log☇︎
1243 entries in 0.741s
asciilifeform: sina: http://wotpaste.cascadianhacker.com/pastes/Zy27g/?raw=true << current ffa.
a111: 40 results for "ffa", http://btcbase.org/log-search?q=ffa
asciilifeform: !#s ffa
asciilifeform: 'from 0' means that it is permissible to have an array of length 0 ( not in ffa ! but in ada )
asciilifeform: mod6: look at ffa.ads
asciilifeform: ( see ffa.ads )
a111: Logged on 2017-06-21 17:08 asciilifeform: also gotta understand what the F in 'ffa' means. it means that N**2 takes exactly same time as N**N
ben_vulpes: asciilifeform: what does 'ffa' stand for?
asciilifeform: mod6: ffa has to stay an algebraic, if you will, whole, or you cannot prove that it works.
asciilifeform: also gotta understand what the F in 'ffa' means. it means that N**2 takes exactly same time as N**N ☟︎
mod6: <+asciilifeform> but keep in mind mod6 et al that where before you had package ALU is new FFA(Width => Width, Word => OS.Machine_Word); you will now have, e.g., package ALU is new FFA(Word => OS.Machine_Word); << thanks for this too
asciilifeform: but keep in mind mod6 et al that where before you had package ALU is new FFA(Width => Width, Word => OS.Machine_Word); you will now have, e.g., package ALU is new FFA(Word => OS.Machine_Word);
asciilifeform: ( there are not and will not be signed arithmetic in ffa )
a111: Logged on 2017-06-17 17:55 asciilifeform: immediate 'argh' because we set out to operate SOLELY with M-bit FFA.
asciilifeform: ...if all of your ffa are 'clean' powerof2-sized, you can lose the floor/ceiling thing
a111: Logged on 2017-06-17 17:55 asciilifeform: immediate 'argh' because we set out to operate SOLELY with M-bit FFA.
asciilifeform: immediate 'argh' because we set out to operate SOLELY with M-bit FFA. ☟︎☟︎
asciilifeform: re 'small', my ffa example ( linked earlier ) is illustrative. basic bignum arithm stack, works on any machine bitness or endianness
asciilifeform: the generalization of the principle is ffa
mod6: I read through ffa, and it 'makes sense' to me as far as what you're doing in most cases -- as I understand the intent of the expressions. but i stand in awe, especially after trying my own hand at Ada, at what you've produced.
mod6: i set out to do something similar to what phf had been doing by writing some tests utilizing ffa. but then i couldn't figure out how to even make a package with function or procedure that used a generic type.
asciilifeform: mod6: somewhat relatedly, chased bug in ffa for a ~week : wanted to make FZ.Z a trimmable ( e.g. length set as optional param ) so as to enable karatsubaization (which does 0 useful without being able to, for it and nothing else, constrain the lenght of FZ)
asciilifeform: in other noose, probably will have to rewrite ffa entirely
mod6: new ffa compiles
asciilifeform: http://wotpaste.cascadianhacker.com/pastes/QTnBx/?raw=true << mod6 , phf , ave1 , et al : most recent ffa
ave1: Hello, I've been reading the ffa code
mod6: anyone else playing with ffa?
mod6: sorry, don't mean to confuse with ffa
asciilifeform: this has 0 to do with the swap trick in ffa, mod6
mod6: and that's the main thing. i really care about trying to fit it in head, as opposed to "compiler" concerns at this time with ffa. however, my other little sample shits i've been trying to make are probably going to go a lot better now that i have a 'sane' environment.
a111: Logged on 2017-05-04 23:33 mod6: I totally should have just checked that the version alignment on the thing. It was weird because simple things would compile.... like really simple things. But when it came to the more complex, ffa for example, was getting weird error messages. But at least, hey, it's out there, it has some sharp edges.
a111: Logged on 2017-05-04 23:33 mod6: I totally should have just checked that the version alignment on the thing. It was weird because simple things would compile.... like really simple things. But when it came to the more complex, ffa for example, was getting weird error messages. But at least, hey, it's out there, it has some sharp edges.
mod6: I totally should have just checked that the version alignment on the thing. It was weird because simple things would compile.... like really simple things. But when it came to the more complex, ffa for example, was getting weird error messages. But at least, hey, it's out there, it has some sharp edges. ☟︎☟︎
mod6: !~later tell mircea_popescu Hi, does seem like my gnat woes were system environment related. Was able to compile alf's ffa after splitting up the file: http://p.bvulpes.com/pastes/Nl0EP/?raw=true
mod6: anyway, stan's ffa is exciting to read through.
phf: package Foo is new ffa(Width => 64*5,
phf: i'm doing 2345x1 on ffa(64*5, u64) and it's giving me overflow 1
mircea_popescu: lol is that ffa implemented in slutty what's hjer name ?
phf: ffa(64*18, Unsigned_64)
trinque: ffa
cazalla: isn't it a ffa down there in SA now?
mircea_popescu: ffa = future farmers association ?
kakobrekla: FFA model did not work out well so im trying the PPS