log
600+ entries in 0.063s
asciilifeform: phf: i aint in particular hurry ( possibly unlike mircea_popescu , who wants to emphasize the imho important larger point ). i won't even have time to properly touch the thing until finished ffa. but plz consider what i suggested today.
asciilifeform: if mircea_popescu or anybody else finds a 'clever' in ffa, plox to ping asciilifeform asap so it can be burned out with hot irons.
asciilifeform: ( user porting ffa to new irons is expected to run the litmus )
asciilifeform: i'm sadly in a pile of saecular shit atm, and on top of that behind sched in ffa. so will be slow.
asciilifeform: ( ffa is not threaded per se, but is thread-safe, dun allocate anyffing other than on stack, i.e. can be used inside a thread safely )
asciilifeform: juliankunkel: as a maths fella, you may also find 'ffa' ( asciilifeform's current item, http://www.loper-os.org/?cat=49 ) interesting, world's only sidechannelism-proof crypto lib, ~80% done
asciilifeform: sometimes abstract work is done by kamikazes who do it from principle, at the expense of sleep and walks in park, the way asciilifeform bakes e.g. ffa. but in general yes , it simply dun happen because nobody pays.
asciilifeform: mircea_popescu: fortunately i dun have any complex subexpressions in exponents anywhere in ffa notebook.
asciilifeform: i refuse to use bitmaps for the matholade in ffa, cuz retarded ( not searchable + looks like shit on 3 out of 4 people's displays no matter what you do )
mircea_popescu: "in the end" ffa is a piece of shit, because events from star date 78987.4
asciilifeform: ( compare vs http://www.loper-os.org/pub/ffa/ch12b_exp_timing.png )
diana_coman: asciilifeform, phf I've signed ch4 and ch5 of ffa: http://ossasepia.com/reference-code-shelf/#selection-555.0-593.39
asciilifeform: ( starting with ch12 all ffa tested strictly on ave1's gnat )
asciilifeform hands 100% full with ffa , presently
asciilifeform: http://btcbase.org/log/2018-12-03#1877916 << to date we've had both types of regrind ( e.g. diana_coman reground 'mpi' into 1 genesis, for use in smg ; ffa on other hand had a 'history-preserving' regrind , http://www.loper-os.org/?p=2743 ; and iirc mod6 is baking a 'history' regrind for trb ; diana_coman had 'history' regrind for eucrypt; and possibly i missed somebody in this list ) ☝︎
asciilifeform: http://btcbase.org/log/2018-11-30#1876362 << i'll admit, i kept half-expecting him to actually read a ch of ffa and go 'hmm... maybe i should throw out my c hairball and mecha-proofisms' but of course no dice. ☝︎
diana_coman: asciilifeform, phf my sig for ffa ch3: http://ossasepia.com/reference-code-shelf/#selection-513.0-513.24
asciilifeform: *ffa
asciilifeform: http://www.loper-os.org/pub/ffa/hypertext/ch13/fz_io__adb.htm#29_14 << subj. presumes existence of 'nibbles'
asciilifeform: btw , q for ffa readership, can anybody think of a way to make the digit slider routine non8bitbyte-clean ? ( beyond having ffa eat its pistol on boot if it finds itself on such machine, lol )
asciilifeform: you can browse whole ch13 at http://www.loper-os.org/pub/ffa/hypertext/ch13/ffa_calc__adb.htm , if lazy
asciilifeform: the statement is quite serious, i and some yet-unknown number of other people will literally bet arse on our grasp of ffa correctness.
asciilifeform: ffa. see esp the passage re parachute.
zx2c4: hey ffa in ada
asciilifeform: at 1 time, was used in the constant entry ( nibble inserter ) routine, then the latter was replaced with rewritten http://www.loper-os.org/pub/ffa/hypertext/ch13/fz_io__adb.htm#29_14 , nao sole remaining use is in the knuth divider.
asciilifeform: diana_coman: i'ma also note, _O_I is used strictly in fz_mod : http://www.loper-os.org/pub/ffa/hypertext/ch13/fz_divis__adb.htm#83_14 ; prolly oughta be inlined ~there~ and abolished as a global (even internally) function.
