RFR: 8333893: Optimization for StringBuilder append boolean & null [v4]

[Shaojin Wen]

On Tue, 11 Jun 2024 11:35:28 GMT, Shaojin Wen <duke at openjdk.org> wrote:

>> After PR https://github.com/openjdk/jdk/pull/16245, C2 optimizes stores into primitive arrays by combining values ​​into larger stores.
>> 
>> This PR rewrites the code of appendNull and append(boolean) methods so that these two methods can be optimized by C2.
>
> Shaojin Wen has updated the pull request incrementally with one additional commit since the last revision:
> 
>   revert

below stand-alone reproducer,  But StringUTF16.putChar is an instrinsic method and cannot be simulated.

import java.nio.ByteOrder;
import java.util.Arrays;

public class AppendNullTest {
    public static class StringBuilder {
        byte[] value;
        byte coder;
        int count;

        public StringBuilder(boolean utf16) {
            value = new byte[0];
            coder = (byte) (utf16 ? 1 : 0);
        }

        public static boolean canEncode(char cp) {
            return cp <= 0xff;
        }

        private StringBuilder appendNull() throws Throwable {
            int count = this.count;
            ensureCapacityInternal(count + 4);
            byte[] val = this.value;
            if (isLatin1()) {
                val[count    ] = 'n';
                val[count + 1] = 'u';
                val[count + 2] = 'l';
                val[count + 3] = 'l';
            } else {
                StringUTF16.putCharsAt(val, count, 'n', 'u', 'l', 'l');
            }
            this.count = count + 4;
            return this;
        }

        public StringBuilder append(boolean b) throws Throwable {
            int count = this.count;
            int spaceNeeded = count + (b ? 4 : 5);
            ensureCapacityInternal(spaceNeeded);
            byte[] val = this.value;
            if (isLatin1()) {
                if (b) {
                    val[count    ] = 't';
                    val[count + 1] = 'r';
                    val[count + 2] = 'u';
                    val[count + 3] = 'e';
                } else {
                    val[count    ] = 'f';
                    val[count + 1] = 'a';
                    val[count + 2] = 'l';
                    val[count + 3] = 's';
                    val[count + 4] = 'e';
                }
            } else {
                if (b) {
                    StringUTF16.putCharsAt(val, count, 't', 'r', 'u', 'e');
                } else {
                    StringUTF16.putCharsAt(val, count, 'f', 'a', 'l', 's', 'e');
                }
            }
            this.count = spaceNeeded;
            return this;
        }

        private void ensureCapacityInternal(int minimumCapacity) {
            int oldCapacity = value.length >> coder;
            if (minimumCapacity - oldCapacity > 0) {
                value = Arrays.copyOf(value,
                        newCapacity(minimumCapacity) << coder);
            }
        }

        private int newCapacity(int minCapacity) {
            int oldLength = value.length;
            int newLength = minCapacity << coder;
            int growth = newLength - oldLength;
            int length = oldLength + Math.max(growth, oldLength + (2 << coder));
            return length >> coder;
        }

        final boolean isLatin1() {
            return coder == 0;
        }

        public void clear() {
            int count = 0;
            coder = 0;
        }
    }

    static class StringUTF16 {
        public static void putCharsAt(byte[] value, int i, char c1, char c2, char c3, char c4) throws Throwable {
            putChar(value, i    , c1);
            putChar(value, i + 1, c2);
            putChar(value, i + 2, c3);
            putChar(value, i + 3, c4);
        }

        public static void putCharsAt(byte[] value, int i, char c1, char c2, char c3, char c4, char c5) throws Throwable {
            putChar(value, i    , c1);
            putChar(value, i + 1, c2);
            putChar(value, i + 2, c3);
            putChar(value, i + 3, c4);
            putChar(value, i + 4, c5);
        }

        private static final int HI_BYTE_SHIFT;
        private static final int LO_BYTE_SHIFT;
        static {
            if (ByteOrder.nativeOrder() == ByteOrder.BIG_ENDIAN) {
                HI_BYTE_SHIFT = 8;
                LO_BYTE_SHIFT = 0;
            } else {
                HI_BYTE_SHIFT = 0;
                LO_BYTE_SHIFT = 8;
            }
        }

        static void putChar(byte[] val, int index, int c) {
            index <<= 1;
            val[index++] = (byte)(c >> HI_BYTE_SHIFT);
            val[index]   = (byte)(c >> LO_BYTE_SHIFT);
        }
    }

    public static void main(String[] args) throws Throwable {
        StringBuilder sb = new StringBuilder(true);
        for (int i = 0; i < 1_000_000; i++) {
            sb.appendNull().appendNull().appendNull().appendNull().appendNull();
//            sb.append(true).append(false).append(true).append(false).append(true).append(false);
            sb.clear();
        }
    }
}

-------------

PR Comment: https://git.openjdk.org/jdk/pull/19626#issuecomment-2162441299