Russ Cox
60a1f54d6b
math: make Hypot special cases look like others
...
The edit makes Hypot's description match the form
used in the other routines in this package.
R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/8003046
2013-03-25 17:01:40 -04:00
Oling Cat
1693e14bc4
math: modify a comment to the convention format.
...
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/8012043
2013-03-25 08:43:51 -07:00
Russ Cox
f6d0adedb6
build: remove dead code
...
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/7877045
2013-03-22 17:28:22 -04:00
Russ Cox
07720b67b3
build: update assembly variable names for vet
...
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/7834046
2013-03-22 12:57:55 -04:00
Oling Cat
79ae1ad489
math: remove repeated comment.
...
R=golang-dev, nigeltao
CC=golang-dev
https://golang.org/cl/7835046
2013-03-22 14:54:20 +11:00
Brad Fitzpatrick
e15c0ac693
all: remove now-unnecessary unreachable panics
...
Take advantage of the new terminating statement rule.
R=golang-dev, r, gri
CC=golang-dev
https://golang.org/cl/7712044
2013-03-11 14:16:55 -07:00
Russ Cox
65ed3f1073
math/big: fix comment
...
The variable is named b not bit.
R=golang-dev, agl
CC=golang-dev
https://golang.org/cl/7699043
2013-03-11 10:20:33 -04:00
Rob Pike
1bf66f081f
all: fix a few more printf arg bugs found by go vet
...
R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/7413045
2013-02-28 11:33:08 -08:00
Alan Donovan
19bb42d637
math/big: make tests faster by reducing coverage in --test.short mode.
...
The time to test all of math/big is now:
default => ~3min
--test.short => 150ms
R=rsc
CC=golang-dev
https://golang.org/cl/7223054
2013-01-30 18:19:58 -05:00
Alan Donovan
961195ae69
math/big: add Rat.{,Set}Float64 methods for IEEE 754 conversions.
...
Added tests, using input data from strconv.ParseFloat.
Thanks to rsc for most of the test code.
math/big could use some good package-level documentation.
R=remyoudompheng, rsc
CC=golang-dev
https://golang.org/cl/6930059
2013-01-28 18:00:15 -05:00
Russ Cox
92b2643c92
math/big: fix typo
...
Fixes #4678 .
TBR=gri
CC=golang-dev
https://golang.org/cl/7135059
2013-01-18 17:30:34 -05:00
Matthew Dempsky
46811d27ce
src: Use bytes.Equal instead of bytes.Compare where possible.
...
bytes.Equal is simpler to read and should also be faster because
of short-circuiting and assembly implementations.
Change generated automatically using:
gofmt -r 'bytes.Compare(a, b) == 0 -> bytes.Equal(a, b)'
gofmt -r 'bytes.Compare(a, b) != 0 -> !bytes.Equal(a, b)'
R=golang-dev, dave, adg, rsc
CC=golang-dev
https://golang.org/cl/7038051
2013-01-07 10:03:49 +11:00
Russ Cox
a3677b5f22
math: handle exponent separately in Log2
...
This guarantees that powers of two return exact answers.
We could do a multiprecision approximation for the
rest of the answer too, but this seems like it should be
good enough.
Fixes #4567 .
R=golang-dev, iant, remyoudompheng
CC=golang-dev
https://golang.org/cl/6943074
2012-12-20 12:23:27 -05:00
Shenghou Ma
d1ef9b56fb
all: fix typos
...
caught by https://github.com/lyda/misspell-check .
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/6949072
2012-12-19 03:04:09 +08:00
Luit van Drongelen
f4fc163d17
math/big: add SetUint64 and Uint64 functions to *Int
...
Implementation is mostly identical to passing a non-negative int64 to
SetInt64, and calling Int64 with a non-negative value in the *Int.
Fixes #4389 .
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6929048
2012-12-11 12:19:10 -05:00
David Symonds
ec45952670
math/rand: fix typo in example comment.
...
R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/6901056
2012-12-08 19:20:38 +11:00
Russ Cox
b99161e41f
math/rand: add example / regression test
...
