log☇︎
54200+ entries in 0.033s
asciilifeform: thus far i've found that ffa speed is almost exact function of cpu clock (for given bitness of cpu)
mod6: *nod* I thought it was pretty speedy tbh.
asciilifeform: pretty decent, even tho broken inlining on sad gnat
mod6: I'll make some blog posts (eventually) when i go through setting up cuntoo and the musltronic GNAT from ave1.
asciilifeform: it's sorta like a 'run mile' time 'wearing gasmask'. but dunhurt, wainot, paste
mod6: Ok, well, maybe this isn't very useful then. But will throw it out here anyway. Good to know
asciilifeform as of ch11 all benchmarks posted are on standard tmsr gnat.
asciilifeform: but sure, paste. it'll be 2-3x-fold dog-slow vs proper inlinetronic tho.
mod6: Haven't had a chance to build it yet. I will do so, and a cuntoo when I get a chance.
asciilifeform: waitasec how come yer not using standard tmsr gnat ?
mod6: The one caviat about the timings is that I had to take out some of the inlines and pragmas to make it compile correctly. my GNAT 2015 version seems to have some funkyness to it -- we've actually discussed this before. Wanna just see a wotpaste?
asciilifeform: mod6: if you already took the time to benchmark, then plox to post, ty
mod6: asciilifeform: congrats btw! I've been attempting very much not to read ahead or anything, I still need to work through the chapters. I did however build all of the latest keccak vpatches in your vtree. I did some timings on my old test blox (core2duo) if you're curious.
mod6: yw, thank ~you~
lobbesbot: BingoBoingo: The operation succeeded.
BingoBoingo: !Qlater tell mod6 http://p.bvulpes.com/pastes/7mJ94/?raw=true
billymg: family is visiting today so won't be around for the rest of the evening though, back online tomorrow
billymg: in the mean time, if anyone else would like to test, the files here should work for you: http://billymg.com/downloads/mp-wp_add-footnotes-and-textselectionjs.vpatch http://billymg.com/downloads/mp-wp_add-footnotes-and-textselectionjs.vpatch.billymg.sig
a111: Logged on 2018-12-27 21:14 diana_coman: billymg, this might help perhaps too: http://ossasepia.com/2018/11/13/v-with-vtools-keccak-hashes-and-its-own-tree/
billymg: http://btcbase.org/log/2018-12-27#1883449 < quick update: using diana_coman's keccak-compatible v.pl, the vpatch i posted before pressed without issue. will write up a blog post with details on the patch, and another post with the patch for the svg file links ☝︎
asciilifeform: ( tldr : miller-rabin still king.. )
asciilifeform: https://arxiv.org/abs/1009.3956 << lol, they're a bourbaki apparently.
a111: Logged on 2017-11-07 16:36 asciilifeform: let's model the ideal prime-shitter. it would be an item that takes integer N , of whatever bitness, and produce the Nth prime ( or eggog if the Nth prime is bigger than the register bitness permitted. )
feedbot: http://bimbo.club/2018/12/the-tippecanoe-river/ << Bimbo.Club -- The Tippecanoe River
BingoBoingo: More on the riot in Malvin. This is what happens when a country likes sports, wants to compete on an elite level regionally, and can only field two teams: https://www.elobservador.com.uy/nota/imm-hara-una-denuncia-penal-por-disturbios-durante-festejo-de-navidad-en-malvin-20181228202844
asciilifeform: aanyways, teatime, bbl.
asciilifeform: ( i.e. sequence where only ADD, SUB, MUL, SHR/SHL, MOV, all the way from head to toe , a,b,c in, a^b mod c comes out )
asciilifeform: the other thing, if one is writing a proggy where only 1 ffawidth is used, it is possible to unroll all of the loops, derecursivize karatsuba, etc. and get straight line of instructions
asciilifeform: ( and it is fortunately much easier to find mistake in 5k loc , with ~one~ execution path, than in 150kloc with 900000 paths... )
asciilifeform: diana_coman: sadly i do not know how to 'guarantee perfection'. all i know how to do is to bake maximally 'fits in head' and bank that the folx here will find mistake if it turns out that i made one. ☟︎
asciilifeform: ( http://btcbase.org/log/2018-12-18#1881560 << see also. technically applies to ~errything after ch6. ) ☝︎
asciilifeform: ( where yer tryin' to save coad space )
asciilifeform: btw it is possible to dispense with FZ_LoMul , and simply discard upper chunk of ordinary karatsuba product. this costs ~10% cpuwise, and may be justified for some proggies.
