Permalink
Browse files

strconv: use % instead of computing the remainder from the quotient

The compiler recognizes that in a sequence q = x/y; r = x%y only
one division is required. Remove prior work-arounds and write
more readable straight-line code (this also results in fewer
instructions, though it doesn't appear to affect the benchmarks
significantly).

name          old time/op  new time/op  delta
FormatInt-8   2.95µs ± 1%  2.92µs ± 5%   ~     (p=0.952 n=5+5)
AppendInt-8   1.91µs ± 1%  1.89µs ± 2%   ~     (p=0.421 n=5+5)
FormatUint-8   795ns ± 2%   782ns ± 4%   ~     (p=0.444 n=5+5)
AppendUint-8   557ns ± 1%   557ns ± 2%   ~     (p=0.548 n=5+5)

https://perf.golang.org/search?q=upload:20170310.1

Also:
- use uint instead of uintptr where we want to guarantee single-
  register operations
- remove some unnecessary conversions (before indexing)
- add more comments and fix some comments

Change-Id: I04858dc2d798a6495879d9c7cfec2fdc2957b704
Reviewed-on: https://go-review.googlesource.com/38071
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
  • Loading branch information...
griesemer committed Mar 10, 2017
1 parent e19f184 commit 95c5227c15da3c0c61eeea70f0a8288088301b98
Showing with 26 additions and 22 deletions.
  1. +26 −22 src/strconv/itoa.go
View
@@ -39,9 +39,9 @@ func AppendUint(dst []byte, i uint64, base int) []byte {
return dst
}
const (
digits = "0123456789abcdefghijklmnopqrstuvwxyz"
)
const host32bit = ^uint(0)>>32 == 0
const digits = "0123456789abcdefghijklmnopqrstuvwxyz"
var shifts = [len(digits) + 1]uint{
1 << 1: 1,
@@ -71,61 +71,65 @@ func formatBits(dst []byte, u uint64, base int, neg, append_ bool) (d []byte, s
}
// convert bits
// We use uint values where we can because those will
// fit into a single register even on a 32bit machine.
if base == 10 {
// common case: use constants for / because
// the compiler can optimize it into a multiply+shift
if ^uintptr(0)>>32 == 0 {
for u > uint64(^uintptr(0)) {
if host32bit {
// convert the lower digits using 32bit operations
for u >= 1e9 {
// the compiler recognizes q = a/b; r = a%b
// and produces only one DIV instruction;
// no need to be clever here
q := u / 1e9
us := uintptr(u - q*1e9) // us % 1e9 fits into a uintptr
us := uint(u % 1e9) // u % 1e9 fits into a uint
for j := 9; j > 0; j-- {
i--
qs := us / 10
a[i] = byte(us - qs*10 + '0')
us = qs
a[i] = byte(us%10 + '0')
us /= 10
}
u = q
}
// u < 1e9
}
// u guaranteed to fit into a uintptr
us := uintptr(u)
// u guaranteed to fit into a uint
us := uint(u)
for us >= 10 {
i--
q := us / 10
a[i] = byte(us - q*10 + '0')
us = q
a[i] = byte(us%10 + '0')
us /= 10
}
// u < 10
// us < 10
i--
a[i] = byte(us + '0')
} else if s := shifts[base]; s > 0 {
// base is power of 2: use shifts and masks instead of / and %
b := uint64(base)
m := uintptr(b) - 1 // == 1<<s - 1
m := uint(base) - 1 // == 1<<s - 1
for u >= b {
i--
a[i] = digits[uintptr(u)&m]
a[i] = digits[uint(u)&m]
u >>= s
}
// u < base
i--
a[i] = digits[uintptr(u)]
a[i] = digits[u]
} else {
// general case
b := uint64(base)
for u >= b {
i--
q := u / b
a[i] = digits[uintptr(u-q*b)]
u = q
a[i] = digits[u%b]
u /= b
}
// u < base
i--
a[i] = digits[uintptr(u)]
a[i] = digits[u]
}
// add sign, if any

0 comments on commit 95c5227

Please sign in to comment.