This makes sure the outputs do not change for a fixed seed.
See also https://golang.org/cl/6905049 .
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6907048
2012-12-07 11:58:59 -05:00
Russ Cox
ff2534e076
undo CL 6845121 / 79603a5e4cda
...
This changes the output of
rand.Seed(0)
perm := rand.Perm(100)
When giving the same seeds to Go 1.0 and Go 1.1 programs
I would like them to generate the same random numbers.
««« original CL description
math/rand: remove noop iteration in Perm
The first iteration always do `m[0], m[0] = m[0], m[0]`, because
`rand.Intn(1)` is 0.
fun note: IIRC in TAOCP version of this algorithm, `i` goes
backward (n-1->1), meaning that the "already" shuffled part of the
array is never altered betweens iterations, while in the current
implementation the "not-yet" shuffled part of the array is
conserved between iterations.
R=golang-dev
CC=golang-dev
https://golang.org/cl/6845121
»»»
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6905049
2012-12-07 11:58:44 -05:00
Johan Euphrosine
d7b0f2a524
math/rand: remove noop iteration in Perm
...
The first iteration always do `m[0], m[0] = m[0], m[0]`, because
`rand.Intn(1)` is 0.
fun note: IIRC in TAOCP version of this algorithm, `i` goes
backward (n-1->1), meaning that the "already" shuffled part of the
array is never altered betweens iterations, while in the current
implementation the "not-yet" shuffled part of the array is
conserved between iterations.
R=golang-dev
CC=golang-dev
https://golang.org/cl/6845121
2012-12-01 14:11:46 -08:00
Shenghou Ma
9e30b708a1
all: set GOMAXPROCS to 1 when counting mallocs
...
also fix an annoying test that relies on $GOROOT be set.
Fixes #3690 .
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/6844086
2012-12-01 00:38:01 +08:00
Robert Griesemer
465b9c35e5
gofmt: apply gofmt -w src misc
...
Remove trailing whitespace in comments.
No other changes.
R=r
CC=golang-dev
https://golang.org/cl/6815053
2012-10-30 13:38:01 -07:00
Adam Langley
73f11171b4
math/big: add 4-bit, fixed window exponentiation.
...
A 4-bit window is convenient because 4 divides both 32 and 64,
therefore we never have a window spanning words of the exponent.
Additionaly, the benefit of a 5-bit window is only 2.6% at 1024-bits
and 3.3% at 2048-bits.
This code is still not constant time, however.
benchmark old ns/op new ns/op delta
BenchmarkRSA2048Decrypt 17108590 11180370 -34.65%
Benchmark3PrimeRSA2048Decrypt 13003720 7680390 -40.94%
R=gri
CC=golang-dev
https://golang.org/cl/6716048
2012-10-17 11:19:26 -04:00
Robert Griesemer
7565726875
math/big: fix big.Exp and document better
...
- always return 1 for y <= 0
- document that the sign of m is ignored
- protect against div-0 panics by treating
m == 0 the same way as m == nil
- added extra tests
Fixes #4239 .
R=agl, remyoudompheng, agl
CC=golang-dev
https://golang.org/cl/6724046
2012-10-16 13:46:27 -07:00
Robert Griesemer
4bee88d45f
math/big: more conservative use of lock for divisor table
...
