encoding/base32: simplify and speed up decoder

First, we can lift the enc.decodeMap nil check out of the loop.

Second, we can make it clear to the compiler that 'in := src[0]' doesn't
need a bounds check, by making len(src)==0 a single if check that always
stops the loop. This is by far the largest speed-up.

Third, we can use a dst slice index instead of reslicing dst, which
removes work from the loop body.

While at it, we can merge the two 'switch dlen' pieces of code, which
simplifies the code and doesn't affect performance.

name            old time/op    new time/op    delta
DecodeString-8    80.2µs ± 0%    67.5µs ± 0%  -15.81%  (p=0.002 n=6+6)

name            old speed      new speed      delta
DecodeString-8   163MB/s ± 0%   194MB/s ± 0%  +18.78%  (p=0.002 n=6+6)

Change-Id: Iefeaae94c03453f8760452b1da706a77b3522718
Reviewed-on: https://go-review.googlesource.com/c/154422
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

1 files changed, 22 insertions, 30 deletions

diff --git a/src/encoding/base32/base32.go b/src/encoding/base32/base32.go
index 3fb6caceab..e14d2d4987 100644
--- a/src/encoding/base32/base32.go
+++ b/src/encoding/base32/base32.go
@@ -284,7 +284,12 @@ func (e CorruptInputError) Error() string {
 // additional data is an error. This method assumes that src has been
 // stripped of all supported whitespace ('\r' and '\n').
 func (enc *Encoding) decode(dst, src []byte) (n int, end bool, err error) {
+	// Lift the nil check outside of the loop.
+	_ = enc.decodeMap
+
+	dsti := 0
 	olen := len(src)
+
 	for len(src) > 0 && !end {
 		// Decode quantum using the base32 alphabet
 		var dbuf [8]byte
@@ -292,17 +297,15 @@ func (enc *Encoding) decode(dst, src []byte) (n int, end bool, err error) {
 
 		for j := 0; j < 8; {
 
-			// We have reached the end and are missing padding
-			if len(src) == 0 && enc.padChar != NoPadding {
-				return n, false, CorruptInputError(olen - len(src) - j)
-			}
-
-			// We have reached the end and are not expecing any padding
-			if len(src) == 0 && enc.padChar == NoPadding {
+			if len(src) == 0 {
+				if enc.padChar != NoPadding {
+					// We have reached the end and are missing padding
+					return n, false, CorruptInputError(olen - len(src) - j)
+				}
+				// We have reached the end and are not expecing any padding
 				dlen, end = j, true
 				break
 			}
-
 			in := src[0]
 			src = src[1:]
 			if in == byte(enc.padChar) && j >= 2 && len(src) < 8 {
@@ -339,37 +342,26 @@ func (enc *Encoding) decode(dst, src []byte) (n int, end bool, err error) {
 		// quantum
 		switch dlen {
 		case 8:
-			dst[4] = dbuf[6]<<5 | dbuf[7]
+			dst[dsti+4] = dbuf[6]<<5 | dbuf[7]
+			n++
 			fallthrough
 		case 7:
-			dst[3] = dbuf[4]<<7 | dbuf[5]<<2 | dbuf[6]>>3
+			dst[dsti+3] = dbuf[4]<<7 | dbuf[5]<<2 | dbuf[6]>>3
+			n++
 			fallthrough
 		case 5:
-			dst[2] = dbuf[3]<<4 | dbuf[4]>>1
+			dst[dsti+2] = dbuf[3]<<4 | dbuf[4]>>1
+			n++
 			fallthrough
 		case 4:
-			dst[1] = dbuf[1]<<6 | dbuf[2]<<1 | dbuf[3]>>4
+			dst[dsti+1] = dbuf[1]<<6 | dbuf[2]<<1 | dbuf[3]>>4
+			n++
 			fallthrough
 		case 2:
-			dst[0] = dbuf[0]<<3 | dbuf[1]>>2
-		}
-
-		if !end {
-			dst = dst[5:]
-		}
-
-		switch dlen {
-		case 2:
-			n += 1
-		case 4:
-			n += 2
-		case 5:
-			n += 3
-		case 7:
-			n += 4
-		case 8:
-			n += 5
+			dst[dsti+0] = dbuf[0]<<3 | dbuf[1]>>2
+			n++
 		}
+		dsti += 5
 	}
 	return n, end, nil
 }