asciilifeform: i'ma add commentary/warnings when diana_coman et al point out good additional places where; but no one should live with expectation that ~all~ possible ways to break ffa by monkeying with internals, will be listed
asciilifeform: idea being , exported routines ( current set is shown in http://www.loper-os.org/pub/ffa/hypertext/ch13/ffa__ads.htm ) are to be 'safe on all electrically possible inputs' , with the exception of div0 (user is commanded to test for div0, as example in http://www.loper-os.org/pub/ffa/hypertext/ch13/ffa_calc__adb.htm#172_17 )
asciilifeform: observe that the operators of fz_shifts are not exported in ffa.ads ( ch11 unified api ) , they are strictly for internal use in the lib.
diana_coman: asciilifeform, http://btcbase.org/patches/ffa_ch3_shifts.kv#L462
diana_coman just ate ffa 3, will sign
asciilifeform: ( ffa has 0 deps outside of gnat itself and system character i/o, currently posix's )
asciilifeform at one point tried to systematically fit in head danielpbarron's theology, but it appears to be at least as complicated as ffa; broke teeth
mircea_popescu: asciilifeform myeah. though in fairness, those corner cases rarely in the ffa or ffa-lite deployed.
mircea_popescu: wouldn't it be ~nice~ if you used some kind of sane naming convention ? trb.adding-ffa.alf ? something ?
asciilifeform: ftr 1 of the many ruins that litter asciilifeform's hdds is a half-written ada ecc ( pre-ffa )
asciilifeform: mircea_popescu: could, for example, use diana_coman's. or a modified permissive ffa. or even the thing that came with gnat. but in any case would be 'sapper errs once' component, gotta be bug-free
mircea_popescu: politically, ffa is fine. practically, it might not work.
mircea_popescu: and yes, ffa majorily useful, and no, not necessarily against writing for it. but there may be a timing issue (trb that takes > minute to check block is useless)
asciilifeform: can be written from scratch ( whoever thinks that he can do it faster than asciilifeform's ffa, is welcome to try, i promise to take off hat ) or on ffa, either.
asciilifeform: if mircea_popescu gives signal that we're fucking done with the old flintlock pistols, then i'ma start welding on ffa in trb as soon as the former is battlefield-ready.
asciilifeform: ( it is certainly indexed in the place where it lives -- http://www.loper-os.org/pub/ffa/hypertext/ch13/fz_basic__adb.htm#72_13 -- but for some reason not in ffa_calc__adb... )
asciilifeform: ( e.g. in above example -- FFA_FZ_Get_Head is not linkable, and several hrs of dig failed to tell me why )
asciilifeform: still mighty useful. ( presently i'm continuing with the http://www.loper-os.org/pub/ffa/hypertext/ch13/ffa_calc__adb.htm thing, to fill that niche )
a111: Logged on 2018-11-26 02:46 asciilifeform: !Q later tell phf plox to snarf http://www.loper-os.org/?p=2822 into http://btcbase.org/patches?patchset=ffa , ty
mircea_popescu: i'm sure if someone defaces ffa series to talk about what-ustards-understood-of-foucaults-sexnonsense-half-century-later someone'll notify you.
asciilifeform: pehbot is simply a wrapper around compiled ffa_calc , currently it gives closest possible picture to what you get if you ran the proggy itself locally.
a111: Logged on 2018-11-26 02:46 asciilifeform: !Q later tell phf plox to snarf http://www.loper-os.org/?p=2822 into http://btcbase.org/patches?patchset=ffa , ty
lobbesbot: phf: Sent 11 hours and 19 minutes ago: <asciilifeform> plox to snarf http://www.loper-os.org/?p=2822 into http://btcbase.org/patches?patchset=ffa , ty
asciilifeform: !Q later tell phf plox to snarf http://www.loper-os.org/?p=2822 into http://btcbase.org/patches?patchset=ffa , ty
asciilifeform: mircea_popescu: now you know what half of ffa illustration sweat loox like !
