aboutsummaryrefslogtreecommitdiff
path: root/optimizations.md
diff options
context:
space:
mode:
authorGravatar jadep <jade.philipoom@gmail.com>2017-01-09 11:14:43 -0500
committerGravatar jadep <jade.philipoom@gmail.com>2017-01-09 11:14:43 -0500
commit71dfb8a5d0677736f496f82af0b8dcbe8cd5c22c (patch)
treec9966fb8c545f33915d8f1f760c450139ddf64cd /optimizations.md
parent0b9634036c1b0f17fe9c21cd4255010dfc56e1bb (diff)
optimizations.md: formatting fix
Diffstat (limited to 'optimizations.md')
-rw-r--r--optimizations.md2
1 files changed, 1 insertions, 1 deletions
diff --git a/optimizations.md b/optimizations.md
index 2e600cfc6..20ac9aaf7 100644
--- a/optimizations.md
+++ b/optimizations.md
@@ -8,4 +8,4 @@
| Field Arithmetic | Precomputed Tables | Not Implemented | n/a, TODO | |
| Elliptic Curve Points | Extended Coordinates | Implemented | [ExtendedCoordinates.v#L258](https://github.com/mit-plv/fiat-crypto/blob/master/src/CompleteEdwardsCurve/ExtendedCoordinates.v#L258) | http://hyperelliptic.org/EFD/g1p/auto-edwards.html |
| Elliptic Curve Points | Point Compression | Implemented | [PointEncodingPre.v#L313](https://github.com/mit-plv/fiat-crypto/blob/master/src/Encoding/PointEncodingPre.v#L313) and [PointEncodingPre.v#L412](https://github.com/mit-plv/fiat-crypto/blob/master/src/Encoding/PointEncodingPre.v#L412) | Instead of transmitting $(x,y)$ to transmit a point, transmit $y$ and a bit representing the sign of $x$. Decode $x$ by solving the curve equation for $x^2$, taking the square root, and picking the square root with the appropriate sign bit |
-| Low-Level | Use Varied-size Registers, Half-Implemented | [MapCastWithCastOp.v#L116](https://github.com/mit-plv/fiat-crypto/blob/master/src/Reflection/MapCastWithCastOp.v#L116) | Rather than using the largest available integer size (e.g., `uint32_t` on x86_32, `uint64_t` on x86_64) for all operations, pick the smallest integer size which is guaranteed to fit the result for each arithmetic operation separately | |
+| Low-Level | Use Varied-size Registers| Half-Implemented | [MapCastWithCastOp.v#L116](https://github.com/mit-plv/fiat-crypto/blob/master/src/Reflection/MapCastWithCastOp.v#L116) | Rather than using the largest available integer size (e.g., `uint32_t` on x86_32, `uint64_t` on x86_64) for all operations, pick the smallest integer size which is guaranteed to fit the result for each arithmetic operation separately | |