl; boy gets in trouble; boy gets face stomped by boot" trope that's repeated throughout dystopian novels; there's probably more to it than that, but if there is, I'm not equipped with the literary baggage to see it. Orwell is fashionable nowadays because pantsuit equates Trump with big brother, and... so
spyked: what?
phf: BingoBoingo: is that archive.is link working for you? i've been getting routing resolution errors from cloudflare, pretty consistently part few months
mircea_popescu: spyked better error messages.
BingoBoingo: phf: It is working at the moment
mircea_popescu: BingoBoingo wait, what ?
phf: "You've requested a page on a website (archive.is) that is on the Cloudflare network. Cloudflare is currently not routing the requested domain (archive.is). There are two potential causes of this:"
BingoBoingo: mircea_popescu: For the joy of seeing an underlinked classic load. http://trilema.com/2014/gotta-love-that-negro-speak/
mircea_popescu: phf ahgahaha
mircea_popescu: there goes eh.
mircea_popescu: i hope we all got zips.
mircea_popescu: also fwiw, i see it.
BingoBoingo: "That’s just motherfuckin’ mother nature" (TM)(R)
phf: well, i've been getting that error for the past two months. while there was still a heavy archive.is exchange in the logs, os i thought it's something to do with russia. i'm still getting it in u.s. though :o
phf: so i wonder if my machine has been fingerprinted somehow
phf: oh never mind false alert, i have an explicit archive.is host in my hosts file
phf: it was confusing because i was getting legitimate cloudflare issues all the while i was in russia. now i wonder if it's some dns "firewall" propagation issue
mircea_popescu: i don't get it. so you had a specific ip, which used to work, but now they changed it and instead of failing they self-advertise ?
spyked: ftr, I have archive.is in hostsfile with a different IP than the one currently returned by DNS, and not getting a cloudflare page.
phf: my specific ip was a cloudflare ip 104.28.25.2. current dig for archive.is resolves to 195.123.218.180, which is a netherlands "mobicom ltd" range. i suspect that archive.is took themselves off cloudflare in the last some months, so now i'm hitting cloudflare proxy servers and they are complaining that the host: is no longer served
phf: spyked: what ip do you have?
spyked: in hosts, the one you mentioned (195.123.218.180), while host archive.is returns 94.242.57.138
phf: odd, that last one is russian, "ooo fishnet" :)
mircea_popescu: phf i see the same ip as you do.
phf: cloud!
mircea_popescu: spyked seems you're poisoned ?
mircea_popescu: that 94 thing isn't anywhere.
phf: spyked: are you in u.k.?
mircea_popescu: the sudden silence of omg.
spyked: phf, nope. ro. so... let's try some more dns servers.
phf: because the ip address is russian, but registered to a u.k hosting provider, redcentric
shinohai: Weirdly enough when I dig archive.is I get:
shinohai: ;; ANSWER SECTION:
shinohai: archive.is. 201 IN A 84.22.118.22
mircea_popescu: this dns thing...
spyked: dig through Google DNS: 91.235.136.108; dig through romtelecom DNS: 91.228.152.189; so yeah, mircea_popescu might be on to something here. spyed flushes dns caches.
mircea_popescu: well, in other lulz, apparently it's 91.228.152.189 in india.
mircea_popescu: and 88.99.177.114 in brazil.
mircea_popescu: phf i dun think they got off cloudflare tbh, this looks like exactly the sort of crap.
mircea_popescu: 94.242.57.138 in mexico.
spyked: speaking of cloud: http://btcbase.org/log/2017-10-05#1720904 <-- lulz. SGX is ripe with side channels. progz can leak data through page faults (and any other "exceptional" interaction with OS, basically).☝︎
a111: Logged on 2017-10-05 00:12 ben_vulpes: > Finally, we abuse Intel SGX to hide the attack entirely from the user and the operating system, making any inspection or detection of the attack infeasible.
mircea_popescu: rife.
mircea_popescu sits here trying to remember the name of the irrelevant dork with the guns. after a while the best lead i have is "hacker lexicon" was it ? google produces nothing but wired crap ; if treated with a -wired sprinking, suddenly catb.org "jargon file" is top result.
mircea_popescu: why does wired exist again ?
spyked liked ripe moar. as in, full of low-hanging fruit
mircea_popescu: oh, to try and parasitize google-reality ? just like any other zynga facebookism ? awww..
mircea_popescu: spyked ok, but there's no such thing as "ripe with".
mircea_popescu: if it's ripe it's ripe.
mircea_popescu: and for that matter rife, as in "rife with fleas" is how english says "full of low hanging fruit".
spyked: ty for explanation, mircea_popescu. oddly, it seems there's no strong etymological relation between the two. (ripe's related to reap, rape; rife, just Germanic for abundancy?)
mircea_popescu: somethinglike that.
asciilifeform: mircea_popescu: http://btcbase.org/log/2017-10-08#1722595 << esr ?☝︎
a111: Logged on 2017-10-08 13:27 mircea_popescu sits here trying to remember the name of the irrelevant dork with the guns. after a while the best lead i have is "hacker lexicon" was it ? google produces nothing but wired crap ; if treated with a -wired sprinking, suddenly catb.org "jargon file" is top result.
mircea_popescu: yeah fished him out eventually.
asciilifeform: aite
mircea_popescu: assistence went "you don't remember esr ?" and i went "what would i remember him for ?" and they went "is this a bit ?"
mircea_popescu: no it's not a fucking bit. even if i sometimes sound just like a character, it's purely fucking accident!
spyked: http://btcbase.org/log/2017-10-05#1720982 <-- considering this as a "learn Ada" project. will share code once I have minimum item worth of showing (processing r5rs and tinyscheme meanwhile)☝︎
a111: Logged on 2017-10-05 02:17 asciilifeform: however it would very conspicuously rock , imho, if one of you folx did this, instead of waiting.
asciilifeform: neato spyked . keep in mind that you gotta use the ada subset displayed in ffa.
phf: spyked: r5rs and tinyscheme are not the right places to start on the other, non-ada end, i'd recommend looking at lisp in small pieces. you can tease out the theory out of tinyscheme, but it's definitely easier not to get bogged on accidentals if you start from theory
asciilifeform: 'small pieces' is pretty good b00k
spyked is going through Kogge's book at the moment, incidentally. much lower level, but it should help on Lisp internals.