asciilifeform: diana_coman et al: http://ossasepia.com/2018/11/24/proposed-change-to-w_borrow-ffa/comment-page-1/#comment-4500
deedbot: http://ossasepia.com/2018/11/24/proposed-change-to-w_borrow-ffa/ << Ossasepia - Proposed Change to W_Borrow (FFA)
phf: asciilifeform: a downside to your renaming the patches is that it breaks all the old btcbase links. e.g. http://btcbase.org/patches/ffa_ch1_genesis is now http://btcbase.org/patches/ffa_ch1_genesis.kv even though the content is the same
ascii_modem: http://btcbase.org/log/2018-11-23#1874276 << longer even! i have time budget to ffa XOR sand and repaint my www, not currently both tho ☝︎
lobbesbot: phf: Sent 3 minutes ago: <asciilifeform> plox to snarf http://www.loper-os.org/?p=2798 into http://btcbase.org/patches?patchset=ffa , ty
asciilifeform: !Q later tell phf plox to snarf http://www.loper-os.org/?p=2798 into http://btcbase.org/patches?patchset=ffa , ty
asciilifeform: this is the reason why asciilifeform is saving the optional asm subroutines in ffa for ~dead last~ ( per http://www.loper-os.org/?p=2735 proclamation ) rather than given in beginning of series
asciilifeform: i've added all kindsa 'new api' in ffa, but the only regrind was when we took up new hash type for vtrons
a111: Logged on 2018-11-18 20:52 diana_coman: http://btcbase.org/log/2018-11-18#1873291 - noted and added to the list though it might still take some time to get up to speed with ffa again
asciilifeform: http://btcbase.org/log/2018-11-18#1873480 << this is entirely ok, diana_coman , there is not a burning hurry ( tho i expect that anybody who intends on firing ffa on battlefield will first take the time to properly eat it ) ☝︎
asciilifeform: ( e.g. ffa genesis, is very small, but contained all of the basic fundamental moving parts for simple arithm )
a111: Logged on 2018-06-25 16:40 mircea_popescu: spyked> I think there's great benefit in the ffa chapter-based approach << that ~started with a genesis~.
a111: Logged on 2018-11-18 16:57 asciilifeform: unrelatedly, ave1 , diana_coman , mircea_popescu , phf , other folx with a gnat and 10min of free time -- asciilifeform would like to see outputs of http://www.loper-os.org/pub/ffa/thousand_muls.txt ( ch12 benchmark ) on various irons
diana_coman: http://btcbase.org/log/2018-11-18#1873291 - noted and added to the list though it might still take some time to get up to speed with ffa again ☝︎
asciilifeform: unrelatedly, ave1 , diana_coman , mircea_popescu , phf , other folx with a gnat and 10min of free time -- asciilifeform would like to see outputs of http://www.loper-os.org/pub/ffa/thousand_muls.txt ( ch12 benchmark ) on various irons
mircea_popescu: http://www.loper-os.org/pub/ffa/ch12_mul_timing.png << so beautifully linear! i would daresay this graph by itself promises corectness.
asciilifeform presently massaging ffa ch12, expects to take most of today
asciilifeform: ffa/p-tron as a dos boot disk would imho rock.
asciilifeform: ( i deliberately wrote ffa in such a way that it oughta work on under 32bit or even plain 16, supposing somehow enuff ram )
asciilifeform: it is on hold pending resolution of http://btcbase.org/log/2018-10-26#1866266 ( and is taking back seat to ffa currently ) ☝︎
asciilifeform: btw i was going through my ffa notebooks and found a margin note that was actually about this, that prolly oughta go in the l0gz : if yer 'pow' is a walk-through-storage, 1 problem is that initial state of the system gives too little material to walk. so 1 interesting answer might be to include a 'ballast', consisting of, say, the first 9 yrs of classical btc blox, 1...N, that is part of the luby'd set (as raw bytes, rather than as m
asciilifeform actually ran 2 printers into the ground since started ffa
asciilifeform: ( no algo in ffa is newer than 1988 or so )
asciilifeform: mircea_popescu: kinda what ffa is
asciilifeform: i wish i could concur, but while baking ffa i had to eat a good 100GB of pdf.