Minor performance impact running sequentially:
benchmark old ns/op new ns/op delta
BenchmarkString10Base2 389 391 +0.51%
BenchmarkString100Base2 1530 1534 +0.26%
BenchmarkString1000Base2 11789 11787 -0.02%
BenchmarkString10000Base2 111443 112030 +0.53%
BenchmarkString100000Base2 1017483 1015347 -0.21%
BenchmarkString10Base8 339 344 +1.47%
BenchmarkString100Base8 753 756 +0.40%
BenchmarkString1000Base8 4618 4641 +0.50%
BenchmarkString10000Base8 43217 43534 +0.73%
BenchmarkString100000Base8 397518 400602 +0.78%
BenchmarkString10Base10 630 630 +0.00%
BenchmarkString100Base10 1975 1960 -0.76%
BenchmarkString1000Base10 10179 10174 -0.05%
BenchmarkString10000Base10 44527 44416 -0.25%
BenchmarkString100000Base10 14404694 14425308 +0.14%
BenchmarkString10Base16 283 288 +1.77%
BenchmarkString100Base16 597 598 +0.17%
BenchmarkString1000Base16 3189 3186 -0.09%
BenchmarkString10000Base16 29403 29364 -0.13%
BenchmarkString100000Base16 265657 265587 -0.03%
Note that due to other improvements (faster assembly routines,
better code generation by compiler), these benchmarks now run
up to 37% faster than they used to at the last time measured (1/9/2012).
Minor performance impact for StringPiParallel running in parallel:
Current CL but with Lock/Unlock commented out (removed):
BenchmarkStringPiParallel 5000 343581 ns/op
BenchmarkStringPiParallel-2 10000 184511 ns/op
BenchmarkStringPiParallel-3 10000 129768 ns/op
BenchmarkStringPiParallel-4 10000 102326 ns/op
Current CL:
BenchmarkStringPiParallel 5000 345169 ns/op
BenchmarkStringPiParallel-2 10000 185827 ns/op
BenchmarkStringPiParallel-3 10000 131168 ns/op
BenchmarkStringPiParallel-4 10000 102353 ns/op
Fixes #4218 .
R=dvyukov, michael.jones, dave
CC=golang-dev
https://golang.org/cl/6643053
2012-10-11 16:04:03 -07:00
Adam Langley
9070d5759f
math/big: avoid some allocation in Exp
...
benchmark old ns/op new ns/op delta
BenchmarkRSA1024Decrypt 745686 644964 -13.51%
BenchmarkRSA2048Decrypt 5517318 5049200 -8.48%
Benchmark3PrimeRSA2048Decrypt 3767386 3288048 -12.72%
R=gri
CC=gobot, golang-dev
https://golang.org/cl/6566043
2012-10-01 17:31:35 -04:00
Russ Cox
10ea6519e4
build: make int 64 bits on amd64
...
The assembly offsets were converted mechanically using
code.google.com/p/rsc/cmd/asmlint. The instruction
changes were done by hand.
Fixes #2188 .
R=iant, r, bradfitz, remyoudompheng
CC=golang-dev
https://golang.org/cl/6550058
2012-09-24 20:57:01 -04:00
Russ Cox
8572950383
all: match asm variable names to Go func declarations
...
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/6548046
2012-09-21 00:35:56 -04:00
Charles L. Dorian
0f8f5d2140
math: Faster Tanh
...
From 159 to 47.6 ns/op; slightly more accurate.
R=rsc, golang-dev, mtj, dave, remyoudompheng
CC=golang-dev
https://golang.org/cl/6500121
2012-09-17 17:18:16 -04:00
Christopher Swenson
baf426f10f
math/big: Replace RCLQ + ANDQ with SETCS in unrolled arithmetic assembly.
...
benchmark old ns/op new ns/op delta
BenchmarkAddVW_1 8 8 +0.60%
BenchmarkAddVW_2 10 9 -8.64%
BenchmarkAddVW_3 10 10 -4.63%
BenchmarkAddVW_4 10 11 +3.67%
BenchmarkAddVW_5 11 12 +5.98%
BenchmarkAddVW_1e1 18 20 +6.38%
BenchmarkAddVW_1e2 129 115 -10.85%
BenchmarkAddVW_1e3 1270 1089 -14.25%
BenchmarkAddVW_1e4 13376 12145 -9.20%
BenchmarkAddVW_1e5 130392 125260 -3.94%
benchmark old MB/s new MB/s speedup
BenchmarkAddVW_1 7709.10 7661.92 0.99x
BenchmarkAddVW_2 12451.10 13604.00 1.09x
BenchmarkAddVW_3 17727.81 18721.54 1.06x
BenchmarkAddVW_4 23552.64 22708.81 0.96x
BenchmarkAddVW_5 27411.40 25816.22 0.94x
BenchmarkAddVW_1e1 34063.19 32023.06 0.94x
BenchmarkAddVW_1e2 49529.97 55360.55 1.12x
BenchmarkAddVW_1e3 50380.44 58764.18 1.17x
BenchmarkAddVW_1e4 47843.59 52696.10 1.10x
BenchmarkAddVW_1e5 49082.60 51093.66 1.04x
R=gri, rsc, r
CC=golang-dev
https://golang.org/cl/6480063
2012-08-28 09:29:45 -07:00
Robert Griesemer
74c6325142
math/big: fix broken comment
...