phf: spyked: i'd also recommend staying away from continuations, they are a cute hack and flow out of some of the classical scheme interpreter designs (i.e. CPS transform), but they are not very useful in production. instead i'd go for a tagbody that gets compiled to a bunch of jmps. in practice tagbody solves 99% of cont problems
phf: http://clhs.lisp.se/Body/s_tagbod.htm
spyked: anyway, I have much of Scheme in head. I'm looking at tinyscheme mostly to figure out what "subset of r5rs" they've implemented. though I'm expecting tinyscheme subset isn't necessarily the same as tmsr-needed subset.
mircea_popescu: in other fucking lulz : http://academic.research.microsoft.com/Paper/1285929.aspx (original title was "beyond the centralized mindset")
asciilifeform: mircea_popescu: i see eggog
mircea_popescu: that'd be the point now wouldn't it.
deedbot: http://trilema.com/2017/the-sexy-problem-formalized/ << Trilema - The sexy problem, formalized
asciilifeform: aa
mircea_popescu: https://llk.media.mit.edu/papers/archive/decentralized/ if you care for some reason. typical mit ocean-toeing.
phf: spyked: keep in mind that all the lispers here are common lisp programmers, so a ~practical~ scripting lisp would be LISP-family themed, rather than an explicit scheme. that's my personal experience with trying to get useful things out of shiva: having to write a bunch of "missing" hyperspec functions. asciilifeform said something similar in the past
mircea_popescu: was candi_lustt cmucl or what was it ?
phf: sbcl
mircea_popescu: aha k.
mircea_popescu: possibly worth linking the s400 thing : https://breakingdefense.com/2017/10/saudis-rock-us-alliance-say-theyll-buy-russias-top-plane-killer-s-400/
phf: fwiw all our production lisp runs on sbcl, including btcbase. as much as i'm pimping cmucl, it's not "modern" enough to host a website on unix. i still think it's a better target for a hypothetical on the iron common lisp
mircea_popescu: phf if you had a blog and time you could sit down to do the whole discussion of that so as to inform future policymaking above and beyond simple amoebic continuation.
spyked: phf ty for lisp in small pieces reference. will look through it.
mircea_popescu: " “The Saudi’s will not be able to link the S-400 with Saudi’s current (US and Europe-sourced) infrastructure, nor will they be able to connect the S-400 with US systems." << says who the everloving fuck.
phf: i've started writing blog posts to "toughen my hand", but it's rough going, i'll add it to the list of things to write about
mircea_popescu: phf hand mollifies in solitude. to toughen must expose.
mircea_popescu: oh, and in random romanian learnings of today, http://trilema.com/2010/care-sunt-iq-tau/
mircea_popescu: if you can read the list so it sounds like a poem, your romanian pronounciation is probably acceptable.
asciilifeform: apparently that easy huh
mircea_popescu: romanian is read as written!
asciilifeform: mircea_popescu's www 'cheats' by omitting the orc letterz
mircea_popescu: ehehe
mircea_popescu: I suppose i could do a phonetics of this.
mircea_popescu: why the hell not, the girls are worth the trouble. brb.
asciilifeform: word for girl vs face say
asciilifeform: differ by an orc glyf
mircea_popescu: tru.
mircea_popescu: asciilifeform http://trilema.com/2010/care-sunt-iq-tau/#comment-123116
asciilifeform: wassis
spyked: lulz!
asciilifeform: oh hey
asciilifeform: speaking of ro, http://perevod99.blogspot.ru/2011/08/blog-post_23.html << ru pro linguist with some decades of ro habitation, various lulzy posts re subj and other.
asciilifeform: ( or possibly i misread re habitation. but still fountain of lulz. )
mircea_popescu: o hey
mircea_popescu: is the kak ?!?! tot acolo joke rendered ?
mircea_popescu: http://btcbase.org/log/2015-11-21#1329338 << guess what, it's in the logs.☝︎
a111: Logged on 2015-11-21 18:55 mircea_popescu: asciilifeform btw ever told you the joke of the muscovite trying to take a shit in bucharest ?
asciilifeform: aayes
mircea_popescu: (cac in romanian being, obviously, off the caco cacare latin root)
mircea_popescu: asciilifeform ftr it was in plenum of the parliament thingee.
asciilifeform went on a tr kick and noticed that dulap is a turkish, i.e. dolap ( crate )
mircea_popescu: also fellow misrepresents the failure of the 1989 bucharest meeting. the events flew more or less thus : timisoara rebelled, ceausescu verbally ordered armed repression, on the basis of some discussion, but (most likely deliberately) omitted to actually issue the proper paperwork. the war minister killed himself. ceausescu ordered a meeting organised in bucharest, to announce (in the dead of a bitter winter, without proper he
mircea_popescu: ating in most homes) a list of easements, the chief buttress of which being an extra 100 lei allowed to pregnant mothers (about 1 dollar in black market rates of the time). then the people were dismissed, and they started to leave, but for some entirely to this day unknown reason they were called back. except, they DIDNT go back in order, which means the politruks didn't know who's who.
mircea_popescu: THEN it exploded.
mircea_popescu: there's no "small voice".
mircea_popescu: asciilifeform yep.
asciilifeform: why would fuhrer deliberately omit the paperwork
mircea_popescu: anyway. classical ceausescu regime "demonstrations" were organised very much on the ddr bank model, where you could eventually find the address of every single byte.
mircea_popescu: but you need the array to do that.
mircea_popescu: asciilifeform because furher was a chickenshit ? like any other 90yo ?
mircea_popescu: who knows.
asciilifeform: gorby wasn't even 60 and already chicken
asciilifeform: ( and/or bought )
mircea_popescu: ceausescu was not a gallant sort of brave fellow. during a 1977 miners' revolt (they beat up the commie locals etc) he went there and very much shat it.
mircea_popescu: he was, however, fundamentally brave, in that what he told teh us style "judges" at his "trial" was that he'd much rather be dead.
asciilifeform: this last part is well-known
asciilifeform: in ru sphere it often is chalked up to the proverbial 'дедушка старый - ему всё равно'
asciilifeform: rather than bravery
mircea_popescu: possibly.
mircea_popescu: still, how man faces death is a hard to paper over factor.
mircea_popescu: https://www.youtube.com/watch?v=KTTVfid33yY << the well publicized moment of "wtf is going on here" with rando politruk going "silence silence" in the background.
mircea_popescu: "alo" being what you say in a phone.
shinohai: !~translate so to en oogi boogi nigga nigger
jhvh1: shinohai: charges associated wound nigga nigger
asciilifeform: mircea_popescu: lol is that the sniper-roof
mircea_popescu: possibly.
