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

Attempt to fix segfault CalculateAverage_zerninv.java (#635)

Browse Source 
* attempt to fix segfault, graal native

* fix last bytes for last line handler

* fix typo

* one more attempt
This commit is contained in:
zerninv 2024-01-31 16:44:50 +00:00 committed by GitHub
parent e2ee4cbc30
commit 11a89d6cb8
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 59 additions and 30 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
calculate_average_zerninv.sh
View File

@ -15,5 +15,11 @@
# limitations under the License.
#
JAVA_OPTS="--enable-preview"
java $JAVA_OPTS --class-path target/average-1.0.0-SNAPSHOT.jar dev.morling.onebrc.CalculateAverage_zerninv
if [ -f target/CalculateAverage_zerninv_image ]; then
echo "Picking up existing native image 'target/CalculateAverage_zerninv_image', delete the file to select JVM mode." 1>&2
target/CalculateAverage_zerninv_image
else
JAVA_OPTS="--enable-preview -Xmx512m -XX:+UseSerialGC -XX:-TieredCompilation"
echo "Chosing to run the app in JVM mode as no native image was found, use prepare_zerninv.sh to generate." 1>&2
java $JAVA_OPTS --class-path target/average-1.0.0-SNAPSHOT.jar dev.morling.onebrc.CalculateAverage_zerninv
fi

7
prepare_zerninv.sh
View File

@ -17,4 +17,9 @@
source "$HOME/.sdkman/bin/sdkman-init.sh"
sdk use java 21.0.1-graal 1>&2
sdk use java 21.0.2-graal 1>&2
if [ ! -f target/CalculateAverage_zerninv_image ]; then
NATIVE_IMAGE_OPTS="--gc=epsilon -O3 -march=native -R:MaxHeapSize=512m -H:-GenLoopSafepoints --enable-preview --initialize-at-build-time=dev.morling.onebrc.CalculateAverage_zerninv"
native-image $NATIVE_IMAGE_OPTS -cp target/average-1.0.0-SNAPSHOT.jar -o target/CalculateAverage_zerninv_image dev.morling.onebrc.CalculateAverage_zerninv
fi

72
src/main/java/dev/morling/onebrc/CalculateAverage_zerninv.java
View File

@ -56,8 +56,7 @@ public class CalculateAverage_zerninv {
tasks[i] = new TaskThread((int) (fileSize / minChunkSize / CORES + 1));
}
var results = new HashMap<String, TemperatureAggregation>();
var chunks = splitByChunks(segment.address(), segment.address() + fileSize, minChunkSize, results);
var chunks = splitByChunks(segment.address(), segment.address() + fileSize, minChunkSize);
for (int i = 0; i < chunks.size() - 1; i++) {
var task = tasks[i % tasks.length];
task.addChunk(chunks.get(i), chunks.get(i + 1));
@ -67,6 +66,7 @@ public class CalculateAverage_zerninv {
task.start();
}
var results = new HashMap<String, TemperatureAggregation>();
for (var task : tasks) {
task.join();
task.collectTo(results);
@ -79,31 +79,8 @@ public class CalculateAverage_zerninv {
}
}
private static List<Long> splitByChunks(long address, long end, long minChunkSize, Map<String, TemperatureAggregation> results) {
// handle last line
long offset = end - 1;
int temperature = 0;
byte b;
int multiplier = 1;
while ((b = UNSAFE.getByte(offset--)) != ';') {
if (b >= '0' && b <= '9') {
temperature += (b - '0') * multiplier;
multiplier *= 10;
}
else if (b == '-') {
temperature = -temperature;
}
}
long cityNameEnd = offset;
while (UNSAFE.getByte(offset - 1) != '\n' && offset > address) {
offset--;
}
var cityName = new byte[(int) (cityNameEnd - offset + 1)];
UNSAFE.copyMemory(null, offset, cityName, Unsafe.ARRAY_BYTE_BASE_OFFSET, cityName.length);
results.put(new String(cityName, StandardCharsets.UTF_8), new TemperatureAggregation(temperature, 1, (short) temperature, (short) temperature));
private static List<Long> splitByChunks(long address, long end, long minChunkSize) {
// split by chunks
end = offset;
List<Long> result = new ArrayList<>((int) ((end - address) / minChunkSize + 1));
result.add(address);
while (address < end) {
@ -278,10 +255,51 @@ public class CalculateAverage_zerninv {
@Override
public void run() {
for (int i = 0; i < begins.size(); i++) {
calcForChunk(begins.get(i), ends.get(i));
var begin = begins.get(i);
var end = ends.get(i) - 1;
while (end > begin && UNSAFE.getByte(end - 1) != '\n') {
end--;
}
calcForChunk(begin, end);
calcLastLine(end);
}
}
private void calcLastLine(long offset) {
long cityOffset = offset;
long lastBytes = 0;
int hashCode = 0;
byte cityNameSize = 0;
byte b;
while ((b = UNSAFE.getByte(offset++)) != ';') {
lastBytes = (lastBytes << 8) | b;
hashCode = hashCode * 31 + b;
cityNameSize++;
}
int temperature;
int 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));
}
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;
}
else if ((word & TWO_DIGITS_MASK) == TWO_DIGITS_MASK) {
temperature = (word & BYTE_MASK) * 10 + ((word >>> 16) & BYTE_MASK) - ZERO * 11;
}
else {
// #.##-
word = (word >>> 8) | (UNSAFE.getByte(offset) << 24);
temperature = ZERO * 111 - ((word & BYTE_MASK) * 100 + ((word >>> 8) & BYTE_MASK) * 10 + ((word >>> 24) & BYTE_MASK));
}
container.put(cityOffset, cityNameSize, hashCode, lastBytes, (short) temperature);
}
private void calcForChunk(long offset, long end) {
long cityOffset, lastBytes, city, masked, hashCode;
int temperature, word, delimiterIdx;