bigint: refactor BigUint ops

As much as possible, ops are forwarded to val-ref so the LHS memory can
be reused for the result.  This reduces the number of clones required.
This commit is contained in:
Josh Stone 2015-11-16 10:56:30 -08:00
parent 7781256041
commit ce3d375b21
1 changed files with 216 additions and 97 deletions

View File

@ -86,7 +86,6 @@ pub type BigDigit = u32;
pub type DoubleBigDigit = u64;
pub const ZERO_BIG_DIGIT: BigDigit = 0;
static ZERO_VEC: [BigDigit; 1] = [ZERO_BIG_DIGIT];
#[allow(non_snake_case)]
pub mod big_digit {
@ -237,7 +236,26 @@ macro_rules! forward_val_val_binop {
#[inline]
fn $method(self, other: $res) -> $res {
(&self).$method(&other)
// forward to val-ref
$imp::$method(self, &other)
}
}
}
}
macro_rules! forward_val_val_binop_commutative {
(impl $imp:ident for $res:ty, $method:ident) => {
impl $imp<$res> for $res {
type Output = $res;
#[inline]
fn $method(self, other: $res) -> $res {
// forward to val-ref, with the larger capacity as val
if self.data.capacity() >= other.data.capacity() {
$imp::$method(self, &other)
} else {
$imp::$method(other, &self)
}
}
}
}
@ -250,7 +268,22 @@ macro_rules! forward_ref_val_binop {
#[inline]
fn $method(self, other: $res) -> $res {
self.$method(&other)
// forward to ref-ref
$imp::$method(self, &other)
}
}
}
}
macro_rules! forward_ref_val_binop_commutative {
(impl $imp:ident for $res:ty, $method:ident) => {
impl<'a> $imp<$res> for &'a $res {
type Output = $res;
#[inline]
fn $method(self, other: $res) -> $res {
// reverse, forward to val-ref
$imp::$method(other, self)
}
}
}
@ -263,58 +296,121 @@ macro_rules! forward_val_ref_binop {
#[inline]
fn $method(self, other: &$res) -> $res {
(&self).$method(other)
// forward to ref-ref
$imp::$method(&self, other)
}
}
}
}
macro_rules! forward_all_binop {
macro_rules! forward_ref_ref_binop {
(impl $imp:ident for $res:ty, $method:ident) => {
impl<'a, 'b> $imp<&'b $res> for &'a $res {
type Output = $res;
#[inline]
fn $method(self, other: &$res) -> $res {
// forward to val-ref
$imp::$method(self.clone(), other)
}
}
}
}
macro_rules! forward_ref_ref_binop_commutative {
(impl $imp:ident for $res:ty, $method:ident) => {
impl<'a, 'b> $imp<&'b $res> for &'a $res {
type Output = $res;
#[inline]
fn $method(self, other: &$res) -> $res {
// forward to val-ref, choosing the larger to clone
if self.data.len() >= other.data.len() {
$imp::$method(self.clone(), other)
} else {
$imp::$method(other.clone(), self)
}
}
}
}
}
// Forward everything to ref-ref, when reusing storage is not helpful
macro_rules! forward_all_binop_to_ref_ref {
(impl $imp:ident for $res:ty, $method:ident) => {
forward_val_val_binop!(impl $imp for $res, $method);
forward_ref_val_binop!(impl $imp for $res, $method);
forward_val_ref_binop!(impl $imp for $res, $method);
forward_ref_val_binop!(impl $imp for $res, $method);
};
}
forward_all_binop!(impl BitAnd for BigUint, bitand);
// Forward everything to val-ref, so LHS storage can be reused
macro_rules! forward_all_binop_to_val_ref {
(impl $imp:ident for $res:ty, $method:ident) => {
forward_val_val_binop!(impl $imp for $res, $method);
forward_ref_val_binop!(impl $imp for $res, $method);
forward_ref_ref_binop!(impl $imp for $res, $method);
};
}
impl<'a, 'b> BitAnd<&'b BigUint> for &'a BigUint {
// Forward everything to val-ref, commutatively, so either LHS or RHS storage can be reused
macro_rules! forward_all_binop_to_val_ref_commutative {
(impl $imp:ident for $res:ty, $method:ident) => {
forward_val_val_binop_commutative!(impl $imp for $res, $method);
forward_ref_val_binop_commutative!(impl $imp for $res, $method);