diana_coman: fwiw I like very much that "pedantic" there
asciilifeform: ( hence the somewhat pedantic treatment of the subj, 2 separate proofolade articles. was going for 'measure 7777 times, cut 1ce' there. )
diana_coman: sadly though no amount of slow is any guarantee of perfection still
asciilifeform: you may find it interesting that asciilifeform's 1st shot at barrett managed to have a subtly fatal mistake in the proof, and correspondingly a proggy that failed on <0.1% of input space.
asciilifeform: this is the place where slow an' careful is the Right Thing.
diana_coman: that's precisely why the slow pace - I need a bit of fresher mind to get back and read through the rest of ffa
asciilifeform: ( the key work for diana_coman et al, from ffaology pov, is to eat the proofs )
diana_coman: asciilifeform, I'll try to get to it but more probably in January really
asciilifeform: diana_coman -- will be interesting to see benchmarks from your irons, if you get the chance.
asciilifeform: ( i.e. thing is siliconizable, in principle. )
a111: Logged on 2017-10-08 00:20 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.
asciilifeform: the other thing we get, is of course http://btcbase.org/log/2017-10-08#1722435 . ☝︎
a111: Logged on 2018-10-12 13:52 asciilifeform: ( bonus is that the closed form is not only constant time, but substantially faster on pc, nomoar branch prediction misses )
asciilifeform: ( http://btcbase.org/log/2018-10-12#1860844 << see also oldthread where the same effect ) ☝︎
asciilifeform: btw there's a 'seekrit' 30% linear speed boost if one uses unrolled combas.
diana_coman: on one hand there is what mircea_popescu says that yes, by now and given the type of rot it's actually to be expected; on the other hand it's of the sort that still sounds...unbelievable, yes
BingoBoingo: Meanwhile in local lulz, confiscated drugs liberated from the police in motorcycle snatch and grab https://www.teledoce.com/telemundo/policiales/en-san-jose-robaron-145-envoltorios-de-droga-que-habian-incautado-a-un-hombre-minutos-antes/
feedbot: http://qntra.net/2018/12/stanislav-datskovskiy-publishes-fully-constant-time-code-for-barretts-modular-reduction-as-part-of-ffa-library/ << Qntra -- Stanislav Datskovskiy Publishes Fully Constant Time Code For Barrett's Modular Reduction As Part Of FFA Library
a111: Logged on 2016-11-26 22:20 mircea_popescu: meanwhile at the slut party, http://68.media.tumblr.com/tumblr_lard124fV41qzw9y2o1_1280.jpg
asciilifeform: http://btcbase.org/log/2016-11-26#1573536 << ditto. and, come to think of it, ~all of'em. ☝︎
asciilifeform: ( if not for irc line char # cap, could even make it do 2048bit widths, rsa etc. will still fire faster than ping lag on typical day..)
asciilifeform: nuffin to write home about; but there, current.
mircea_popescu: "gotta pick something" and all that.
mircea_popescu: not even ff reached 255 yet. but anyway, i think this is exhausted
asciilifeform: my reasoning is 'must pick a number, distinguish item from the endless moron march of firefox etc'
mircea_popescu: so your reasoning was that if some bits get left out the koch is gonna eat them ?
mircea_popescu: see ? there's ~reasoning~.
asciilifeform: e.g. you gotta keep'em from linking up with neighbouring pile o'jerries, they'll encircle you and yer done
mircea_popescu: well what's the incentive ?
asciilifeform: let's say scout reported a 1k-man garrison in the point of origin. and nobody knows precisely how many hauptmann fuckface will be sending.
mircea_popescu: any details re this convoy ?
asciilifeform: let's hear mircea_popescu's logical end of the stick tho. mircea_popescu finds himself partizan commander again, in the field. has 300 men. how he decides how many to send to ambush nazi convoy ?
mircea_popescu: my objection inter alia is that this is not satisfied here ; makes us look like a bunch of retards.
mircea_popescu: is thgat it'll be applied to like things
mircea_popescu: the fundamental promise involved in the concept of number, as a scalar measure of something
asciilifeform: i'll be surprised if i haven't nailed the last comment typo by the time done adding chapter-sized knobs, tbh
a111: Logged on 2014-04-29 22:56 asciilifeform: 'Dijkstra said angrily. (He’d been visibly shaking his head through out the talk even before this outburst.) “How many bugs are we going to tolerate?” he demanded. “Seven,” Teitelman shot back.'
