Commit Graph

99 Commits

Author SHA1 Message Date
Josh Stone 2f8f952d1d clean up unused macros 2017-09-21 17:36:21 -07:00
Josh Stone 2a9750ada4 bigint: remove an unused mut 2017-09-21 17:28:37 -07:00
Josh Stone 7679cb86fb Remove `#[must_use]` on `__add2`
It doesn't actually work on functions yet, and nightly now warns that it
is experimental, behind `#[feature(fn_must_use)]`.
2017-09-21 17:26:20 -07:00
bors[bot] 4896746fec Merge #328
328: Optimizing BigUint and Bigint multiplication with the Toom-3 algorithm r=cuviper a=kompass

Hi !

I finally implemented the Toom-3 algorithm ! I first tried to minimize the memory allocations by allocating the `Vec<BigDigit>` myself, as was done for Toom-2, but Toom-3 needs more complex calculations, with negative numbers. So I gave up this method, to use `BigInt` directly, and it's already faster ! I also chose a better threshold for the Toom-2 algorithm.

Before any modification :
```
running 4 tests
test multiply_0        ... bench:         257 ns/iter (+/- 25)
test multiply_1        ... bench:      30,240 ns/iter (+/- 1,651)
test multiply_2        ... bench:   2,752,360 ns/iter (+/- 52,102)
test multiply_3        ... bench:  11,618,575 ns/iter (+/- 266,286)
```

With a better Toom-2 threshold (16 instead of 4) :
```
running 4 tests
test multiply_0        ... bench:         130 ns/iter (+/- 8)
test multiply_1        ... bench:      19,772 ns/iter (+/- 1,083)
test multiply_2        ... bench:   1,340,644 ns/iter (+/- 17,987)
test multiply_3        ... bench:   7,302,854 ns/iter (+/- 82,060)
```

With the Toom-3 algorithm (with a threshold of 300):
```
running 4 tests
test multiply_0        ... bench:         123 ns/iter (+/- 3)
test multiply_1        ... bench:      19,689 ns/iter (+/- 837)
test multiply_2        ... bench:   1,189,589 ns/iter (+/- 29,101)
test multiply_3        ... bench:   3,014,225 ns/iter (+/- 61,222)
```

I think this could be optimized, but it's a first step !
2017-09-20 20:53:40 +00:00
Josh Stone 1ddbee7f37 bigint mac3: tweak thresholds between algorithms
It's not too rigorous, but thresholds 32 and 256 give me better results.

Before:

     test multiply_0        ... bench:          87 ns/iter (+/- 0)
     test multiply_1        ... bench:      11,926 ns/iter (+/- 19)
     test multiply_2        ... bench:     772,178 ns/iter (+/- 3,068)
     test multiply_3        ... bench:   2,034,237 ns/iter (+/- 9,618)

After:

     test multiply_0        ... bench:          87 ns/iter (+/- 0)
     test multiply_1        ... bench:      11,927 ns/iter (+/- 64)
     test multiply_2        ... bench:     672,440 ns/iter (+/- 3,570)
     test multiply_3        ... bench:   1,577,065 ns/iter (+/- 11,137)
2017-09-20 13:19:00 -07:00
Josh Stone 28d84ca3ac Toom-3: operate more on values where possible 2017-09-20 13:17:06 -07:00
Josh Stone 2c2e46c8df Add comments about multiplication strategy 2017-09-20 13:15:44 -07:00
Josh Stone 05dc87c041 Improve mac_digit bounds checking
By starting with `split_at_mut`, the hot multiplication loop runs with
no bounds checking at all!  The remaining carry loop has a slightly
simpler check for when the remaining iterator runs dry.
2017-09-20 11:41:59 -07:00
Josh Stone 3c1c84307c bigint: make sure bigand is normalized 2017-09-19 17:53:13 -07:00
Josh Stone 952dba4e06 remove unused lifetime from forward_val_assign 2017-09-19 17:36:17 -07:00
Josh Stone 953087fe4e Fix biguint assert_assign_op for rust 1.8
The `$op` only seems to be accepted when used indirectly through another
macro, in this case `assert_eq`.  How nice.
2017-09-19 14:00:51 -07:00
Josh Stone 98a3f17db6 Fix impl_rem_assign_scalar for rust 1.8 2017-09-19 13:47:59 -07:00
Alice Ryhl ff8f106186 Test *Assign for BigUint 2017-09-03 10:44:17 +02:00
Alice Ryhl 555dab7d33 Implement ShrAssign and ShlAssign for BigUint 2017-09-03 09:56:29 +02:00
Alice Ryhl 2f6c0bf354 Implement BitXorAssign for BigUint 2017-09-03 09:53:33 +02:00
Alice Ryhl 8c3b2de11c Implement BitAndAssign and BitOrAssign for BigUint 2017-09-02 23:37:54 +02:00
Alice Ryhl 23085800e0 Implement RemAssign for BigUint 2017-09-02 23:29:34 +02:00
Alice Ryhl e85ab24567 Forward by-value biguint and scalars to {Add,Mul,Sub,Div}Assign 2017-09-02 22:38:12 +02:00
Alice Ryhl 04117fafe9 Implement DivAssign for BigUint 2017-08-26 14:50:01 +02:00
Alice Ryhl 17030ea412 Implement SubAssign for BigUint 2017-08-26 14:50:01 +02:00
Alice Ryhl 8c47ca00c7 Implement MulAssign for BigUint 2017-08-26 14:50:01 +02:00
Alice Ryhl 03d717f26f Implement AddAssign for BigUint 2017-08-26 14:50:01 +02:00
Nicolas Kirchner bcd76c55e8 Optimize mac_digit 2017-08-15 21:14:15 +02:00
Nicolas Kirchner 243bc6fe4c Optimize Toom-3 algorithm 2017-08-15 19:27:22 +02:00
Nicolas Kirchner b43c1ab258 Replace the use of a feature not yet implemented in rust 1.8 2017-08-15 01:39:43 +02:00
Nicolas Kirchner c9c40b9402 Optimize and clean the Toom-3 algorithm and choose better thresholds 2017-08-14 20:07:35 +02:00
Nicolas Kirchner d7554ad931 Naive implementation of the Toom-3 algorithm
The Toom-2 algorithm is bypassed for tests.
2017-08-13 23:32:03 +02:00
Nicolas Kirchner cbdaf8f6f9 Optimize `BigInt::from_biguint` and `BigInt::assign_from_slice`
It removes useless memory allocations.
2017-08-02 22:25:21 +02:00
Nicolas Kirchner 5106fcc95a Replace `Vec::copy_from_slice` not implemented in rust 1.8 by another method 2017-08-02 16:38:38 +02:00
Nicolas Kirchner 9b56d6667c Implement assign_from_slice 2017-08-02 16:12:04 +02:00
Josh Stone ea2158f3dd bigint: bump to 0.1.40 2017-07-14 17:32:18 -07:00
bors[bot] ebc36b3e55 Merge #313
313: Scalar operations across all integer types r=cuviper

