Henry de Valence
285e57f2ff
Merge pull request #128 from hdevalence/feature/avx2-docs
...
Update docs for AVX2 backend
2018-04-08 17:16:05 -07:00
Henry de Valence
67ba201835
Update AVX2 documentation
2018-04-08 16:59:57 -07:00
Henry de Valence
6bb2c02a1f
Merge pull request #127 from hdevalence/feature/generalize-naf
...
Generalize NAF code to wider window sizes.
2018-04-08 16:39:45 -07:00
Henry de Valence
1464c4101d
change LSB to least significant bit
2018-04-08 16:29:54 -07:00
Henry de Valence
9fc5602ce7
Split AVX2 docs into markdown file
2018-04-08 15:15:23 -07:00
Henry de Valence
85dd3329fb
Merge branch 'master' into develop
2018-04-06 15:14:06 -07:00
Henry de Valence
5d1f0a4ca6
Merge branch 'release/0.16.2'
2018-04-06 15:13:53 -07:00
Henry de Valence
dc563c7415
Bump version to 0.16.2
...
Only change is better Ristretto docs
2018-04-06 15:11:08 -07:00
Henry de Valence
d2cdc878d6
Merge pull request #124 from hdevalence/feature/ristretto-docs
...
Update & expand Ristretto docs
2018-04-06 15:07:58 -07:00
Henry de Valence
68bbd1bd03
Document algorithm for NAFs
2018-04-06 12:11:59 -07:00
Henry de Valence
6b768c2a1a
Change AVX2 backend to use width-8 tables
2018-04-05 16:20:47 -07:00
Henry de Valence
e30d764b7f
Add eq testing stub
2018-04-05 14:52:01 -07:00
Henry de Valence
9aa208959d
split into new Ristretto Group section
2018-04-05 14:52:01 -07:00
Henry de Valence
dbca373639
Promote headers
2018-04-05 14:52:01 -07:00
Henry de Valence
f446ea82e0
Explicit decoding formulas
2018-04-05 14:52:01 -07:00
Henry de Valence
35c18cfc51
Stub out remaining sections, add overview of batching
2018-04-05 14:52:01 -07:00
Henry de Valence
b713274890
Describe encoding better and more generally
2018-04-05 14:52:01 -07:00
Henry de Valence
5385399854
Use Q_4 instead of T_4
...
T_4 is nicer because it sounds like "torsion", but T is also used as a variable in ext. coordinates, so Q is better overall since it avoids conflicts later.
2018-04-05 14:52:01 -07:00
Henry de Valence
34a6de83e0
Split out torquing section to cover cof=4 case better
2018-04-05 14:52:01 -07:00
Henry de Valence
fbee5e252c
Split explanation of encoding formulas
2018-04-05 14:52:01 -07:00
Henry de Valence
25024a63bc
Add dual isogeny
2018-04-05 14:52:01 -07:00
Henry de Valence
3f68349480
Fix whitespace characters
2018-04-05 14:52:01 -07:00
Henry de Valence
862dd33695
Add link to curve models
2018-04-05 14:52:01 -07:00
Henry de Valence
83c3d97651
clarify wording
2018-04-05 14:52:01 -07:00
Henry de Valence
c2d409788a
Add links, tweak intro
2018-04-05 14:52:01 -07:00
Henry de Valence
ff787c2f12
Clarify Ristretto / Decaf relation in user docs
2018-04-05 14:52:00 -07:00
Henry de Valence
c99bd62b25
Add example to RistrettoBasepointTable
2018-04-05 14:52:00 -07:00
Henry de Valence
46048e451b
Clarify wording on canonical scalars
2018-04-05 14:52:00 -07:00
Henry de Valence
35e1b07e72
Clarify abstraction layers in Ristretto intro
2018-04-05 14:52:00 -07:00
Henry de Valence
c841998b07
Extract Ristretto notes into a markdown file
2018-04-05 14:52:00 -07:00
Henry de Valence
ef0dae241a
Add an explanatory note that's built on stable
...
