| Age | Commit message (Collapse) | Author |
|---|
|
This will let low-level things depend on the canonical routines,
even for floating-point printing.
Change-Id: I31207dc6584ad90d4e365dbe6eaf20f8662ed22d
Reviewed-on: https://go-review.googlesource.com/c/go/+/716000
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
|
|
Change-Id: I372233d8494665b3300f9a186c883a4254435e1c
Reviewed-on: https://go-review.googlesource.com/c/go/+/710617
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
|
|
Calling FormatFloat with an invalid value of fmt is expected
to return a string containing '%' and the input fmt character.
Since even before Go 1.0, the code has been panicking in the
case where prec=0.
Fixes #52187
Change-Id: I74fec601eedb7fe28efc5132c4253674661452aa
Reviewed-on: https://go-review.googlesource.com/c/go/+/402817
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Run-TryBot: Ian Lance Taylor <iant@google.com>
Reviewed-by: Carlos Amedee <carlos@golang.org>
Auto-Submit: Emmanuel Odeke <emmanuel@orijtech.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
|
|
This patch implements the algorithm from Ulf Adams,
"Ryū: Fast Float-to-String Conversion" (doi:10.1145/3192366.3192369)
for formatting floating-point numbers with a fixed number of decimal
digits.
It is not a direct translation of the reference C implementation
but still follows the original paper. In particular, it uses full
128-bit powers of 10, which allows for more precision in the other
modes (fixed ftoa, atof).
name old time/op new time/op delta
AppendFloat/Decimal-4 49.6ns ± 3% 59.3ns ± 0% +19.59% (p=0.008 n=5+5)
AppendFloat/Float-4 122ns ± 1% 91ns ± 1% -25.92% (p=0.008 n=5+5)
AppendFloat/Exp-4 89.3ns ± 1% 100.0ns ± 1% +11.98% (p=0.008 n=5+5)
AppendFloat/NegExp-4 88.3ns ± 2% 97.1ns ± 1% +9.87% (p=0.008 n=5+5)
AppendFloat/LongExp-4 143ns ± 2% 103ns ± 0% -28.17% (p=0.016 n=5+4)
AppendFloat/Big-4 144ns ± 1% 110ns ± 1% -23.26% (p=0.008 n=5+5)
AppendFloat/BinaryExp-4 46.2ns ± 2% 46.0ns ± 1% ~ (p=0.603 n=5+5)
AppendFloat/32Integer-4 49.1ns ± 1% 58.7ns ± 1% +19.57% (p=0.008 n=5+5)
AppendFloat/32ExactFraction-4 95.6ns ± 1% 88.6ns ± 1% -7.30% (p=0.008 n=5+5)
AppendFloat/32Point-4 122ns ± 1% 87ns ± 1% -28.63% (p=0.008 n=5+5)
AppendFloat/32Exp-4 88.6ns ± 2% 95.0ns ± 1% +7.29% (p=0.008 n=5+5)
AppendFloat/32NegExp-4 87.2ns ± 1% 91.3ns ± 1% +4.63% (p=0.008 n=5+5)
AppendFloat/32Shortest-4 107ns ± 1% 82ns ± 0% -24.08% (p=0.008 n=5+5)
AppendFloat/Slowpath64-4 1.00µs ± 1% 0.10µs ± 0% -89.92% (p=0.016 n=5+4)
AppendFloat/SlowpathDenormal64-4 34.1µs ± 3% 0.1µs ± 1% -99.72% (p=0.008 n=5+5)
Fixes #15672
Change-Id: Ib90dfa245f62490a6666671896013cf3f9a1fb22
Reviewed-on: https://go-review.googlesource.com/c/go/+/170080
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Trust: Nigel Tao <nigeltao@golang.org>
Run-TryBot: Emmanuel Odeke <emmanuel@orijtech.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Nigel Tao <nigeltao@golang.org>
|
|
This patch implements a simplified version of Ulf Adams,
"Ryū: Fast Float-to-String Conversion" (doi:10.1145/3192366.3192369)
for formatting floating-point numbers with a fixed number of decimal
digits.
It uses the same principles but does not need to handle
the complex task of finding a shortest representation.
This allows to handle a few more cases than Grisu3, notably
formatting with up to 18 significant digits.
name old time/op new time/op delta
AppendFloat/32Fixed8Hard-4 72.0ns ± 2% 56.0ns ± 2% -22.28% (p=0.000 n=10+10)
AppendFloat/32Fixed9Hard-4 74.8ns ± 0% 64.2ns ± 2% -14.16% (p=0.000 n=8+10)
AppendFloat/64Fixed1-4 60.4ns ± 1% 54.2ns ± 1% -10.31% (p=0.000 n=10+9)
AppendFloat/64Fixed2-4 66.3ns ± 1% 53.3ns ± 1% -19.54% (p=0.000 n=10+9)
AppendFloat/64Fixed3-4 61.0ns ± 1% 55.0ns ± 2% -9.80% (p=0.000 n=9+10)
AppendFloat/64Fixed4-4 66.9ns ± 0% 52.0ns ± 2% -22.20% (p=0.000 n=8+10)
AppendFloat/64Fixed12-4 95.5ns ± 1% 76.2ns ± 3% -20.19% (p=0.000 n=10+9)
AppendFloat/64Fixed16-4 1.62µs ± 0% 0.07µs ± 2% -95.69% (p=0.000 n=10+10)
AppendFloat/64Fixed12Hard-4 1.27µs ± 1% 0.07µs ± 1% -94.83% (p=0.000 n=9+9)
AppendFloat/64Fixed17Hard-4 3.68µs ± 1% 0.08µs ± 2% -97.86% (p=0.000 n=10+9)
AppendFloat/64Fixed18Hard-4 3.67µs ± 0% 3.72µs ± 1% +1.44% (p=0.000 n=9+10)
Updates #15672
Change-Id: I160963e141dd48287ad8cf57bcc3c686277788e8
Reviewed-on: https://go-review.googlesource.com/c/go/+/170079
Reviewed-by: Emmanuel Odeke <emmanuel@orijtech.com>
Trust: Emmanuel Odeke <emmanuel@orijtech.com>
Trust: Nigel Tao <nigeltao@golang.org>
Trust: Robert Griesemer <gri@golang.org>
Run-TryBot: Emmanuel Odeke <emmanuel@orijtech.com>
TryBot-Result: Go Bot <gobot@golang.org>
|
|
In ParseComplex, the "size" passed to parseFloatPrefix should be 64 for
complex128, not 128. It still works because of how parseFloatPrefix
is forgiving about the size if it's not 32, but worth fixing anyway.
