hash/adler32: Unroll loop for extra performance.

name         old time/op    new time/op    delta
Adler32KB-4     592ns ± 0%     447ns ± 0%  -24.49%  (p=0.000 n=19+20)

name         old speed      new speed      delta
Adler32KB-4  1.73GB/s ± 0%  2.29GB/s ± 0%  +32.41%  (p=0.000 n=20+20)

Change-Id: I38990aa66ca4452a886200018a57c0bc3af30717
Reviewed-on: https://go-review.googlesource.com/21880
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

diff --git a/src/hash/adler32/adler32.go b/src/hash/adler32/adler32.go
index 0c733f751af5742d72977be67ca6960762c00fb3..21d6a2e1dc660fdf7633eed080fa1a9e889a4222 100644 (file)
--- a/src/hash/adler32/adler32.go
+++ b/src/hash/adler32/adler32.go
@@ -42,7 +42,7 @@ func New() hash.Hash32 {
 
 func (d *digest) Size() int { return Size }
 
-func (d *digest) BlockSize() int { return 1 }
+func (d *digest) BlockSize() int { return 4 }
 
 // Add p to the running checksum d.
 func update(d digest, p []byte) digest {
@@ -52,6 +52,17 @@ func update(d digest, p []byte) digest {
                if len(p) > nmax {
                        p, q = p[:nmax], p[nmax:]
                }
+               for len(p) >= 4 {
+                       s1 += uint32(p[0])
+                       s2 += s1
+                       s1 += uint32(p[1])
+                       s2 += s1
+                       s1 += uint32(p[2])
+                       s2 += s1
+                       s1 += uint32(p[3])
+                       s2 += s1
+                       p = p[4:]
+               }
                for _, x := range p {
                        s1 += uint32(x)
                        s2 += s1