R=iant, iant
CC=golang-dev
https://golang.org/cl/6485064
2012-08-24 13:50:09 -07:00
Robert Griesemer
3bd8684fac
math/big: minor tweaks to assembly code (slightly better performance)
...
Benchmarks run on 2.8GHz Quad-Code Intel Xeon,
4GB 800MHz DDR2 FB-DIMM ("PowerMac").
benchmark old ns/op new ns/op delta
BenchmarkAddVV_1 7 7 -0.82%
BenchmarkAddVV_2 8 8 -3.46%
BenchmarkAddVV_3 10 9 -4.81%
BenchmarkAddVV_4 9 9 -1.89%
BenchmarkAddVV_5 11 10 -5.22%
BenchmarkAddVV_1e1 17 18 +4.05%
BenchmarkAddVV_1e2 117 115 -1.71%
BenchmarkAddVV_1e3 1095 1090 -0.46%
BenchmarkAddVV_1e4 13149 12679 -3.57%
BenchmarkAddVV_1e5 135133 129482 -4.18%
BenchmarkAddVW_1 6 6 -1.14%
BenchmarkAddVW_2 7 7 +3.78%
BenchmarkAddVW_3 8 8 +0.12%
BenchmarkAddVW_4 8 8 -6.52%
BenchmarkAddVW_5 9 8 -3.70%
BenchmarkAddVW_1e1 14 13 -4.29%
BenchmarkAddVW_1e2 97 96 -1.33%
BenchmarkAddVW_1e3 953 940 -1.36%
BenchmarkAddVW_1e4 9776 9527 -2.55%
BenchmarkAddVW_1e5 102396 97738 -4.55%
benchmark old MB/s new MB/s speedup
BenchmarkAddVV_1 8702.84 8774.56 1.01x
BenchmarkAddVV_2 14739.60 15277.82 1.04x
BenchmarkAddVV_3 18375.37 19398.16 1.06x
BenchmarkAddVV_4 26935.44 27464.68 1.02x
BenchmarkAddVV_5 27754.04 29423.30 1.06x
BenchmarkAddVV_1e1 37050.89 35629.72 0.96x
BenchmarkAddVV_1e2 54289.15 55533.24 1.02x
BenchmarkAddVV_1e3 58428.83 58682.53 1.00x
BenchmarkAddVV_1e4 48670.55 50475.99 1.04x
BenchmarkAddVV_1e5 47360.54 49427.66 1.04x
BenchmarkAddVW_1 10397.27 10502.23 1.01x
BenchmarkAddVW_2 17279.03 16654.13 0.96x
BenchmarkAddVW_3 23858.39 23825.89 1.00x
BenchmarkAddVW_4 29799.42 31895.06 1.07x
BenchmarkAddVW_5 34781.83 36105.11 1.04x
BenchmarkAddVW_1e1 45629.88 47597.42 1.04x
BenchmarkAddVW_1e2 65341.93 66240.04 1.01x
BenchmarkAddVW_1e3 67153.67 68069.83 1.01x
BenchmarkAddVW_1e4 65464.60 67173.83 1.03x
BenchmarkAddVW_1e5 62501.88 65480.66 1.05x
R=iant
CC=golang-dev
https://golang.org/cl/6484056
2012-08-24 10:51:39 -07:00
Robert Griesemer
35422bc11f
math/big: faster (add|sub)V(V|W) routines
...
