Playground/1brc
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

change temperature parsing approach (#405)

Browse Source
This commit is contained in:
zerninv 2024-01-14 19:47:42 +00:00 committed by GitHub
parent 5fb1218064
commit 990f884ff1
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 33 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

48
src/main/java/dev/morling/onebrc/CalculateAverage_zerninv.java
View File

@ -33,11 +33,18 @@ import java.util.concurrent.Future;
public class CalculateAverage_zerninv {
private static final String FILE = "./measurements.txt";
private static final int MIN_FILE_SIZE = 1024 * 1024 * 16;
private static final char DELIMITER = ';';
private static final char LINE_SEPARATOR = '\n';
private static final char ZERO = '0';
private static final char NINE = '9';
private static final char MINUS = '-';
// #.##
private static final int THREE_DIGITS_MASK = 0x2e0000;
// #.#
private static final int TWO_DIGITS_MASK = 0x2e00;
// #.#-
private static final int TWO_NEGATIVE_DIGITS_MASK = 0x2e002d;
private static final int BYTE_MASK = 0xff;
private static final int ZERO = '0';
private static final byte DELIMITER = ';';
private static final byte LINE_SEPARATOR = '\n';
private static final Unsafe UNSAFE = initUnsafe();
@ -111,7 +118,7 @@ public class CalculateAverage_zerninv {
var results = new MeasurementContainer();
long cityOffset;
int hashCode, temperature, multiplier;
int hashCode, temperature, word;
byte cityNameSize, b;
while (offset < end) {
@ -122,18 +129,27 @@ public class CalculateAverage_zerninv {
}
cityNameSize = (byte) (offset - cityOffset - 1);
multiplier = 1;
temperature = UNSAFE.getByte(offset++) - ZERO;
if (temperature == MINUS - ZERO) {
multiplier = -1;
temperature = 0;
word = UNSAFE.getInt(offset);
offset += 4;
if ((word & TWO_NEGATIVE_DIGITS_MASK) == TWO_NEGATIVE_DIGITS_MASK) {
word >>>= 8;
temperature = ZERO * 11 - ((word & BYTE_MASK) * 10 + ((word >>> 16) & BYTE_MASK));
}
while ((b = UNSAFE.getByte(offset++)) != LINE_SEPARATOR) {
if (b >= ZERO && b <= NINE) {
temperature = temperature * 10 + (b - ZERO);
}
else if ((word & THREE_DIGITS_MASK) == THREE_DIGITS_MASK) {
temperature = (word & BYTE_MASK) * 100 + ((word >>> 8) & BYTE_MASK) * 10 + ((word >>> 24) & BYTE_MASK) - ZERO * 111;
}
results.put(cityOffset, cityNameSize, hashCode, (short) (temperature * multiplier));
else if ((word & TWO_DIGITS_MASK) == TWO_DIGITS_MASK) {
temperature = (word & BYTE_MASK) * 10 + ((word >>> 16) & BYTE_MASK) - ZERO * 11;
offset--;
}
else {
// #.##-
word = (word >>> 8) | (UNSAFE.getByte(offset++) << 24);
temperature = ZERO * 111 - ((word & BYTE_MASK) * 100 + ((word >>> 8) & BYTE_MASK) * 10 + ((word >>> 24) & BYTE_MASK));
}
offset++;
results.put(cityOffset, cityNameSize, hashCode, (short) temperature);
}
return results.toStringMap();
}