RFR: 8347009: Speed ​​up parseInt and parseLong [v6]

[Raffaello Giulietti]

On Tue, 14 Jan 2025 15:53:36 GMT, Shaojin Wen <swen at openjdk.org> wrote:

>> This is an optimization for decimal Integer.parseInt and Long.parseLong, which improves performance by about 10%. The optimization includes:
>> 1. Improve performance by parsing 2 numbers at a time, which has performance improvements for numbers with length >= 3.
>> 2. It uses charAt(0) for the first number. Assuming that the optimization can eliminate boundary checks, this will be more friendly to parsing numbers with length 1.
>> 3. It removes the reliance on the Character.digit method and eliminates the reliance on the CharacterDataLatin1#DIGITS cache array, which avoids performance degradation caused by cache misses.
>
> Shaojin Wen has updated the pull request incrementally with two additional commits since the last revision:
> 
>  - remove unused
>  - Update src/java.base/share/classes/jdk/internal/util/DecimalDigits.java
>    
>    Co-authored-by: Chen Liang <liach at openjdk.org>

I believe the following implementation of `Integer.parseInt(String,int)` is correct, shorter, and more readable.
On my M1 Pro, the benchmark is even slightly faster than the proposed one, but I didn't check on other platforms.

The same holds for `Long.parseLong(String,int)` (simply copy the text and change the type or `result` to be `long`).

Please have a look.


    public static int parseInt(String s, int radix)
                throws NumberFormatException {
        int len;
        byte[] value;
        if (s == null || radix != 10 || (len = (value = s.value()).length) == 0 || !s.isLatin1()) {
            return parseInt0(s, radix);
        }
        int fc = value[0];
        int result = Integer.isDigitLatin1(fc)
                ? '0' - fc
                : len != 1 && (fc == '-' || fc == '+')
                ? 0
                : 1;  // or any value > 0
        int i = 1;
        int d;
        while (i + 1 < len
                && (d = DecimalDigits.digit2(value, i)) != -1
                && MIN_VALUE / 100 <= result) {
            result = result * 100 - d;  // overflow from d => result > 0
            i += 2;
        }
        if (i < len
                && Integer.isDigitLatin1(d = value[i])
                && MIN_VALUE / 10 <= result && result <= 0) {
            result = result * 10 + '0' - d;  // overflow from '0' - d => result > 0
            i += 1;
        }
        if (i == len
                & result <= 0
                & (MIN_VALUE < result || fc == '-')) {
            return fc == '-' ? result : -result;
        }
        throw NumberFormatException.forInputString(s);
    }

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

PR Comment: https://git.openjdk.org/jdk/pull/22919#issuecomment-2613033520