asciilifeform: mod6: have you considered to get a diff heathen gig ? the kind where you dun have to go anywhere ? in '16 asciilifeform pulled this off, yes it took having to run on batteries for a spell, but was 100% worth it, it's the only reason i had strength to actually do FG, ffa, etc
asciilifeform: i was hoping to avoid baking hashing into ffa/p , but loox like it isn't escapable if we're doing oaep
diana_coman: the table with FFA patches looks great; I'll look into carving again some space to re-start on it
asciilifeform: i think just his bigendian vs littlendian speshulcases (why he has any?! ffa doesn't..! ) weigh moar than all of ffa together
asciilifeform: i'd like to be able to put a link in the 'ffa regrind' article, 'and you can press it with ~this~'
a111: Logged on 2018-11-13 14:21 asciilifeform: when i started ffa, i did not plan to bake any asm speedups at all. but there's 2 reasons to do it, eventually : one is that on e.g. x86/x64, getting the upper half of a word-sized multiplication, without asm, takes ~four~ MULs plus a buncha additions : http://www.loper-os.org/pub/ffa/hypertext/ch11/w_mul__adb.htm#95_14
asciilifeform: nao, this being said, my objective is still to give acceptably-performing, re erryday ops, ffa ~without asm~. but asmism will make a serious diff re e.g. key generation, make minutes instead of day+.
ave1: http://www.loper-os.org/pub/ffa/hypertext/ch11/w_mul__adb.htm#33_13, this is the egyptian mul?
asciilifeform: ( there is also the fact that such a simple thing as addition with carry takes not 1 ADD instruction, but an entire http://www.loper-os.org/pub/ffa/hypertext/ch11/word_ops__adb.htm#36_13 orchestra )
asciilifeform: the other is that on iron such as certain ARM ( i have not yet investigated ~which~ ) , and ppc, and certain others, there does not even exist a constant-time MUL, and one is stuck with some variant of http://www.loper-os.org/pub/ffa/hypertext/ch11/w_mul__adb.htm#33_13 -- which really begs to be asmed, is riotously inefficient
asciilifeform: when i started ffa, i did not plan to bake any asm speedups at all. but there's 2 reasons to do it, eventually : one is that on e.g. x86/x64, getting the upper half of a word-sized multiplication, without asm, takes ~four~ MULs plus a buncha additions : http://www.loper-os.org/pub/ffa/hypertext/ch11/w_mul__adb.htm#95_14
asciilifeform: ( but eventually i do want to write e.g., asmtronic MUL, for arm64 ffa )
asciilifeform: i was able to regrind ffa today, using phf's vdiff, but atm cannot yet press and confirm that it actually presses to same thing as the classical
bvt: though i have nothing against work on bignum multiplication and modexp -- but as i see it, it could be a side branch of ffa. ffa already provides a solid foundation for such algorithm exploration.
asciilifeform in the process of keccakizing ffa, hence practical q
asciilifeform: when i have the choice, i'd rather write proggy to need no such cassettes ( observe, streams are not used in ffa )
mod6: So usually that leaves 30min-1hour per night, sometimes more depending on the size. Sometimes I fall down a well in the logs if i'm looking for something specific related to any number of things: pizarro, foundation, ada, ffa, who knows.
asciilifeform: mircea_popescu: after ffa/p beta rollout, i'ma be at yer service for the next thing ( fg2? tmsr.mips ? radio ? or other, take yer pick )
asciilifeform: ( i got one coming in a week or so, tho, for ffa exhaustive tests )
asciilifeform: ^ mircea_popescu , diana_coman , other potential ffa eaters ^
asciilifeform finally finished rereading ffa, nao preparing ch12 + estimate of just-how-long-to-battlefield promised earlier
asciilifeform: phf: i prolly won't get a chance to actually put it to use in near term, hands pretty full with ffa; but would like to have it. if it's there to be had.
asciilifeform: ( it's not ~completely~ fucktarded, all of ffa is still reachable via http://www.loper-os.org/pub/ffa/hypertext/ch11/ffa__ads.htm . but still frustrating )
asciilifeform: phf: it's definitely not perfect, e.g. why the FUCK is FFA_FZ_ZeroP not clickable ?