mircea_popescu: shinohai lol!
shinohai: ^ I heard the above was edited by sjw on Google translate. It used to be "Take a look at the nigger"
mircea_popescu: asciilifeform anyway, it's a great postmodern moment you know, failed mangod and his broken phoneline to reality. "hello ? hello ? *mutters under breath* this is a provocation"
mircea_popescu: ("provocation", in ourdemocracy lingo of the time, is what you called enemy maneuvers.)
mircea_popescu: (because of reasons discussed in http://trilema.com/2014/the-problem-of-ideal-social-systems-reprint/ socialisms can't have categorical terms, defined in the normal manner, but must always include the ethical color of all words in the words. so "movement" becomes either provocation (bad) or progres (good) and so following for everything, stalin's cup is named by a different cupword than hitler's self-same identical cup)
asciilifeform: aufklärung!
mircea_popescu: rather.
mircea_popescu: anyway, there's a pile of disinfo and general crap surrounding the events. as an example : on 2-3 dec gorbachev hung out with bush on a soviet ship. on 4th, there was the wasaw pact meeting. gorbachev was well excited of whatever, the new bulgarian (mladenov, his college pal) and generally the western press coverage.
mircea_popescu: he wrote bucharest a general letter about it, proposing to explain what happened on the 4th. ceausescu responded that unless they have a private meeting he's not coming altogether.
mircea_popescu: this was reported (via the entirely irrelevant "sources" of talbot, beshloss & co) as "the agitation of stalinist tirant" at warsaw, "unhappy with the us-euro offensive to liquidate communism".
mircea_popescu: in point of fact, ceausescu refused to sign off on some paper establishing 20 years after the fact that the invasion of czechoslovakia was a mistake. for the fucking obvious reason that he condemned the russians at the time, and according to readily forgotten "consensus" at the time, at no small personal risk.
mircea_popescu: and he also wasn't impressed with gorbachev;s verbiage as to "disarmament", seeing how romania was spending about 3% of pib on arms in the 85-90 5year plan, whereas ussr had never went under 30% yet.
mircea_popescu: and so on.
asciilifeform: very easy to 3 instead of 30 when you dun gotta do the rocket, n00kz, etc
mircea_popescu: no argument there.
mircea_popescu: but you have to also compare and contrast with romania's declared, and tirelessly promoted official philosophy of pace and "unmingling in internal affairs of sovereign states".
mircea_popescu: guy never saw himself as much more of a su ally than saudis see themselves us allies i dun suspect.
asciilifeform: possibly fancied himself tito ?
mircea_popescu: something. consider actual live events : gorbachev says at the meeting, once they move on past his insistence on having visited the pope as if anyone gave a shit about that "we are all here, who were implicated in the czech affair, except romania, that had exited then".
mircea_popescu: guy replied "romania exited nothing, romania didn't go in, so it had not what to exit."
asciilifeform: in tito's case , and for that matter kim ir sen's -- 'throne is mine, i won it as partizan commander in the war, took no payola from foreign devils' was tru. but how did the shoemaker get ~his~ throne
mircea_popescu: "nu-i adevarat (that's not true), romania nu a iesti din aceasta problema (romania didn't exit the matter), romania nu a intrat in cehoslovacia (it never entered) asa ca nu avea de unde sa iasa (had not what to exit)"
mircea_popescu: asciilifeform elected!
asciilifeform: lollected
mircea_popescu: (this isn't even false, classical case of "sickly old pope")
mircea_popescu: no, actually.
mircea_popescu: the conclave of cardinals picked the least competent of themselves to sit on throne.
mircea_popescu: anyway, the whole meeting went in that vein, ceausescu pointed out to soviet troops still at praga, gorby was liek "oh, that is a bilateral matter" "da, stiu, este un acord bilateral incheiat dupa ocuparea cehoslovaciei" (yea, i know... post-occupation bilateral). then gorby says they can't agree in this matter and ceausescu agrees with him.
mircea_popescu: the important point for romania was that gorby wasn't going to deliver all the oil romania had contracted (and paid for). so ceausescu went to iran ; where he got ~40mn barrels with a further option, to be paid in romanian agricultural machinery.
asciilifeform: waitasec why was ro petroimporting
mircea_popescu: overgrown industrial base, romania imported a lot of energy exported a lot of high tech stuff.
asciilifeform: ah hm.
mircea_popescu: kinda the model since comunists came to their senses post ww2.
mircea_popescu: before that, romania exported oil. after that, imported about 50% of consumption.
mircea_popescu: and of course, "Alan Green, named ambassador to Romania by President George Bush [who called him "a good friend", "intransigent", "well introduced to my take on freedom and democracy"], died Friday in his home in Palm Springs, Calif. He was 75. Mr. Bush selected Mr. Green in 1989, and he moved into the American Embassy in Bucharest just two weeks before the dictator, Nicolae Ceausescu, was executed."
mircea_popescu: (state dept distinguished honor award recipient)
mircea_popescu: and in entirely other lulz : http://trilema.com/wp-content/uploads/2011/05/jsexnetrwork-soft-on-demand-s.jpg the exhibitionists are those girls who make the obviously deliberate effort of not covering their cunt with their hands.
asciilifeform: lolwassat
danielpbarron: !!v A9B8E39BDFDB6F827D07E95D156AD3BBE11A4165A89EC9FF957F61CDE4C47DF0
deedbot: danielpbarron rated cruciform 1 << bought a couple FUCKGOATS from me
asciilifeform: mircea_popescu: for some reason i can't help but think of the old lul with chinese on footstools synchronously jumping
mircea_popescu: random bucnha naked azn girls.
danielpbarron: http://btcbase.org/log/2017-10-07#1722406 << it is in the files.tar.gz linked on my blog, and deeded here http://deedbot.org/deed-482547-1.txt☝︎
a111: Logged on 2017-10-07 22:12 ben_vulpes: danielpbarron: wouldja mind sharing that stage3 you build your eulora gentoos with?
asciilifeform: well yes but loox like intends to be a coherentwave of gurlz rather than randopolarized, if you will.
mircea_popescu: i don't think such a thing as randfomly polarized female wave ever existed or ever could exist.
mircea_popescu: females are naturally-coherent.
mircea_popescu: even to the degree of http://trilema.com/2017/zuleika-dobson-or-an-proper-love-story/#selection-3413.588-3413.746 ; or rather... starting with.
mircea_popescu: and in other lulz, http://boards.4chan.org/biz/thread/3790332
danielpbarron: !!deposit 1
danielpbarron: what am i doing wrong?