forward_ref_ref_binop_commutative!(impl $imp for $res, $method);
};
}
forward_all_binop_to_val_ref_commutative!(impl BitAnd for BigUint, bitand);
impl<'a> BitAnd<&'a BigUint> for BigUint {
type Output = BigUint;
#[inline]
fn bitand(self, other: &BigUint) -> BigUint {
BigUint::new(self.data.iter().zip(other.data.iter()).map(|(ai, bi)| *ai & *bi).collect())
let mut data = self.data;
for (ai, &bi) in data.iter_mut().zip(other.data.iter()) {
*ai &= bi;
}
data.truncate(other.data.len());
BigUint::new(data)
}
}
forward_all_binop!(impl BitOr for BigUint, bitor);
forward_all_binop_to_val_ref_commutative!(impl BitOr for BigUint, bitor);
impl<'a, 'b> BitOr<&'b BigUint> for &'a BigUint {
impl<'a> BitOr<&'a BigUint> for BigUint {
type Output = BigUint;
fn bitor(self, other: &BigUint) -> BigUint {
let zeros = ZERO_VEC.iter().cycle();
let (a, b) = if self.data.len() > other.data.len() { (self, other) } else { (other, self) };
let ored = a.data.iter().zip(b.data.iter().chain(zeros)).map(
|(ai, bi)| *ai | *bi
).collect();
return BigUint::new(ored);
let mut data = self.data;
for (ai, &bi) in data.iter_mut().zip(other.data.iter()) {
*ai |= bi;
}
if other.data.len() > data.len() {
let extra = &other.data[data.len()..];
data.extend(extra.iter().cloned());
}
BigUint::new(data)
}
}
forward_all_binop!(impl BitXor for BigUint, bitxor);
forward_all_binop_to_val_ref_commutative!(impl BitXor for BigUint, bitxor);
impl<'a, 'b> BitXor<&'b BigUint> for &'a BigUint {
impl<'a> BitXor<&'a BigUint> for BigUint {
type Output = BigUint;
fn bitxor(self, other: &BigUint) -> BigUint {
let zeros = ZERO_VEC.iter().cycle();
let (a, b) = if self.data.len() > other.data.len() { (self, other) } else { (other, self) };
let xored = a.data.iter().zip(b.data.iter().chain(zeros)).map(
|(ai, bi)| *ai ^ *bi
).collect();
return BigUint::new(xored);
let mut data = self.data;
for (ai, &bi) in data.iter_mut().zip(other.data.iter()) {
*ai ^= bi;
}
if other.data.len() > data.len() {
let extra = &other.data[data.len()..];
data.extend(extra.iter().cloned());
}
BigUint::new(data)
}
}
@ -332,7 +428,7 @@ impl<'a> Shl<usize> for &'a BigUint {
fn shl(self, rhs: usize) -> BigUint {
let n_unit = rhs / big_digit::BITS;
let n_bits = rhs % big_digit::BITS;
return self.shl_unit(n_unit).shl_bits(n_bits);
self.shl_unit(n_unit).shl_bits(n_bits)
}
}
@ -350,7 +446,7 @@ impl<'a> Shr<usize> for &'a BigUint {
fn shr(self, rhs: usize) -> BigUint {
let n_unit = rhs / big_digit::BITS;
let n_bits = rhs % big_digit::BITS;
return self.shr_unit(n_unit).shr_bits(n_bits);
self.shr_unit(n_unit).shr_bits(n_bits)
}
}
@ -369,70 +465,85 @@ impl One for BigUint {
impl Unsigned for BigUint {}
forward_all_binop!(impl Add for BigUint, add);
forward_all_binop_to_val_ref_commutative!(impl Add for BigUint, add);
impl<'a, 'b> Add<&'b BigUint> for &'a BigUint {
impl<'a> Add<&'a BigUint> for BigUint {
type Output = BigUint;
fn add(self, other: &BigUint) -> BigUint {
let zeros = ZERO_VEC.iter().cycle();
let (a, b) = if self.data.len() > other.data.len() { (self, other) } else { (other, self) };
let mut sum = self.data;
if other.data.len() > sum.len() {
let additional = other.data.len() - sum.len();
sum.reserve(additional);
sum.extend(repeat(ZERO_BIG_DIGIT).take(additional));
}
let other_iter = other.data.iter().cloned().chain(repeat(ZERO_BIG_DIGIT));
let mut carry = 0;
let mut sum: Vec<BigDigit> = a.data.iter().zip(b.data.iter().chain(zeros)).map(|(ai, bi)| {
let (hi, lo) = big_digit::from_doublebigdigit(
(*ai as DoubleBigDigit) + (*bi as DoubleBigDigit) + (carry as DoubleBigDigit));
for (a, b) in sum.iter_mut().zip(other_iter) {