Benchmarks run on 3.06GHz Intel Core 2 Duo,
4GB 800MHz DDR2 SDRAM ("iMac").
benchmark old ns/op new ns/op delta
BenchmarkAddVV_1 6 6 +2.75%
BenchmarkAddVV_2 9 7 -19.71%
BenchmarkAddVV_3 9 9 +2.25%
BenchmarkAddVV_4 10 8 -20.46%
BenchmarkAddVV_5 12 10 -19.53%
BenchmarkAddVV_1e1 23 15 -32.48%
BenchmarkAddVV_1e2 213 107 -49.77%
BenchmarkAddVV_1e3 2088 993 -52.44%
BenchmarkAddVV_1e4 20874 12027 -42.38%
BenchmarkAddVV_1e5 209858 121480 -42.11%
BenchmarkAddVW_1 5 5 +0.90%
BenchmarkAddVW_2 11 11 -3.51%
BenchmarkAddVW_3 7 7 -0.27%
BenchmarkAddVW_4 8 7 -6.32%
BenchmarkAddVW_5 9 8 -10.89%
BenchmarkAddVW_1e1 17 12 -26.01%
BenchmarkAddVW_1e2 155 89 -42.32%
BenchmarkAddVW_1e3 1479 873 -40.97%
BenchmarkAddVW_1e4 13838 8764 -36.67%
BenchmarkAddVW_1e5 147353 89560 -39.22%
benchmark old MB/s new MB/s speedup
BenchmarkAddVV_1 9765.57 9508.55 0.97x
BenchmarkAddVV_2 13077.63 16284.97 1.25x
BenchmarkAddVV_3 20599.58 20156.67 0.98x
BenchmarkAddVV_4 23591.58 29516.02 1.25x
BenchmarkAddVV_5 24920.95 31194.10 1.25x
BenchmarkAddVV_1e1 27393.76 40621.71 1.48x
BenchmarkAddVV_1e2 29911.96 59592.99 1.99x
BenchmarkAddVV_1e3 30650.73 64429.84 2.10x
BenchmarkAddVV_1e4 30660.09 53213.08 1.74x
BenchmarkAddVV_1e5 30496.74 52683.46 1.73x
BenchmarkAddVW_1 11503.39 11405.98 0.99x
BenchmarkAddVW_2 11203.56 11586.92 1.03x
BenchmarkAddVW_3 26173.45 26224.75 1.00x
BenchmarkAddVW_4 30560.30 32621.94 1.07x
BenchmarkAddVW_5 33183.81 37269.94 1.12x
BenchmarkAddVW_1e1 36991.75 50098.53 1.35x
BenchmarkAddVW_1e2 41087.14 71549.93 1.74x
BenchmarkAddVW_1e3 43266.42 73279.83 1.69x
BenchmarkAddVW_1e4 46246.74 73021.97 1.58x
BenchmarkAddVW_1e5 43433.00 71459.96 1.65x
Benchmarks run on 2.8GHz Quad-Code Intel Xeon,
4GB 800MHz DDR2 FB-DIMM ("PowerMac").