danielpbarron: !!balance
mircea_popescu: danielpbarron should work.
asciilifeform: mircea_popescu: 404 eggog ?
mircea_popescu: aww ? they beleeted it already ?
mircea_popescu: look at that, and before the archival bot kicked in, even.
mircea_popescu: I GUESS THE WORLD WILL NEVER KNOW.
danielpbarron: mircea_popescu, i get no response from the bot
mircea_popescu: odd. mayne trinque has something to say
phf: !!deposit 0
phf: i was going to say maybe it's hardcoded to mircea_popescu, but lobbes was using it too
asciilifeform: mircea_popescu: what was it
trinque: nope, just something amiss. one sec.
trinque: danielpbarron: give it another try
trinque: today I learn autossh will sometimes exit wtf
trinque: thought it *was* a retry loop for ssh
trinque: !!balance
deedbot: http://p.bvulpes.com/pastes/IkZcb/?raw=true
phf: trinque: i'll fix http://btcbase.org/log/2017-10-05#1721087 in a bit. obviously such a change will require deedbot to not ever quote arbitrary log lines in channel☝︎
a111: Logged on 2017-10-05 16:19 trinque: hm no a111 quote?
trinque: yep, it wont.
danielpbarron: !!deposit 1
deedbot: Get your OTP: http://p.bvulpes.com/pastes/WYvmi/?raw=true
danielpbarron: !!v C6941C8796AD7EBE588C4C2A97DFFD5B3BDF2B2553A01E66AD2F210E28B2C45B
deedbot: http://p.bvulpes.com/pastes/mpjJR/?raw=true
danielpbarron: !!balance
deedbot: http://p.bvulpes.com/pastes/ojr7G/?raw=true
trinque: isn't reflected until I actually credit the account
trinque: will do after a few blox
danielpbarron: why is the deposit operation a 2-part thing? couldn't it just encrypt an address to my key with the amount to send?
trinque: it's already in the logs.
trinque: that'll change, but as part of limiting the thing to this channel only
danielpbarron: ah ok, i did read that part
trinque: plenty of improvements left to make, for sure.
asciilifeform: !!up apeloyee
deedbot: apeloyee voiced for 30 minutes.
apeloyee: thanks
apeloyee: http://btcbase.org/log/2017-10-07#1722395 << why do karatsuba when you can just shift and add them, like in your W_Mul☝︎
a111: Logged on 2017-10-07 21:14 apeloyee: http://btcbase.org/log/2017-10-07#1722289 << and the point of doing karatsuba is? you do 2 recursive calls to Mul_Karatsuba_TopOnly and one to Mul_Karatsuba. should've simply calculated upper_part(XLo*YHi), upper_part(YLo*XHi) and XHi*YHi
apeloyee: I mean, W_Mul doesn't do karatsuba
asciilifeform: describe in detail ?
apeloyee: compute the higher part of product X*Y as XHi*YHi+ShiftRight(XLo*YHi+XHi*YLo, K), where K is size of XLo and YLo
apeloyee: no computation of |XHi-XLo|
apeloyee: are you disappointed by the savings of computing just the higher part yet?
asciilifeform: not yet, considering that it dun work yet, lol
apeloyee: see, it does three recursive calls, meaning the speedup is wholly dependent on the speedup of comba for half-multiply
apeloyee: three, just like normal caratsuba mul
apeloyee: and most products for which the comba is called, are full products, not half products
apeloyee: 2 half products out of 3 on the first level of recursion, 4 of 9 on second, and 8 of 27 on third, assuming 64-bit words and unrealistic 2-fold speedup of comba for half-multiply, and no overhead in karatsuba,
apeloyee: !~calc 4/27
jhvh1: apeloyee: 4/27 = 0.14814814814814814
apeloyee: 15% speedup
apeloyee: for 64-bit words, 4096-bit operands
apeloyee: the best case pretty much.
asciilifeform: currently i'm aiming for <1sec (opteron 3GHz) 4096b modexp, with minimal new moving parts. after that -- releasing.
apeloyee: which pretty much resolves http://btcbase.org/log/2017-10-07#1722070☝︎
a111: Logged on 2017-10-07 00:38 asciilifeform: mod6: you will notice that the barrett in 'crc handbook' is more complicated : it shrinks the x and then compensates later. this relies on normalization , and constanttimeized incarnation of it would have to work as apeloyee described ( i'ma try it much later, once i see what can be had re speed strictly from having asymmetric karatsuba instead of the current mega-waste )
apeloyee: so, are you putting off the top-half-multiply
asciilifeform: i'ma try it next
deedbot: http://btcbase.org/log/2017-10-08#1722757 << http://p.bvulpes.com/pastes/SMb9H/?raw=true
trinque: danielpbarron: ^
mircea_popescu: apeloyee i'm curious, what do you do for a living ?
apeloyee: i'ma try it next<< for less than 15% speedup (or 10% for 8192-bit operands)? converting to classical barrett should be much more productive
asciilifeform: apeloyee: my strategy so far was to introduce moving parts very, very reluctantly ( started with egyptian multiplier, for example ) when there is absolutely no choice.
asciilifeform: for instance unrolled comba wins 20-25% speed, but i did not use it in place of the generic because it is longer and harder to read.
apeloyee: what do you do for a living ?<<sit before computer, sometimes participating in writing of, er, physics papers.
asciilifeform: ( unrolled comba would have explicit unrolled cases for 1,2,...,8-word operands )
mircea_popescu: ha.
apeloyee: so top-half-multiplier isn't a moving part?
asciilifeform: it is! but much smaller than, for instance, the secretshift-barrett.
asciilifeform: which i will also make, and decide if it was worth the cost
asciilifeform: and then bernsteinian karatsuba, possibly, and whatever else i can think of.
asciilifeform: in the end might even release different variants that have different complexity tradeoffs.
danielpbarron: trinque, ty
asciilifeform: ( reader can pick which he wants )
mircea_popescu: meanwhile in wetware engine work, http://78.media.tumblr.com/99f358816497d4f932800e6fa98aeaff/tumblr_nakmetMb3y1tk1rqio1_500.gif
apeloyee: do you agree that top-half-mul wins 10% at most for 8192-bit operands?
trinque: apeloyee: try upping yourself
asciilifeform: ( a concrete example : http://wotpaste.cascadianhacker.com/pastes/bP0Qt/?raw=true vs http://wotpaste.cascadianhacker.com/pastes/YBnZR/?raw=true knuthianmod )
apeloyee: !!v 2E65968AB6D7D411ADAAA70B8E530119A400D844AC246114A50D95567C482A12
deedbot: You are now voiced in #trilema
trinque: great. obviously when not voiced, will have to pm.