let d = (*a as DoubleBigDigit)
+ (b as DoubleBigDigit)
+ (carry as DoubleBigDigit);
let (hi, lo) = big_digit::from_doublebigdigit(d);
carry = hi;
lo
}).collect();
*a = lo;
}
if carry != 0 { sum.push(carry); }
return BigUint::new(sum);
BigUint::new(sum)
}
}
forward_all_binop!(impl Sub for BigUint, sub);
forward_all_binop_to_val_ref!(impl Sub for BigUint, sub);
impl<'a, 'b> Sub<&'b BigUint> for &'a BigUint {
impl<'a> Sub<&'a BigUint> for BigUint {
type Output = BigUint;
fn sub(self, other: &BigUint) -> BigUint {
let new_len = cmp::max(self.data.len(), other.data.len());
let zeros = ZERO_VEC.iter().cycle();
let (a, b) = (self.data.iter().chain(zeros.clone()), other.data.iter().chain(zeros));
let mut diff = self.data;
let other = &other.data;
assert!(diff.len() >= other.len(), "arithmetic operation overflowed");
let mut borrow = 0isize;
let diff: Vec<BigDigit> = a.take(new_len).zip(b).map(|(ai, bi)| {
let (hi, lo) = big_digit::from_doublebigdigit(
big_digit::BASE
+ (*ai as DoubleBigDigit)
- (*bi as DoubleBigDigit)
- (borrow as DoubleBigDigit)
);
let mut borrow: DoubleBigDigit = 0;
for (a, &b) in diff.iter_mut().zip(other.iter()) {
let d = big_digit::BASE - borrow
+ (*a as DoubleBigDigit)
- (b as DoubleBigDigit);
let (hi, lo) = big_digit::from_doublebigdigit(d);
/*
hi * (base) + lo == 1*(base) + ai - bi - borrow
=> ai - bi - borrow < 0 <=> hi == 0
*/
borrow = if hi == 0 { 1 } else { 0 };
lo
}).collect();
*a = lo;
}
assert!(borrow == 0,
"Cannot subtract other from self because other is larger than self.");
return BigUint::new(diff);
for a in &mut diff[other.len()..] {
if borrow == 0 { break }
let d = big_digit::BASE - borrow
+ (*a as DoubleBigDigit);
let (hi, lo) = big_digit::from_doublebigdigit(d);
borrow = if hi == 0 { 1 } else { 0 };
*a = lo;
}
assert!(borrow == 0, "arithmetic operation overflowed");
BigUint::new(diff)
}
}
forward_all_binop!(impl Mul for BigUint, mul);
forward_all_binop_to_val_ref_commutative!(impl Mul for BigUint, mul);
impl<'a, 'b> Mul<&'b BigUint> for &'a BigUint {
impl<'a> Mul<&'a BigUint> for BigUint {
type Output = BigUint;
fn mul(self, other: &BigUint) -> BigUint {
if self.is_zero() || other.is_zero() { return Zero::zero(); }
let (s_len, o_len) = (self.data.len(), other.data.len());
if s_len == 1 { return mul_digit(other, self.data[0]); }
if s_len == 1 { return mul_digit(other.clone(), self.data[0]); }
if o_len == 1 { return mul_digit(self, other.data[0]); }
// Using Karatsuba multiplication
@ -442,7 +553,7 @@ impl<'a, 'b> Mul<&'b BigUint> for &'a BigUint {
// a0*b0
let half_len = cmp::max(s_len, o_len) / 2;
let (s_hi, s_lo) = cut_at(self, half_len);
let (o_hi, o_lo) = cut_at(other, half_len);
let (o_hi, o_lo) = cut_at(other.clone(), half_len);
let ll = &s_lo * &o_lo;
let hh = &s_hi * &o_hi;
@ -459,27 +570,30 @@ impl<'a, 'b> Mul<&'b BigUint> for &'a BigUint {
return ll + mm.shl_unit(half_len) + hh.shl_unit(half_len * 2);
fn mul_digit(a: &BigUint, n: BigDigit) -> BigUint {
fn mul_digit(a: BigUint, n: BigDigit) -> BigUint {
if n == 0 { return Zero::zero(); }
if n == 1 { return a.clone(); }
if n == 1 { return a; }
let mut carry = 0;
let mut prod: Vec<BigDigit> = a.data.iter().map(|ai| {
let (hi, lo) = big_digit::from_doublebigdigit(
(*ai as DoubleBigDigit) * (n as DoubleBigDigit) + (carry as DoubleBigDigit)
);
let mut prod = a.data;
for a in &mut prod {
let d = (*a as DoubleBigDigit)
* (n as DoubleBigDigit)
+ (carry as DoubleBigDigit);
let (hi, lo) = big_digit::from_doublebigdigit(d);
carry = hi;
lo
}).collect();
*a = lo;
}
if carry != 0 { prod.push(carry); }
return BigUint::new(prod);
BigUint::new(prod)