Make ParseComplex and ParseFloat return a bit size error for anything
other than 128 or 64 (for ParseComplex), or 64 or 32 (for ParseFloat).
Add "InvalidBitSize" tests for these cases.
Add tests for ParseComplex with bitSize==64: this is done in a similar
way to how the ParseFloat 32-bit tests work, re-using the tests for the
larger bit size.
Add tests for FormatComplex -- there were none before.
Fixes #40706
Change-Id: I16ddd546e5237207cc3b8c2181dd708eca42b04f
Reviewed-on: https://go-review.googlesource.com/c/go/+/248219
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Minux Ma <minux@golang.org>
Trust: Robert Griesemer <gri@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
|
|
Float formatting uses a multiprecision fallback path where Grisu3
algorithm fails. This has a bug during the rounding phase: the
difference between the decimal value and the upper bound is examined
byte-by-byte and doesn't properly handle the case where the first
divergence has a difference of 1.
For instance (using an example from #29491), for the number
498484681984085570, roundShortest examines the three decimal values:
lower: 498484681984085536
d: 498484681984085568
upper: 498484681984085600
After examining the 16th digit, we know that rounding d up will fall
within the bounds unless all remaining digits of d are 9 and all
remaining digits of upper are 0:
d: ...855xx
upper: ...856xx
However, the loop forgets that d and upper have already diverged and
then on the next iteration sees that the 17th digit of d is actually
lower than the 17th digit of upper and decides that we still can't round
up:
d: ...8556x
upper: ...8560x
Thus the original value is incorrectly rounded down to
498484681984085560 instead of the closer (and equally short)
498484681984085570.
Thanks to Brian Kessler for diagnosing this bug.
Fix it by remembering when we've seen divergence in previous digits.
This CL also fixes another bug in the same loop: for some inputs, the
decimal value d or the lower bound may have fewer digits than the upper
bound, yet the iteration through the digits starts at i=0 for each of
them. For instance, given the float64 value 1e23, we have
d: 99999999999999991611392
upper: 100000000000000000000000
but the loop starts by comparing '9' to '1' rather than '0' to '1'.
I haven't found any cases where this second bug causes incorrect output
because when the digit comparison fails on the first loop iteration the
upper bound always has more nonzero digits (i.e., the expression
'i+1 < upper.nd' is always true).
Fixes #29491
Change-Id: I58856a7a2e47935ec2f233d9f717ef15c78bb2d0
Reviewed-on: https://go-review.googlesource.com/c/go/+/157697
Run-TryBot: Caleb Spare <cespare@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rémy Oudompheng <remyoudompheng@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
|
|
This CL updates FormatFloat to format
standard hexadecimal floating-point constants,
using the 'x' and 'X' verbs.
See golang.org/design/19308-number-literals for background.
For #29008.
Change-Id: I540b8f71d492cfdb7c58af533d357a564591f28b
Reviewed-on: https://go-review.googlesource.com/c/160242
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
|
|
Each URL was manually verified to ensure it did not serve up incorrect
content.
Change-Id: I4dc846227af95a73ee9a3074d0c379ff0fa955df
Reviewed-on: https://go-review.googlesource.com/115798
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
|
|
The relevant benchmark (on an Intel i7-4510U machine):
name old time/op new time/op delta
FormatFloat/Slowpath64-4 68.6µs ± 0% 44.1µs ± 2% -35.71% (p=0.000 n=13+15)
Change-Id: I67eb0e81ce74ed57752d0280059f91419f09e93b
Reviewed-on: https://go-review.googlesource.com/30099
Reviewed-by: Robert Griesemer <gri@golang.org>
|
|
This serves as an example of table-driven benchmarks which are analoguous to the common pattern for table-driven tests.
Change-Id: I47f94c121a7117dd1e4ba03b3f2f8bcb5da38063
Reviewed-on: https://go-review.googlesource.com/23470
Run-TryBot: Marcel van Lohuizen <mpvl@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
|
|
Change-Id: If886f15468680f7e1c589873066b4391eb9784b5
Reviewed-on: https://go-review.googlesource.com/14856
Reviewed-by: Alan Donovan <adonovan@google.com>
|
|
Use optimized formatBits function to format mantissa and exponent.
Add benchmark for binary exponent float format.
on darwin/386
benchmark old ns/op new ns/op delta
BenchmarkAppendFloatBinaryExp 520 122 -76.54%
on darwin/amd64
benchmark old ns/op new ns/op delta
BenchmarkAppendFloatBinaryExp 76.9 84.3 +9.62%
Change-Id: If543552f1960e1655bed3a4130914e5eaa3aac69
Reviewed-on: https://go-review.googlesource.com/5600
Reviewed-by: Robert Griesemer <gri@golang.org>
|
|
Preparation was in CL 134570043.
This CL contains only the effect of 'hg mv src/pkg/* src'.
For more about the move, see golang.org/s/go14nopkg.
|