asciilifeform: apeloyee: theoretically. but cache locality win from smaller memory segment sometimes gives surprising winning. the example above, for instance, gives 2x speedup rather than my predicted 25%.
mircea_popescu: ^
mircea_popescu: i was going to say, the calculated %s rarely match. but theoretically, it should be less than 0.1
apeloyee: thanks trinque
trinque: yw
asciilifeform: if ffa can be made to do 4096b modexp in 0.5s on typical comp, that gives ~1byte/msec purersa payload. which is enough for many purposes, e.g. voice.
asciilifeform: a 2sec modexp is already a wholly fine replacement for koch's gpg, say.
asciilifeform: ( karatsuba, i will note for n00bz, parallelizes , but i deliberately omitted parallelization logic because i want ffa buildable on msdos and for machines with 1 cpu )
apeloyee: http://btcbase.org/log/2017-10-07#1722397 << I was unclear. Let A be the number to be reduced mod N, R the approximate reciprocal, K the ffa bitness fitting the modulus, then we know that 0<A - N*floor(A*R/4^K) < 2*N <2^(K+1). So might as well calculate A - N*floor(A*R/4^K) modulo 2^(K+1).☝︎
a111: Logged on 2017-10-07 21:25 apeloyee: the multiply-by-approximate quotient in barrett's also needs only the lower part (plus 2 extra bits to the left), and lower part of product can be computed exactly (since rounding is not a problem)
apeloyee: i.e. the lower part.
apeloyee: modulo 2^(K+2) for classical barrett.
asciilifeform: it isn't clear to me exactly how
apeloyee: right, unclear again. the muliply of N and floor(A*R/4^K) can be calculated mod 2^(K+1)
apeloyee: *product
asciilifeform: how? k is discrete ( multiple of e.g. 64 )
apeloyee: ok, modulo 2^(K+64).
asciilifeform: all of my mult is xbit * xbit -> 2xbit . ( after using apeloyee's method, also can have xbit * xbit -> upperhalf(2xbit) )
asciilifeform: i even threw out ability to have mul operands that are not power of 2.
asciilifeform: ( karatsuba assumes length always divisible by 2 )
asciilifeform: *that are now power of 2 size
asciilifeform: *not
asciilifeform: so how do you propose to multiply anything modulo 2^(k+64) ?
apeloyee: your choice to not implement it. i've come to the conclusion that it's not worth the effort anyway.
asciilifeform: so far almost all of my theoretical predictions re which optimizations will be worth the effort, were wrong
asciilifeform: ( as in above case with knuth divider )
asciilifeform: because on pc most of the wait time is for memory access.
apeloyee: on a different topic, http://btcbase.org/log/2017-10-08#1722429 and http://btcbase.org/log/2017-10-05#1721484 seem to contradict each other. what's an initial sieve for if the algo must run in fixed time? i've interpreted it as "successful test must run in fixed time, failures can be variable-time", and make proposal accordingly.☝︎☝︎
a111: Logged on 2017-10-08 00:16 asciilifeform: the ONLY correct method of generating cryptoprimes, is to 1) get N bits from FUCKGOATS 2) determine, in fixed spacetime every single time, whether that string of bits constitutes a usable prime.
a111: Logged on 2017-10-05 19:38 asciilifeform: for the initial sieve ~prior~ to miller-rabin
asciilifeform: apeloyee: no contradiction. the variability of time is in the ~test~, not the output result , which naturally will vary depending on what rng gave you
asciilifeform: there must be no variability in the time the ~test~ takes.
asciilifeform: *variability allowed not in the test, but in output
apeloyee: so if the number fails initial sieve, do you proceed to miller-rabin?
mircea_popescu: http://wifc.com/news/articles/2017/oct/06/russia-greek-ruling-on-russian-bitcoin-fraud-suspect-is-illegal/ in other coke machine competitions.
asciilifeform: apeloyee: i see your point. either we dispense with the sieve, or decide to count from the moment after sieve.
asciilifeform: otherwise sieve is waste of time.
mircea_popescu: i don't see what the problem is, practically. so you leak ... how many times you had to try to get a prime ?
mircea_popescu: leaking rng quality is more of a concern for debian/prngs.
asciilifeform: possibly constantly, depending on the rsa keying system
asciilifeform: if we generate keys continuously, it is a problem.
mircea_popescu: hm
asciilifeform: the time taken by heathen miller-rabin , in fact leaks key bits.
asciilifeform: ( i linked to a concrete algo for this attack some months ago )
asciilifeform: rejecting rng result that doesn't pass the gcd sieve -- leaks nothing
asciilifeform: because all you do is get NEXT N bits from rng, they have ( if rng is proper ) 0 relation to previous N
mircea_popescu: right.
apeloyee: yes. same for miller-rabin
mircea_popescu: so this is more a r-m problem altogether. as that's not linear.
asciilifeform: apeloyee: no, because there is another possibility
mircea_popescu: apeloyee no, because as he well points out, the time it takes is not unrelated to the key.
asciilifeform: for miller rabin you may end up ACCEPTING the test
asciilifeform: and then if your miller-rabin is variable time, you have leaked key bits.
asciilifeform: via the interval it took you to ACCEPT.
asciilifeform: this is why ultimately entire primality test algo must be constant time, just like the other pieces.
asciilifeform: if you leak in one place, the rest of the places are worthless
apeloyee: well, I thought it's not a problem, each round of m-r can be implemented by slightly different version of extant modexp
mircea_popescu: ugh
apeloyee: leaking number of rounds is not a problem
asciilifeform: what?!
asciilifeform: no 'different versions'
asciilifeform: understand, this thing is 800 lines right now and i consider it too big.
asciilifeform: ANY LEAK IS A PROBLEM
asciilifeform: there is to be no compromise on leak.
asciilifeform: if you want 'compromise' rsa, use koch's.
mircea_popescu: the true problem here is that there's not going to be a fixtime r-m
asciilifeform: yes there is.
mircea_popescu: how ?
asciilifeform: same method as constant gcd.
asciilifeform: all steps that were previously conditional, happen muxed.
mircea_popescu: but its a probabilistic algo.
apeloyee: if you have N ffa-eligible tests, bailing early out after one of them failed is not a problem.as per above.
mircea_popescu: it naturally makes assumptions about the item you're testing.