With my apologies for opening a new PR, and also for the 8 month delay, this continues the work started in #237 - the discussion there outlines the goals I was aiming for. I suppose this supersedes that PR and the other one can now be closed.

This PR adds support for Add, Sub, Mul, Div and Rem operations involving one BigInt/BigUint and one primitive integer, with operands in either order, and any combination of owned/borrowed arguments.
2017-07-12 05:16:29 +00:00
Josh Stone e5434dc659 Add assert_scalar_op! for DRYer testing 2017-07-11 21:59:10 -07:00
Josh Stone 6afac825d9 test and fix more scalar add cases 2017-07-11 17:27:19 -07:00
Josh Stone 18a5bfcd0b fix endianness of to/from_doublebigdigit calls 2017-07-11 17:22:11 -07:00
Josh Stone 18cc1902fb inline i32_abs_as_u32 and i64_abs_as_u64 2017-07-11 17:01:15 -07:00
Keith Wansbrough ef83e851e7 Add keywords and categories to subcrates too. 2017-07-11 09:30:42 +01:00
Sam Cappleman-Lynes 1fb03ca18a Make new code work on rustc-1.8.0
- Don't apply attributes to statements (1.12.0)
 - Don't use checked_abs (1.13.0)
2017-06-30 00:39:37 +01:00
Sam Cappleman-Lynes 2a3cd41820 Add scalar ops for all remaining integer types 2017-06-29 22:18:54 +01:00
Sam Cappleman-Lynes fd87d87db3 Fix normalization in scalar addition 2017-06-29 20:40:54 +01:00
Sam Cappleman-Lynes 99873d06e5 Scalar operations on integer types up to 32 bits 2017-06-29 18:29:14 +01:00
Sam Cappleman-Lynes 94d570697c Add operations on i32 to BigInt 2017-06-29 17:20:17 +01:00
Sam Cappleman-Lynes 9b0392d235 Add scalar division to BigInt 2017-06-29 16:19:11 +01:00
Sam Cappleman-Lynes 8b1288ea01 Add scalar multiplication to BigInt 2017-06-29 15:46:07 +01:00
Sam Cappleman-Lynes 79448cbdf9 Add scalar subtraction to BigInt 2017-06-29 15:15:59 +01:00
Sam Cappleman-Lynes 80feea2722 Also implement scalar addition for BigInt 2017-06-29 14:07:44 +01:00
Sam Cappleman-Lynes 1e26bdde81 Remove unnecessary normalization 2017-06-29 13:53:08 +01:00
Sam Cappleman-Lynes d0bfb54eee All variants of dividing BigUint by BigDigit
Allow the division to occur with either operand order and with any
combination of owned and borrowed arguments.
2017-06-29 13:38:00 +01:00
Sam Cappleman-Lynes 51408a9b3b All variants of subtracting BigDigit from BigUint
Allow the subtraction to occur with either operand order and with any
combination of owned and borrowed arguments.
2017-06-29 10:12:53 +01:00
Sam Cappleman-Lynes 5738141b7c Distinction for commutative scalar ops 2017-06-29 09:59:42 +01:00