When building on stable, the README.md is not included in the documentation,
leaving a bare entry. This adds a warning stub, pointing people to use nightly
rust.
2018-04-05 14:52:00 -07:00
Henry de Valence
51f9823c91
Create a new docs/ subtree and move assets into it
2018-04-05 14:52:00 -07:00
Henry & Isis
62d43752df
Add NafLookupTable8 and use for pre-computed basepoint table generation.
2018-04-05 05:12:13 +00:00
Henry & Isis
753a0292de
Rename OddLookupTable to NafLookupTable5.
...
An OddLookupTable corresponds to a non-adjacent form of width 5.
2018-04-05 04:48:28 +00:00
Henry de Valence
7e0ddf6b98
Rewrite NAF code to work with more window sizes
...
Change Scalar::non_adjacent_form() to take a width parameter.
This rewrite also makes it faster, although it's probably a ways off
from optimal. I don't know how much it matters.
TODO: write up description of why this computes the same thing.
Thanks to @oleganza for pointing out an error reading bits across words
in an earlier version of this code.
2018-04-04 21:13:09 -07:00
Henry de Valence
ece4bd715a
Merge branch 'master' into develop
2018-04-04 15:35:16 -07:00
Henry de Valence
bf41bde917
Merge branch 'release/0.16.1'
2018-04-04 15:34:40 -07:00
Henry de Valence
207fbb640e
Bump version to 0.16.1
2018-04-04 15:30:12 -07:00
Henry de Valence
05c1d073d1
Merge pull request #123 from hdevalence/feature/drop-bench-travis
...
Drop benchmarks from Travis
2018-04-04 15:23:50 -07:00
Henry de Valence
46d8f9c45d
Merge pull request #122 from hdevalence/feature/drop-stdsimd
...
Move from external `stdsimd` crate to `core::{arch, simd}` intrinsics
2018-04-04 11:06:23 -07:00
Henry de Valence
5f136fbd0c
Remove some warnings.
...
Not all of the warnings are removed, since although this code works, it still
needs a significant amount of cleanup, editing, and polish.
2018-04-04 10:25:42 -07:00
Henry de Valence
1048100cd0
Reduce benchmark runs
2018-04-04 10:16:32 -07:00
Henry de Valence
f830268128
Drop benchmarks from Travis
2018-04-04 10:00:18 -07:00
Henry de Valence
b0bda0278c
Drop the stdsimd crate in favor of core::{simd, arch}.
...
This change required some work, because the to-be-stabilized SIMD functions
don't allow non-constant `imm8`s. Previously, the `stdsimd` functions had a
constifying macro that ensured that the immediates were known. The dalek code
used this to build helper functions which would be inlined into different
places where the immediates were known. Unfortunately, since constexprs aren't
fully supported in Rust yet, this is done by a hidden compiler attribute, and
there's no way to propagate "constness".
To deal with this, some of the functions are specialized (e.g.,
`square_and_negate_D` instead of taking a mask), and others use an enum.
2018-04-03 17:58:23 -07:00
Henry de Valence
2465e0f962
Merge pull request #121 from isislovecruft/feature/montgomery-mul-variants
...
Define mul variants for MontgomeryPoints.
2018-04-03 10:06:48 -07:00
Isis Lovecruft
3281708965
Define mul variants for MontgomeryPoints.
...
This results in less changes to the x25519-dalek code to upgrade to
the latest version.
2018-04-02 21:09:41 +00:00
Isis Lovecruft
1a8cf7beed
Merge remote-tracking branch 'hdevalence/feature/refactor-scalar-mul_r1' into develop
2018-04-02 21:03:17 +00:00
Henry de Valence
4a648df713
Feature-gate multiscalar impls on alloc
2018-03-26 17:58:31 -07:00
Henry de Valence
7ef6a1e6fa
Reorganize AVX2 point code
2018-03-26 17:41:05 -07:00
Henry de Valence
e8b053b281
Remove AVX2 addition formulas
...
Only the readdition formulas are actually used by scalar multiplication, so
there's no reason to implement vectorized addition.
2018-03-26 17:41:05 -07:00