asciilifeform: mircea_popescu: all that means is that one of the inputs comes from rng.
asciilifeform: it doesn't mean that we are forced to BRANCH on it.
mircea_popescu: hm
asciilifeform: can instead -- mux.
asciilifeform: just like i did everywhere else.
asciilifeform: apeloyee: correct.
asciilifeform: apeloyee: each individual test has to be fixedtime though. then -- yes.
asciilifeform: then all ACCEPTED primes took exactly same number of cpu clocks, to produce.
asciilifeform: mircea_popescu: review the mr algo , it is actually surprisingly easy to ffaize, just replace all 'return true' with flag := flag OR true, etc
mircea_popescu: incidentally, since we're on m-r : do we actually pick 4096 bit bases to avoid the arnault number problem ? to leverage the ffa flatness, as in http://btcbase.org/log/2017-10-07#1722376 ?☝︎
a111: Logged on 2017-10-07 19:28 asciilifeform: http://btcbase.org/log/2017-10-07#1722358 << point was exactly to compare like items. i.e. heathendom does NOT get to 'win' by 'oh hey the hamming weight of exponent is only 2, not 4096, so we only do 4 modexps and not 8192'
asciilifeform: y'mean exponents ?
mircea_popescu: Z/nZ < n factor.
asciilifeform: waiwat
apeloyee: each round of miller-rabin is mostly a modexp which makes some tests on the intermediate results. so I don't see how you can avoid a different version of modexp
mircea_popescu: asciilifeform https://www.researchgate.net/publication/220161766_Constructing_Carmichael_Numbers_which_are_Strong_Pseudoprimes_to_Several_Bases (guy named arnault gave example of number for which all tests up to ~300 were misleading)
asciilifeform: apeloyee: i don't see what is wrong with standard modexp
mircea_popescu: but we don't have to start low. and we don't really want to, either.
asciilifeform: mircea_popescu: chance of these without sabotaged rng is < chance of meteorite
asciilifeform: ( meteorite specifically on your head, not elsewhere )
mircea_popescu: depends how you run RM, doesn't it.
mircea_popescu: famously, maple misidentified the guy's number. not because of rng, eiher.
asciilifeform: item paywalled
asciilifeform: anybody got copy ?
mircea_popescu: asciilifeform https://mathcrypto.wordpress.com/2014/11/29/large-carmichael-numbers-that-are-strong-pseudoprimes-to-several-bases/
asciilifeform: ty
mircea_popescu: 2887148238050771212671429597130393991977609459279722700926516024197432303799152733116328983144639225941977803110929349655578418949441740933805615113979999421542416933972905423711002751042080134966731755152859226962916775325475044445856101949404200039904432116776619949629539250452698719329070373564032273701278453899126120309244841494728976885406024976768122077071687938121709811322297802059565867 for permanent record.
asciilifeform: i suspect that for any probabilistic test, you can construct a boojum (e.g. you know that he will do 300 rounds, you make one that needs 301 )
mircea_popescu: this is specifically "all prime bases up to 300".
mircea_popescu: because (((idiots))) start from 1.
mircea_popescu: which i assume we won't be doing ?
asciilifeform: i can't think of why to do any such thing
mircea_popescu: but maple did do it.
apeloyee: maple did a deterministic test
apeloyee: and made it wrong
mircea_popescu: (and ftr -- aleph (carmichael numbers) = aleph(N) iirc.)
mircea_popescu: apeloyee wasn't it exactly r-m restricted to first 10 primes or such ?
apeloyee: I think so.
mircea_popescu: well that's what i'm saying.
asciilifeform: and moreover for long-term key genning, imho a week or longer probabilistic primality test is not inappropriate.
mircea_popescu: "10 rounds, start from 1".
mircea_popescu: asciilifeform should prolly be user knob tbh.
asciilifeform: ( and naturally it parallelizes without any effort on all iron you might have , just set it up on each )
asciilifeform: this requirement is somewhat in tension with classical airgapism 'this key was born in this tin can, and must die in it' however
asciilifeform: so operator must decide for himself
asciilifeform: which he'd rather have -- key that he genned inside 50cent chip, staying there, or primality-torture on his fleet of pentiums etc
apeloyee: at that cost, may also do the deterministic miller test then.
mircea_popescu: ^
mircea_popescu: probably should.
mircea_popescu: and in other bucharest tax office, https://i2.wp.com/www.cronicipebune.ro/wp-content/uploads/IMG_1539.jpg
jurov: hi mircea_popescu, s.qntr is still traded? i have got some frozen mpex orders
mircea_popescu: jurov shall look into it.
asciilifeform: i'm not aware of a fully deterministic test that doesn't run in geological (e.g. saxena) time
asciilifeform: ( very often abuse of terminology, what people actually mean by 'deterministic version' is 'probabilistic with prng supplying the random' )
apeloyee: doesn't run in geological (e.g. saxena) time << if you have faith in generalized riemann hypothesis and correctness of work on deterministic miller test - you have it. I don't, but running test for a week is imo greater crackpottery than believing in that.
mircea_popescu: well, the running maybe not, but ~believing~ that it achieved something, surely.
asciilifeform: i'll take the p(failure) to the week's power, over the possibility of hypothesis falling and ALL keys fucked.
mircea_popescu: as per the ancient "doctor, random things in the house are talking to me, am i losing it ?" "have you started answering ?" "not yet" "then not yet"
asciilifeform: ( i.e. i regard the proof behind strength of the probabilistic ver, as fundamentally stronger than the other's )
mircea_popescu: asciilifeform yes but this is just the artistic side in you.
asciilifeform: and incidentally if there existed an UNBIASED constructor of primes, i'd use that
asciilifeform: and dispense with tests.
asciilifeform: but there does not.
mircea_popescu: oh and in other news
mircea_popescu: !#s simona halep
a111: 21 results for "simona halep", http://btcbase.org/log-search?q=simona%20halep
mircea_popescu: ^ now #1.
asciilifeform: unbiased -- in this case -- would mean that it eats ANY bitstring from rng, R, and maps it to UNIQUE prime , P
mircea_popescu: no, i get it. you want a prime index function.
asciilifeform: where P has same bitness as R.
mircea_popescu: but function, not hash table
asciilifeform: correct.
mircea_popescu: this incidentally is a fine statement of what a prime even fucking is.