benchmark old ns/op new ns/op delta
BenchmarkAddVV_1 7 7 +2.51%
BenchmarkAddVV_2 8 8 +3.70%
BenchmarkAddVV_3 10 10 +4.00%
BenchmarkAddVV_4 11 9 -19.49%
BenchmarkAddVV_5 14 11 -18.44%
BenchmarkAddVV_1e1 23 17 -27.00%
BenchmarkAddVV_1e2 234 117 -50.00%
BenchmarkAddVV_1e3 2284 1095 -52.06%
BenchmarkAddVV_1e4 22906 13149 -42.60%
BenchmarkAddVV_1e5 229860 135133 -41.21%
BenchmarkAddVW_1 6 6 +1.15%
BenchmarkAddVW_2 7 7 +1.37%
BenchmarkAddVW_3 7 8 +1.00%
BenchmarkAddVW_4 9 8 -6.93%
BenchmarkAddVW_5 10 9 -13.21%
BenchmarkAddVW_1e1 18 14 -24.32%
BenchmarkAddVW_1e2 170 97 -42.41%
BenchmarkAddVW_1e3 1619 953 -41.14%
BenchmarkAddVW_1e4 15142 9776 -35.44%
BenchmarkAddVW_1e5 160835 102396 -36.33%
benchmark old MB/s new MB/s speedup
BenchmarkAddVV_1 8928.95 8702.84 0.97x
BenchmarkAddVV_2 15298.84 14739.60 0.96x
BenchmarkAddVV_3 19116.52 18375.37 0.96x
BenchmarkAddVV_4 21644.30 26935.44 1.24x
BenchmarkAddVV_5 22771.64 27754.04 1.22x
BenchmarkAddVV_1e1 27017.62 37050.89 1.37x
BenchmarkAddVV_1e2 27326.09 54289.15 1.99x
BenchmarkAddVV_1e3 28016.84 58428.83 2.09x
BenchmarkAddVV_1e4 27939.38 48670.55 1.74x
BenchmarkAddVV_1e5 27843.00 47360.54 1.70x
BenchmarkAddVW_1 10510.97 10397.27 0.99x
BenchmarkAddVW_2 17499.71 17279.03 0.99x
BenchmarkAddVW_3 24093.93 23858.39 0.99x
BenchmarkAddVW_4 27733.08 29799.42 1.07x
BenchmarkAddVW_5 30267.17 34781.83 1.15x
BenchmarkAddVW_1e1 34566.78 45629.88 1.32x
BenchmarkAddVW_1e2 37521.89 65341.93 1.74x
BenchmarkAddVW_1e3 39513.18 67153.67 1.70x
BenchmarkAddVW_1e4 42263.80 65464.60 1.55x
BenchmarkAddVW_1e5 39792.21 62501.88 1.57x
R=iant, remyoudompheng, nightlyone, minux.ma
CC=golang-dev
https://golang.org/cl/6482062
2012-08-24 09:20:44 -07:00
Robert Griesemer
053b448d61
math/big: added benchmarks to establish baseline for core functions
...
BenchmarkAddVV_1 500000000 7.24 ns/op 8844.11 MB/s
BenchmarkAddVV_2 100000000 10.4 ns/op 12290.41 MB/s
BenchmarkAddVV_3 100000000 10.7 ns/op 17966.58 MB/s
BenchmarkAddVV_4 100000000 12.3 ns/op 20848.67 MB/s
BenchmarkAddVV_5 100000000 14.5 ns/op 21993.82 MB/s
BenchmarkAddVV_1e1 100000000 24.0 ns/op 26720.65 MB/s
BenchmarkAddVV_1e2 10000000 246 ns/op 26014.58 MB/s
BenchmarkAddVV_1e3 1000000 2416 ns/op 26485.06 MB/s
BenchmarkAddVV_1e4 100000 23874 ns/op 26806.36 MB/s
BenchmarkAddVV_1e5 10000 241155 ns/op 26538.87 MB/s
BenchmarkAddVW_1 500000000 6.12 ns/op 10461.91 MB/s
BenchmarkAddVW_2 200000000 11.0 ns/op 11596.63 MB/s
BenchmarkAddVW_3 200000000 8.97 ns/op 21409.82 MB/s
BenchmarkAddVW_4 100000000 10.8 ns/op 23696.72 MB/s
BenchmarkAddVW_5 100000000 12.5 ns/op 25524.88 MB/s
BenchmarkAddVW_1e1 100000000 21.5 ns/op 29786.32 MB/s
BenchmarkAddVW_1e2 10000000 168 ns/op 37925.36 MB/s
BenchmarkAddVW_1e3 1000000 1658 ns/op 38579.15 MB/s
BenchmarkAddVW_1e4 100000 16492 ns/op 38805.85 MB/s
BenchmarkAddVW_1e5 10000 172155 ns/op 37175.69 MB/s
BenchmarkAddMulVVW_1 100000000 12.9 ns/op 4968.49 MB/s
