152900+ entries in 0.088s
mircea_popescu: is it patriotic
to leak
the dnc's self-important bullshit leading
to
the republic sinking clinton ?
shinohai: The sound of
that bell instantly alerts patriotfags and sends Cuban diplomats running, complaining of sonic attacks.
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 ?
jhvh1: lobbes:
The operation succeeded.
lobbes: !~later
tell mircea_popescu ^^ 'help sexpr' and 'help json' also working. lobbesbot has been brought up
to spec
mats: saudis join
turks in s-400 purchases
shinohai: They even bothered
to vanitygen a custom
tor addy
shinohai: TOP KEK "Anonymous access
through
tor browser"!!!!
mircea_popescu: as no 0 led or 0
terminated string will ever pass anyway
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 ?
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
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: yes, but
then would you rather 999 r-m or 995 primorial gcd and 4 r-m ?
mircea_popescu: recall diana_coman 's
trick of "multiply by 6" ? pretty much
the inverse of
the same idea.
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.
☟︎ 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.
BingoBoingo: Well, he works in
the retail industry. What should he expect?
mats: he put it in almost four months in advance and still can’t
take a few days off
mats: l0l an amzn frontend engineer friend has
to work all
through christmas week, got his vacation request denied by upper mgmt
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
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
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
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 )
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 2017-10-07 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)
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
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.
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
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'
phf: a whole new bignum
that is
a111: Logged on 2017-10-07 19:30 asciilifeform: i also suspect
that
they are in fact slower for maxhammingweight case of exponentiation and modulus, vs ffa.
ben_vulpes: danielpbarron: wouldja mind sharing
that stage3 you build your eulora gentoos with?
☟︎ 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
☟︎☟︎ a111: Logged on 2017-10-05 19:38 asciilifeform: want
to gcd(candidate, biggestprimorialthatfitsintheffabitness)
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 2017-10-05 19:38 asciilifeform: want
to gcd(candidate, biggestprimorialthatfitsintheffabitness)
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: 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.
☟︎ jhvh1: apeloyee:
The operation succeeded.
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'
a111: Logged on 2017-10-07 16:26 phf: asciilifeform: wait,
that seems like a cheap sleight of hand. obviously increasing number of iterations in an iterative algorithm
that you gave is going
to increase run
time
a111: Logged on 2017-10-07 16:49 mircea_popescu: my guess is
that it's as close
to closed form solutions as possible, hence all
the barrett fucking etc, but
then again i'm a weak programmer and a very dubious mathematician.
mircea_popescu: my guess is
that it's as close
to closed form solutions as possible, hence all
the barrett fucking etc, but
then again i'm a weak programmer and a very dubious mathematician.
☟︎ phf: i'm
trying
to figure it out from first principles :) (i haven't had
time
to look at
the recent, i.e. past month, versions yet)