mircea_popescu: "not a root of 1st degree polynomial with smaller parameters than it"
asciilifeform: apeloyee: i don't actually see how 'test for a week' is crackpottery when speaking about a key that is intended to stand up for 50 years ( or longer )
asciilifeform: when submarine is built, meant to last maybe 20yrs, test takes much longer than week
asciilifeform: and costs considerably more
apeloyee: pray that GNFS will never be improved upon!!
asciilifeform: whereas some keys are more valuable than any submarine
mircea_popescu: asciilifeform but the test that takes longer and costs more does not consist of manic re-measuring of the same one length, repeated millions of times.
asciilifeform: apeloyee: why not also say 'pray that p != np '
asciilifeform: mircea_popescu: if ~probabilistic~, not 'same' test
apeloyee: asciilifeform: yes, that too.
mircea_popescu: if your expectation is that the fifth attempt did not resolve the problem in a manner such as the fifth million would, there's deeper problems.
asciilifeform: mircea_popescu: this is demonstrably true re r-m test tho.
asciilifeform: whole idea of probabilistic algo.
mircea_popescu: apeloyee the p/np thing is kinda the label used here for all these, zfc, gnfs, etc.
mircea_popescu: asciilifeform convergence on it is much narrower than mn tries or w/e a week provides.
asciilifeform: ^ to rephrase, we don't actually know if hard problems exist as a hard law of nature.
asciilifeform: it remains possible that -- somehow -- they do not
asciilifeform: but in light of this, a correct rsatron is still one that stands on nothing BUT the assumption that rsa is hard.
mircea_popescu: well, so defined it must gcd as primality test.
asciilifeform: rather than, e.g., 'rsa broken OR aes broken OR prng broke OR riemann is false OR ...'
asciilifeform: mircea_popescu: if it were physically possible as the sole primality test, we'd all use.
mircea_popescu: i know no proof of r-m convergence in terms of factorization.
apeloyee: hey, before quadratic sieve was invented, they used to say that breaking 512-bit rsa will take eleventy zillion years and it's therefore Totally Secure (tm)
mircea_popescu: aha. and even after it was invented and before it was published.
asciilifeform: fall of rsa is roughly same item as 'global warming'
asciilifeform: in that it may be an actual problem , but NONE of the folx who ever publicly discussed it, have any business being taken seriously.
asciilifeform: all 'selling something'.
asciilifeform: and all of what they're selling -- stinks.
apeloyee: why should then "ras key for 50n years" should be taken seriously then?
apeloyee: *rsa key for 50 years
asciilifeform: because it is the null hypothesis.
asciilifeform: you get choice between 1) rsa 2) public key crypto does not exist
asciilifeform: people offering a (3) are sellingsomething.
mircea_popescu: then 2 seems the null hypothesis ?
asciilifeform: problem is that the historical period where crypto was a contest of bullet vs armour, rather than 'absolute bullet exists'/'absolute armour exists' is not over.
asciilifeform: situation where rsa is breakable, but no one can yet break it, makes it the sane option . because alternative is to become a donkey fucker ( rely on face to face for all comms , hope that nobody invents listening bug, etc )
apeloyee: anyway, I was saying that, if spending a week, may spend a small fraction of the time on the supposed-deterministic test
asciilifeform: why not both ?
apeloyee: of course boith
apeloyee: *both
asciilifeform: when i say 'week' it does not mean on a particular test.
asciilifeform: can be on any combination of whatever known tests.
apeloyee: http://btcbase.org/log/2017-10-07#1722400 << hey, I offered you an idea for GCD. you: "it stinks". I point you to bernstein ( https://facthacks.cr.yp.to/batchtrial.html ). you: "it stinks". maybe GCD is not a sane option ofter all, eh?☝︎
a111: Logged on 2017-10-07 21:28 apeloyee: http://btcbase.org/log/2017-10-05#1721485 << i thought bernstein's "how to find smooth parts of integers" suggests a remainder tree, not gcd?
asciilifeform: gcd is for the pre-mr sieve, is all.
asciilifeform: ( and potentially for other primality tests, though i can think of some cryptosystems where it is handy )
asciilifeform: remember that ffa is not strictly for rsa.
asciilifeform: apeloyee: what does remainder tree win when you are testing only 2 numbers ?
asciilifeform: i used bernstein's tree in phuctor, where it made actual sense
apeloyee: what two numbers?
asciilifeform: x and y.
asciilifeform: ( in our concrete case, r, a random , and p, a primorial -- for the pre-mr litmus test )
asciilifeform: if gcd(r, p) == 1 -- then worth m-r, otherwise not )
apeloyee: why p, and not p1, p2, ... p419?
asciilifeform: because i'm on a chip with 32kB of memory, say.
asciilifeform: why the hell should i keep random crud in a table to pick up later.
asciilifeform: but hypothetically it may even be possible to ffaize bernstein's tree. or even to do it in such a way that doesn't wipe out the cpu winning from it. and even possibly to prove that it works and doesn't leak bits and doesn't let composites through once in a while.
apeloyee: because the acceptably fast algorithms are simpler.
asciilifeform: but this would weigh more than all of ffa to date !
asciilifeform: when you ffaize 'simpler' is not always what initially looks like .
apeloyee: bernstein's batch trial division would seem to straightforwardly ffaize. where's the problem?
asciilifeform: the fact that i don't need the batch aspect for anything, for starters
asciilifeform: the fact that divisions are dog slow, for seconds
apeloyee: the fact that i don't need the batch aspect for anything, for starters << so don't.
apeloyee: the fact that divisions are dog slow, for seconds << what barrett's reduction is for.
asciilifeform: in practice on pc speed appears to be inversely proportional to memory used, rather than the cpu cycle count.
asciilifeform: and gcd wins vs however-many trial divisions with barrett.
asciilifeform: (i.e. you still win if you take 500x the cpu cycles, so long as you don't get cache-evicted)
asciilifeform: barrett needs large scratch buffer for the mults; gcd can happen in-place.
apeloyee: but you said you was wrong in estimating performance every time!! :P
asciilifeform: lol
mircea_popescu: ahaha
asciilifeform: and incidentally i was not joking when said 32kb, it is fully my intention to eventually put whole thing on fpga where there will be certainly not even half MB of working space.
asciilifeform: not even quarter mb.
apeloyee: I think http://btcbase.org/log/2017-10-07#1722402 will win over quadratic gcd.☝︎
a111: Logged on 2017-10-07 21:48 apeloyee: http://btcbase.org/log/2017-10-05#1721485 << alternatively, can *construct* numbers which don't have very small factors. pick a nonzero remainder mod 2, mod 3, ... mod largest-prime-fit-in-your-primorial and find what number of primorial is congruent to it using chinese remainder theorem
asciilifeform: not if it needs 8x more temp space
asciilifeform: and doesn't even stand a chance of fitting in fpga.