BenchmarkAddMulVVW_2 100000000 15.5 ns/op 8279.42 MB/s
BenchmarkAddMulVVW_3 100000000 13.4 ns/op 14340.53 MB/s
BenchmarkAddMulVVW_4 100000000 15.8 ns/op 16194.94 MB/s
BenchmarkAddMulVVW_5 100000000 18.9 ns/op 16906.61 MB/s
BenchmarkAddMulVVW_1e1 50000000 32.3 ns/op 19838.35 MB/s
BenchmarkAddMulVVW_1e2 10000000 285 ns/op 22427.28 MB/s
BenchmarkAddMulVVW_1e3 1000000 2777 ns/op 23040.42 MB/s
BenchmarkAddMulVVW_1e4 100000 27901 ns/op 22938.01 MB/s
BenchmarkAddMulVVW_1e5 10000 281087 ns/op 22768.73 MB/s
R=r
CC=golang-dev
https://golang.org/cl/6478055
2012-08-23 15:56:14 -07:00
Charles L. Dorian
a5c4e0fa2a
math: update definition of NaN in assembly language files
...
R=rsc, minux.ma, golang-dev, nigeltao
CC=golang-dev
https://golang.org/cl/6461047
2012-08-09 09:40:05 +10:00
Shenghou Ma
6e9506a7b4
math, runtime: use a NaN that matches gcc's
...
our old choice is not working properly at least on VFPv2 in
ARM1136JF-S (it's not preserved across float64->float32 conversions).
Fixes #3745 .
R=dave, rsc
CC=golang-dev
https://golang.org/cl/6344078
2012-08-07 09:57:14 +08:00
Robert Griesemer
6a135a0894
mat/big: more optimal Karatsuba threshold
...
benchmark old ns/op new ns/op delta
BenchmarkHilbert 6253043 6267289 +0.23%
BenchmarkMul 45355940 39490633 -12.93%
R=r
CC=golang-dev
https://golang.org/cl/6355104
2012-07-12 14:19:09 -07:00
Robert Griesemer
98ca655919
math/big: minor performance tuning
...
Reuse temporary slice to avoid extra allocations
(originally done correctly by remyoudompheng@gmail.com
in https://golang.org/cl/6345075/ ).
benchmark old ns/op new ns/op delta
BenchmarkHilbert 6252790 6262304 +0.15%
BenchmarkMul 45827438 45301002 -1.15%
R=r
CC=golang-dev
https://golang.org/cl/6346097
2012-07-12 14:12:50 -07:00
Rémy Oudompheng
ac12131649
math/big: correct quadratic space complexity in Mul.
...
The previous implementation used to have a O(n) recursion
depth for unbalanced inputs. A test is added to check that a
reasonable amount of bytes is allocated in this case.
Fixes #3807 .
R=golang-dev, dsymonds, gri
CC=golang-dev, remy
https://golang.org/cl/6345075
2012-07-12 10:18:24 -07:00
Matthew William Jibson
f83a47cbb9
doc: various "the the" and other typos
...
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/6355076
2012-07-08 11:57:04 +10:00
Rémy Oudompheng
1a0a09dafe
math/rand: avoid use of math.Pow in tests.
...
The use of math.Pow for mere squaring can be extremely
slow on soft-float ARM. Even on systems with hardware
floating-point, a speedup in test duration is observed.
On amd64
Before: ok math/rand 2.009s
After: ok math/rand 0.340s
Fixes #3740 .
R=dave, golang-dev, r, r
CC=golang-dev
https://golang.org/cl/6348061
2012-07-04 00:38:01 +02:00
David G. Andersen
917f764382
math/big: Remove unnecessary test from nat.go multiplication
...
The switch at the beginning of the function already ensures n > 1,
so testing for n < 2 is redundant.
