flate: Improve speed in big stateless blocks. (#718)

* flate: Improve speed in big stateless blocks.

Don't re-alloc and copy dict for every block when compressing more than 32KB.

```
benchmark                         old ns/op     new ns/op     delta
BenchmarkEncodeDigitsSL1e4-32     52954         52850         -0.20%
BenchmarkEncodeDigitsSL1e5-32     781061        745420        -4.56%
BenchmarkEncodeDigitsSL1e6-32     8143640       7715674       -5.26%
BenchmarkEncodeTwainSL1e4-32      68150         68415         +0.39%
BenchmarkEncodeTwainSL1e5-32      715140        687326        -3.89%
BenchmarkEncodeTwainSL1e6-32      7718175       7339694       -4.90%

benchmark                         old MB/s     new MB/s     speedup
BenchmarkEncodeDigitsSL1e4-32     188.84       189.21       1.00x
BenchmarkEncodeDigitsSL1e5-32     128.03       134.15       1.05x
BenchmarkEncodeDigitsSL1e6-32     122.80       129.61       1.06x
BenchmarkEncodeTwainSL1e4-32      146.74       146.17       1.00x
BenchmarkEncodeTwainSL1e5-32      139.83       145.49       1.04x
BenchmarkEncodeTwainSL1e6-32      129.56       136.25       1.05x

benchmark                         old allocs     new allocs     delta
BenchmarkEncodeDigitsSL1e4-32     0              0              +0.00%
BenchmarkEncodeDigitsSL1e5-32     3              0              -100.00%
BenchmarkEncodeDigitsSL1e6-32     41             0              -100.00%
BenchmarkEncodeTwainSL1e4-32      0              0              +0.00%
BenchmarkEncodeTwainSL1e5-32      3              0              -100.00%
BenchmarkEncodeTwainSL1e6-32      41             0              -100.00%

benchmark                         old bytes     new bytes     delta
BenchmarkEncodeDigitsSL1e4-32     0             0             +0.00%
BenchmarkEncodeDigitsSL1e5-32     92929         9             -99.99%
BenchmarkEncodeDigitsSL1e6-32     1298964       97            -99.99%
BenchmarkEncodeTwainSL1e4-32      0             0             +0.00%
BenchmarkEncodeTwainSL1e5-32      92928         8             -99.99%
BenchmarkEncodeTwainSL1e6-32      1298871       92            -99.99%
```

* Pin garble to v0.7.2

Author: klauspost

.github/workflows/go.yml

@@ -85,7 +85,7 @@ jobs:
       run: go build github.com/klauspost/compress/s2/cmd/s2c && go build github.com/klauspost/compress/s2/cmd/s2d&&./s2c -verify s2c &&./s2d s2c.s2&&rm ./s2c&&rm s2d&&rm s2c.s2
 
     - name: install garble
-      run: go install mvdan.cc/[email protected].0
+      run: go install mvdan.cc/[email protected].2
 
     - name: goreleaser deprecation
       run: curl -sfL https://git.io/goreleaser | VERSION=v1.9.2 sh -s -- check

.github/workflows/release.yml

@@ -21,7 +21,7 @@ jobs:
           go-version: 1.19.x
       -
         name: install garble
-        run: go install mvdan.cc/[email protected].1
+        run: go install mvdan.cc/[email protected].2
       -
         name: Run GoReleaser
         uses: goreleaser/goreleaser-action@v2

.goreleaser.yml

@@ -3,7 +3,7 @@
 before:
   hooks:
     - ./gen.sh
-    - go install mvdan.cc/garble@latest
+    - go install mvdan.cc/garble@v0.7.2
 
 builds:
   -

flate/stateless.go

@@ -86,11 +86,19 @@ func StatelessDeflate(out io.Writer, in []byte, eof bool, dict []byte) error {
 		dict = dict[len(dict)-maxStatelessDict:]
 	}
 
+	// For subsequent loops, keep shallow dict reference to avoid alloc+copy.
+	var inDict []byte
+
 	for len(in) > 0 {
 		todo := in
-		if len(todo) > maxStatelessBlock-len(dict) {
+		if len(inDict) > 0 {
+			if len(todo) > maxStatelessBlock-maxStatelessDict {
+				todo = todo[:maxStatelessBlock-maxStatelessDict]
+			}
+		} else if len(todo) > maxStatelessBlock-len(dict) {
 			todo = todo[:maxStatelessBlock-len(dict)]
 		}
+		inOrg := in
 		in = in[len(todo):]
 		uncompressed := todo
 		if len(dict) > 0 {
@@ -102,7 +110,11 @@ func StatelessDeflate(out io.Writer, in []byte, eof bool, dict []byte) error {
 			todo = combined
 		}
 		// Compress
-		statelessEnc(&dst, todo, int16(len(dict)))
+		if len(inDict) == 0 {
+			statelessEnc(&dst, todo, int16(len(dict)))
+		} else {
+			statelessEnc(&dst, inDict[:maxStatelessDict+len(todo)], maxStatelessDict)
+		}
 		isEof := eof && len(in) == 0
 
 		if dst.n == 0 {
@@ -119,7 +131,8 @@ func StatelessDeflate(out io.Writer, in []byte, eof bool, dict []byte) error {
 		}
 		if len(in) > 0 {
 			// Retain a dict if we have more
-			dict = todo[len(todo)-maxStatelessDict:]
+			inDict = inOrg[len(uncompressed)-maxStatelessDict:]
+			dict = nil
 			dst.Reset()
 		}
 		if bw.err != nil {