a111: Logged on 2017-07-18 22:58 asciilifeform: whaack is quite likely thinking of the bulk of the b00k, which consists of blockcipher liquishit which is complicated for no reason at all other than the religion where 'it is confusing to ME, author, and therefore Must Be Hard To Break'
mircea_popescu: in short : "why not say the magic number 3 ?" "because maybe i need to fix a typo, and i dun wanna lose 1 out of 3 for this, but i do wanna fix it" "so how do you know 1 in 255 is ok to lose for this ?" "um http://btcbase.org/log/2017-07-18#1686299 " ☝︎
a111: Logged on 2018-12-28 17:15 mircea_popescu: seems to me the unspoken heuristic is, "large enough so it's not meaningful [and therefore large enough to not bother] and small enough so it's not larger than some other number i thought about".
asciilifeform: http://btcbase.org/log/2018-12-28#1883605 << it is impossible to defend against this criticism, as formulated, so i won't try ☝︎
mircea_popescu: right. and such a thing as typos in those comments is possible. this entire line is the result of your ineffectual sideswiping.
mircea_popescu: not proof ; comments. like knuth wrote them.
asciilifeform: imho the haskellism where 'program carries proof' is generally a sad thing and results in soup
mircea_popescu: seems to me the unspoken heuristic is, "large enough so it's not meaningful [and therefore large enough to not bother] and small enough so it's not larger than some other number i thought about". ☟︎☟︎
asciilifeform: mircea_popescu: if someone knows how to make ffa moar 'docs is the program'-y, oughta write in and tell asciilifeform , cuz i came as close to it as i knew how.
mircea_popescu: yet our ideal is literate code, as in, docs in the program.
asciilifeform: mircea_popescu: if i v-versioned the ~docs~, we'd be at 50-ish, nao, i suppose
mircea_popescu: asciilifeform it's more than that : a) that there's such a thing as a correct program ; b) to which one can converge ; c) in a finite number of itemized steps ; d) the count therewhich can be known in advance.
asciilifeform: ( where it dun make sense to make anymoar changes without simply baking new item out of the dismembered corpse of old )
mircea_popescu: the vpatch that adds a "except if number is 5577, math doesn't work for 5577" is a VERY different patch from the vpatch that adds "here's a sentence ended by a comma turned ended by a dot instead"
asciilifeform: the declaration i want to make with 'kelvins' is that there is such a thing as a correct program.
asciilifeform: which can carry some qty of change ( i hold to the practice of keeping'em reasonably compact )
a111: Logged on 2018-12-19 17:09 mircea_popescu: natural language, however, ablates the trees for "convenience" so to speak, ie, uses commonly what's known in computing as sparse trees.
mircea_popescu: in short, putting numbers on the human mind's http://btcbase.org/log/2018-12-19#1881724 process is gonna be painful. ☝︎
asciilifeform: a patch in the context is a vpatch.
mircea_popescu: the larger problem is "what is meant by a patch". i doubt the item can survive ~even one~ major alterations. meanwhile "hey, you spelled theyre with an apostrophe" needn't be counted (but need be fixed)
asciilifeform: mircea_popescu: how does bridge constructor decide 'it should hold 50 tanks time margin of 3' and not 55 and not 45.
mircea_popescu: turns out he doesn't know ; in context probably dunbar's number is the saner approach.
asciilifeform: fwiw i gave orig vtron, 'v.py' , 100. it's on 97 (phf's revision) and when i get around to massaging it, i suppose will be 96, and i dunsee why to touch it again after that..
mircea_popescu: asciilifeform the discussion spawned from me going "wtf is the POINT of even counting to 255, nobody can fit such in head" that then merged into "how do you know ?" which resulted in the predictable "i wonder how the fuck does ~he~ know"
mircea_popescu: there's great cause to attack at dawn : morale superiority. there's also great cause to attack at night, note eg dracula's not famous for attacking the turkish camp at dawn and nearly making sultan run off.
asciilifeform: incidentally i expected mircea_popescu to say '255 is laughably generous'
mircea_popescu: looky, management is a rational activity. there's nothing wrong with hanging kids for having failed to master spanish in two weeks -- just as long as you can answer as to how you came up wih the two weeks.
asciilifeform: i.e. 'if this thing needs >255 vpatches after today, then asciilifeform is a tard and somebody else ought to do the whole fucking job over'
mircea_popescu: nature found 273.15 i suppose the same way it computed ocean level ; but we're not quite equal to nature.
mircea_popescu: i'm not even sure how that kelvin thing's supposed to work ; notably bothj knuth and mod6 attempted it, but the problem with starting with a number is that you don't know what to pick.