R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/6350051
2012-07-02 15:30:00 -07:00
Charles L. Dorian
35c3afdb62
math: improve Atan, Asin and Acos accuracy
...
pkg/math/all_test.go tests Atan (and therefore Asin and Acos) to a
relative accuracy of 4e-16, but the test vector misses values where
the old algorithm was in error by more than that. For example:
x newError oldError
0.414215746 1.41e-16 -4.24e-16
0.414216076 1.41e-16 -4.24e-16
0.414217632 1.41e-16 -4.24e-16
0.414218770 1.41e-16 -4.24e-16
0.414225466 0 -5.65e-16
0.414226244 1.41e-16 -4.24e-16
0.414228756 0 -5.65e-16
0.414235089 0 -5.65e-16
0.414237070 0 -5.65e-16
R=rsc, golang-dev
CC=golang-dev
https://golang.org/cl/6302093
2012-06-24 19:39:07 -04:00
Robert Griesemer
10b88888f6
math/big: correctly test for positive inputs in Int.GCD
...
Also: better GCD tests.
R=rsc
CC=golang-dev
https://golang.org/cl/6295076
2012-06-13 13:54:36 -07:00
Robert Griesemer
f4240666be
math/big: fix binaryGCD
...
R=rsc
CC=golang-dev
https://golang.org/cl/6297085
2012-06-13 10:29:06 -07:00
Robert Griesemer
b7c5e23df0
math/big: various cleanups
...
R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/6295072
2012-06-13 09:37:47 -07:00
Christopher Swenson
38735b957c
math/big: Implemented binary GCD algorithm
...
benchmark old ns/op new ns/op delta
BenchmarkGCD10x10 4383 2126 -51.49%
BenchmarkGCD10x100 5612 2124 -62.15%
BenchmarkGCD10x1000 8843 2622 -70.35%
BenchmarkGCD10x10000 17025 6576 -61.37%
BenchmarkGCD10x100000 118985 48130 -59.55%
BenchmarkGCD100x100 45328 11683 -74.23%
BenchmarkGCD100x1000 50141 12678 -74.72%
BenchmarkGCD100x10000 110314 26719 -75.78%
BenchmarkGCD100x100000 630000 156041 -75.23%
BenchmarkGCD1000x1000 654809 137973 -78.93%
BenchmarkGCD1000x10000 985683 159951 -83.77%
BenchmarkGCD1000x100000 4920792 366399 -92.55%
BenchmarkGCD10000x10000 16848950 3732062 -77.85%
BenchmarkGCD10000x100000 55401500 4675876 -91.56%
BenchmarkGCD100000x100000 1126775000 258951800 -77.02%
R=gri, rsc, bradfitz, remyoudompheng, mtj
CC=golang-dev
https://golang.org/cl/6305065
2012-06-13 09:31:20 -07:00
Robert Griesemer
008c62b2cd
math/big: optimize common case of Int.Bit(0)
...
R=golang-dev, dsymonds, rsc
CC=golang-dev
https://golang.org/cl/6306069
2012-06-12 09:36:35 -07:00
Robert Griesemer
014d036d84
math/big: added nat.trailingZeroBits, simplified some code
...
Will simplify implementation of binaryGCD.
R=rsc, cswenson
CC=golang-dev
https://golang.org/cl/6299064
2012-06-08 13:00:49 -07:00
Robert Griesemer
cc1890cbe3
math/big: improved karatsuba calibration code, better mul benchmark
...
An attempt to profit from CL 6176043 (fix to superpolinomial
runtime of karatsuba multiplication) and determine a better
karatsuba threshold. The result indicates that 32 is still
a reasonable value. Left the threshold as is (== 32), but
made some minor changes to the calibrate code which are
worthwhile saving (use of existing benchmarking code for
better results, better use of package time).
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6260062
2012-06-04 09:48:27 -07:00
Shenghou Ma
d186d07eda
cmd/5a, cmd/5l, math: add CLZ instruction for ARM
...
Supported in ARMv5 and above.
R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/6284043
2012-06-03 03:08:49 +08:00
