8280124: Reduce branches decoding latin-1 chars from UTF-8 encoded bytes

Reviewed-by: rriggs, alanb, naoto

src/java.base/share/classes/java/lang/String.java

@@ -541,8 +541,7 @@ public final class String
                             offset++;
                             continue;
                         }
-                        if ((b1 == (byte)0xc2 || b1 == (byte)0xc3) &&
+                        if ((b1 & 0xfe) == 0xc2 && offset + 1 < sl) { // b1 either 0xc2 or 0xc3
-                                offset + 1 < sl) {
                             int b2 = bytes[offset + 1];
                             if (!isNotContinuation(b2)) {
                                 dst[dp++] = (byte)decode2(b1, b2);
@@ -698,8 +697,7 @@ public final class String
                         offset++;
                         continue;
                     }
-                    if ((b1 == (byte) 0xc2 || b1 == (byte) 0xc3) &&
+                    if ((b1 & 0xfe) == 0xc2 && offset + 1 < sl) { // b1 either 0xc2 or 0xc3
-                            offset + 1 < sl) {
                         int b2 = bytes[offset + 1];
                         if (!isNotContinuation(b2)) {
                             dst[dp++] = (byte) decode2(b1, b2);

test/micro/org/openjdk/bench/java/lang/StringDecode.java

@@ -40,59 +40,120 @@ import java.util.concurrent.TimeUnit;
 
 @BenchmarkMode(Mode.AverageTime)
 @OutputTimeUnit(TimeUnit.NANOSECONDS)
-@Fork(value = 3, jvmArgs = "-Xmx1g")
+@Fork(value = 3)
 @Warmup(iterations = 5, time = 2)
 @Measurement(iterations = 5, time = 3)
 @State(Scope.Thread)
 public class StringDecode {
 
-    @BenchmarkMode(Mode.AverageTime)
+    @Param({"US-ASCII", "ISO-8859-1", "UTF-8", "MS932", "ISO-8859-6", "ISO-2022-KR"})
-    @OutputTimeUnit(TimeUnit.NANOSECONDS)
+    private String charsetName;
-    @Fork(value = 3, jvmArgs = "-Xmx1g")
-    @Warmup(iterations = 5, time = 2)
-    @Measurement(iterations = 5, time = 2)
-    @State(Scope.Thread)
-    public static class WithCharset {
 
-        @Param({"US-ASCII", "ISO-8859-1", "UTF-8", "MS932", "ISO-8859-6", "ISO-2022-KR"})
+    private Charset charset;
-        private String charsetName;
+    private byte[] asciiString;
-
+    private byte[] utf16String;
-        private Charset charset;
+    private byte[] longUtf16String;
-        private byte[] asciiString;
+    private byte[] longUtf16StartString;
-        private byte[] utf16String;
+    private byte[] longLatin1String;
-
-        @Setup
-        public void setup() {
-            charset = Charset.forName(charsetName);
-            asciiString = "ascii string".getBytes(charset);
-            utf16String = "UTF-\uFF11\uFF16 string".getBytes(charset);
-        }
-
-        @Benchmark
-        public void decodeCharsetName(Blackhole bh) throws Exception {
-            bh.consume(new String(asciiString, charsetName));
-            bh.consume(new String(utf16String, charsetName));
-        }
-
-        @Benchmark
-        public void decodeCharset(Blackhole bh) throws Exception {
-            bh.consume(new String(asciiString, charset));
-            bh.consume(new String(utf16String, charset));
-        }
-    }
-
-    private byte[] asciiDefaultString;
-    private byte[] utf16DefaultString;
 