}
#[inline]
fn cut_at(a: &BigUint, n: usize) -> (BigUint, BigUint) {
fn cut_at(mut a: BigUint, n: usize) -> (BigUint, BigUint) {
let mid = cmp::min(a.data.len(), n);
(BigUint::from_slice(&a.data[mid ..]),
BigUint::from_slice(&a.data[.. mid]))
let hi = BigUint::from_slice(&a.data[mid ..]);
a.data.truncate(mid);
(hi, BigUint::new(a.data))
}
#[inline]
@ -494,7 +608,7 @@ impl<'a, 'b> Mul<&'b BigUint> for &'a BigUint {
}
forward_all_binop!(impl Div for BigUint, div);
forward_all_binop_to_ref_ref!(impl Div for BigUint, div);
impl<'a, 'b> Div<&'b BigUint> for &'a BigUint {
type Output = BigUint;
@ -506,7 +620,7 @@ impl<'a, 'b> Div<&'b BigUint> for &'a BigUint {
}
}
forward_all_binop!(impl Rem for BigUint, rem);
forward_all_binop_to_ref_ref!(impl Rem for BigUint, rem);
impl<'a, 'b> Rem<&'b BigUint> for &'a BigUint {
type Output = BigUint;
@ -587,10 +701,10 @@ impl Integer for BigUint {
fn div_mod_floor(&self, other: &BigUint) -> (BigUint, BigUint) {
if other.is_zero() { panic!() }
if self.is_zero() { return (Zero::zero(), Zero::zero()); }
if *other == One::one() { return ((*self).clone(), Zero::zero()); }
if *other == One::one() { return (self.clone(), Zero::zero()); }
match self.cmp(other) {
Less => return (Zero::zero(), (*self).clone()),
Less => return (Zero::zero(), self.clone()),
Equal => return (One::one(), Zero::zero()),
Greater => {} // Do nothing
}
@ -1007,48 +1121,53 @@ impl BigUint {
#[inline]
fn shl_unit(&self, n_unit: usize) -> BigUint {
if n_unit == 0 || self.is_zero() { return (*self).clone(); }
if n_unit == 0 || self.is_zero() { return self.clone(); }
let mut v = repeat(ZERO_BIG_DIGIT).take(n_unit).collect::<Vec<_>>();
let mut v = vec![0; n_unit];
v.extend(self.data.iter().cloned());
BigUint::new(v)
}
#[inline]
fn shl_bits(&self, n_bits: usize) -> BigUint {
if n_bits == 0 || self.is_zero() { return (*self).clone(); }
fn shl_bits(self, n_bits: usize) -> BigUint {
if n_bits == 0 || self.is_zero() { return self; }
assert!(n_bits < big_digit::BITS);
let mut carry = 0;
let mut shifted: Vec<BigDigit> = self.data.iter().map(|elem| {
let (hi, lo) = big_digit::from_doublebigdigit(
(*elem as DoubleBigDigit) << n_bits | (carry as DoubleBigDigit)
);
carry = hi;
lo
}).collect();
if carry != 0 { shifted.push(carry); }
return BigUint::new(shifted);
let mut shifted = self.data;
for elem in shifted.iter_mut() {
let new_carry = *elem >> (big_digit::BITS - n_bits);
*elem = (*elem << n_bits) | carry;
carry = new_carry;
}
if carry != 0 {
shifted.push(carry);
}
BigUint::new(shifted)
}
#[inline]
fn shr_unit(&self, n_unit: usize) -> BigUint {
if n_unit == 0 { return (*self).clone(); }
if n_unit == 0 { return self.clone(); }
if self.data.len() < n_unit { return Zero::zero(); }
BigUint::from_slice(&self.data[n_unit ..])
}
#[inline]
fn shr_bits(&self, n_bits: usize) -> BigUint {
if n_bits == 0 || self.data.is_empty() { return (*self).clone(); }
fn shr_bits(self, n_bits: usize) -> BigUint {
if n_bits == 0 || self.data.is_empty() { return self; }
assert!(n_bits < big_digit::BITS);
let mut borrow = 0;
let mut shifted_rev = Vec::with_capacity(self.data.len());
for elem in self.data.iter().rev() {
shifted_rev.push((*elem >> n_bits) | borrow);
borrow = *elem << (big_digit::BITS - n_bits);
let mut shifted = self.data;
for elem in shifted.iter_mut().rev() {
let new_borrow = *elem << (big_digit::BITS - n_bits);
*elem = (*elem >> n_bits) | borrow;
borrow = new_borrow;
}
let shifted = { shifted_rev.reverse(); shifted_rev };
return BigUint::new(shifted);
BigUint::new(shifted)
}
/// Determines the fewest bits necessary to express the `BigUint`.