asciilifeform: ( and probably not even on pc, where it will get thrown out of cache )
asciilifeform: but you're welcome to try.
apeloyee: on fpga, can bit-pack.
asciilifeform: elaborate?
apeloyee: pack p1, p2, ... pn so it doesn't take much more space, than the product thereof.
asciilifeform: how??
asciilifeform: we're still doing constant times.
apeloyee: what, the small primes are a secret?
asciilifeform: understand that in fpga 'secret shift' is NOT a function that can be 'called', but a physical object that gets instantiated, using thousands of cells, every time you use it.
asciilifeform: same as adder, etc.
apeloyee: oh noes, the enemy will learn that the first primes are 2,3,5,7,11,..!!!
asciilifeform: ( i.e. every time you write down a '+' that's a minimum of 2N LUTs used up, that cannot be used for anything else, where N is the operand width )
asciilifeform: apeloyee: no compromises.
asciilifeform: on leak.
asciilifeform: no camel's nose in the tent.
asciilifeform: who wants to fuck camels, can go use openssl.
apeloyee: why not lak waht's already public: the list of small primes. that' not a compromise.
asciilifeform: because go and prove that it is ALL you leaked.
asciilifeform: not to mention that we do not have such a thing as a small integer.
asciilifeform: we only have gigantic integers, and single bits. that's it.
asciilifeform bbl
apeloyee: as usual, if primitives don't leak- this won't leak anything but the small primes. this is obvious. or are we doing associativity proof in peano arithmetic or other hair-shirt arcana (c)
apeloyee: not if it needs 8x more temp space << and karatsuba totally doesn't.
apeloyee: you trade temp space for performance.
apeloyee: not 8x, but nether does proposed method
apeloyee: nether does proposed method require 8x
apeloyee: *neither
apeloyee: your fpga impl will necessarily have multipliers for various sizes of int
apeloyee: can reuse that
apeloyee: multipliers for mod-exp
mircea_popescu: i expect you just get one mult for the largest size and reuse it indefinitely, nfi what he's on about with the 2n luts for every sum
apeloyee: yes, that would fit no fpga one can buy
mircea_popescu: and in continuing funz, http://78.media.tumblr.com/8292708d5f10c32dd03e840cbf38b52f/tumblr_msoh7wWpev1snc48co1_400.gif
mircea_popescu: http://www.beatmort.ro/aberatii/internalnote.php << in ancient lulz, rather related to http://btcbase.org/log/2017-08-01#1691973 ; this was genuinely a major piece of ro internet culture late 90s, when ustards started to "invest"/move in.☝︎
a111: Logged on 2017-08-01 17:22 asciilifeform: http://btcbase.org/log/2017-08-01#1691747 << this almost suffices to qualify as crane operator!111
mircea_popescu: sadly, nobody had the foresight to fuck them with a spiked pole then.
mircea_popescu: samovar seemed harmless.
hanbot: nor should you ever substitute "May I help you?" with "ce pula mea vrei"? << ahahaha
mircea_popescu: pai nu ?
hanbot has doubts adhering to that list of 'rules' is even possible for most speakers
mircea_popescu: from what i hear they've a new model now.
mircea_popescu: they adhere to anything.
hanbot: at least the walls'll say muie long after the mouths've stopped
mircea_popescu: sooo... just had a lightning so close, made the landline phone ring.
shinohai: "Hey MP, this is Thor .... what's shakin?"
asciilifeform: woah landline
asciilifeform: they still got'em in mircea_popescustan?!
asciilifeform: http://btcbase.org/log/2017-10-08#1723058 << which op is it that only leaks 1 of its 2 operands ?☝︎
a111: Logged on 2017-10-08 20:12 apeloyee: why not lak waht's already public: the list of small primes. that' not a compromise.
asciilifeform: http://btcbase.org/log/2017-10-08#1723064 << gcd does not need a karatsuba, the karatsubatron can be doing something else while gcd happens☝︎
a111: Logged on 2017-10-08 20:43 apeloyee: not if it needs 8x more temp space << and karatsuba totally doesn't.
asciilifeform: http://btcbase.org/log/2017-10-08#1723072 << mul is not add, can't do this, the wider register costs you nonlinearly moar time☝︎
a111: Logged on 2017-10-08 20:57 mircea_popescu: i expect you just get one mult for the largest size and reuse it indefinitely, nfi what he's on about with the 2n luts for every sum
asciilifeform: and large multron is made of a quantity of half-sized ones, they - half-sized again, and so on.
asciilifeform: is all you get. ( see current ffa src, it is illustrative )
asciilifeform: http://btcbase.org/log/2017-10-08#1722777 << this btw does NOT work☝︎
a111: Logged on 2017-10-08 18:17 apeloyee: compute the higher part of product X*Y as XHi*YHi+ShiftRight(XLo*YHi+XHi*YLo, K), where K is size of XLo and YLo
asciilifeform: if it worked, i'd use it in W_Mul -- but it does NOT
asciilifeform: there is a missing term
asciilifeform: http://wotpaste.cascadianhacker.com/pastes/lXrJ8/?raw=true << W_Mul, for reference.
asciilifeform: ^ represented by 'CL' in above.
asciilifeform: or hm, there's gotta be a mistake in my test, bbl
asciilifeform: now, good noose is that it in fact worx...
asciilifeform: bad is that apeloyee was not only right re 'no more than 10%' but in fact there is NO observed speedup:
asciilifeform: http://wotpaste.cascadianhacker.com/pastes/MceAy/?raw=true .
asciilifeform: same ~2sec.
asciilifeform: ( conceivably on a box with smaller cache, there could be speedup. )
asciilifeform: though i suspect that this top-only multer oughta recurse...
asciilifeform: ( notice, we only use the bottom halves of XHi*YLo and XLo*YHi )
asciilifeform: grr, top, halves
asciilifeform: ...top halves of XHi*YLo and XLo*YHi.
asciilifeform makes a http://wotpaste.cascadianhacker.com/pastes/bnsZv/?raw=true and then notices that it not only is always off by smallint, but there is no way to fix, because the carries from the ( not used for anything else ) bottom halves of the muls, now are lost .
asciilifeform: because what sucks moar than 'ugly' ? how about 'ugly AND wrong answers'
asciilifeform bbl, will continue the misadventure later