     @Setup
     public void setup() {
-        asciiDefaultString = "ascii string".getBytes();
+        charset = Charset.forName(charsetName);
-        utf16DefaultString = "UTF-\uFF11\uFF16 string".getBytes();
+        asciiString = "ascii string".getBytes(charset);
+        utf16String = "UTF-\uFF11\uFF16 string".getBytes(charset);
+        longUtf16String = """
+             Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aliquam ac sem eu
+             urna egestas placerat. Etiam finibus ipsum nulla, non mattis dolor cursus a.
+             Nulla nec nisl consectetur, lacinia neque id, accumsan ante. Curabitur et
+             sapien in magna porta ultricies. Sed vel pellentesque nibh. Pellentesque dictum
+             dignissim diam eu ultricies. Class aptent taciti sociosqu ad litora torquent
+             per conubia nostra, per inceptos himenaeos. Suspendisse erat diam, fringilla
+             sed massa sed, posuere viverra orci. Suspendisse tempor libero non gravida
+             efficitur. Vivamus lacinia risus non orci viverra, at consectetur odio laoreet.
+             Suspendisse potenti.
+
+             Phasellus vel nisi iaculis, accumsan quam sed, bibendum eros. Sed venenatis
+             nulla tortor, et eleifend urna sodales id. Nullam tempus ac metus sit amet
+             sollicitudin. Nam sed ex diam. Praesent vitae eros et neque condimentum
+             consectetur eget non tortor. Praesent bibendum vel felis nec dignissim.
+             Maecenas a enim diam. Suspendisse quis ligula at nisi accumsan lacinia id
+             hendrerit sapien. \uFF11Donec aliquam mattis lectus eu ultrices. Duis eu nisl\uFF11
+             euismod, blandit mauris vel, \uFF11placerat urna. Etiam malesuada enim purus,
+             tristique mollis odio blandit quis.\uFF11 Vivamus posuere. \uFF11
+             \uFF11
+            """.getBytes(charset);
+        longUtf16StartString = """
+             \uFF11
+             Lorem ipsum dolor sit amet, \uFF11consectetur adipiscing elit. Aliquam ac sem eu
+             urna egestas \uFF11placerat. Etiam finibus ipsum nulla, non mattis dolor cursus a.
+             Nulla \uFF11nec nisl consectetur, lacinia neque id, accumsan ante. Curabitur et
+             sapien in \uFF11magna porta ultricies. \uFF11Sed vel pellentesque nibh. Pellentesque dictum
+             dignissim diam eu ultricies. Class aptent taciti sociosqu ad litora torquent
+             per conubia nostra, per inceptos himenaeos. Suspendisse erat diam, fringilla
+             sed massa sed, posuere viverra orci. Suspendisse tempor libero non gravida
+             efficitur. Vivamus lacinia risus non orci viverra, at consectetur odio laoreet.
+             Suspendisse potenti.
+
+             Phasellus vel nisi iaculis, accumsan quam sed, bibendum eros. Sed venenatis
+             nulla tortor, et eleifend urna sodales id. Nullam tempus ac metus sit amet
+             sollicitudin. Nam sed ex diam. Praesent vitae eros et neque condimentum
+             consectetur eget non tortor. Praesent bibendum vel felis nec dignissim.
+             Maecenas a enim diam. Suspendisse quis ligula at nisi accumsan lacinia id
+             hendrerit sapien. Donec aliquam mattis lectus eu ultrices. Duis eu nisl
+             euismod, blandit mauris vel, placerat urna. Etiam malesuada enim purus,
+             tristique mollis odio blandit quis. Vivamus posuere.
+            """.getBytes(charset);
+
+        longLatin1String = """
+             a\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             b\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             c\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             d\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             e\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             f\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             g\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             h\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             i\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6
+             j\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6\u00F6
+             k\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6\u00F6
+             l\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6\u00F6
+             m\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00F6\u00B6\u00F6\u00F6
+            """.getBytes(charset);
     }
 
     @Benchmark
-    public void decodeDefault(Blackhole bh) throws Exception {
+    public String decodeAsciiCharsetName() throws Exception {
-        bh.consume(new String(asciiDefaultString));
+        return new String(asciiString, charsetName);
-        bh.consume(new String(utf16DefaultString));
+    }
+
+    @Benchmark
+    public String decodeAscii() throws Exception {
+        return new String(asciiString, charset);
+    }
+
+    @Benchmark
+    public String decodeLatin1Long() throws Exception {
+        return new String(longLatin1String, charset);
+    }
+
+    @Benchmark
+    public String decodeUTF16Short() throws Exception {
+        return new String(utf16String, charset);
+    }
+
+    @Benchmark
+    public String decodeUTF16LongEnd() throws Exception {
+        return new String(longUtf16String, charset);
+    }
+
+    @Benchmark
+    public String decodeUTF16LongStart() throws Exception {
+        return new String(longUtf16StartString, charset);
+    }
+
+    @Benchmark
+    public void decodeUTF16LongMixed(Blackhole bh) throws Exception {
+        bh.consume(new String(longUtf16StartString, charset));
+        bh.consume(new String(longUtf16String, charset));
     }
 }