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

Inline and optimize value parsing code for each of the four semicolon position processing branches. This provides a small but noticeable speed-up. It also expands and obfuscates the code, unfortunately. (#563)

Browse Source
This commit is contained in:
Elliot Barlas 2024-01-23 07:31:04 -08:00 committed by GitHub
parent e16ad7660d
commit df891354d3
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 112 additions and 47 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

159
src/main/java/dev/morling/onebrc/CalculateAverage_ebarlas.java
View File

@ -185,48 +185,114 @@ public class CalculateAverage_ebarlas {
long keyAddr = keyBaseAddr; // address for next int long keyAddr = keyBaseAddr; // address for next int
int keyArrLen = 0; // number of key 4-byte ints int keyArrLen = 0; // number of key 4-byte ints
int keyLastBytes; // occupancy in last byte (1, 2, 3, or 4) int keyLastBytes; // occupancy in last byte (1, 2, 3, or 4)
byte b0, b1, b2, b3; int val;
while (true) { while (true) {
int n = UNSAFE.getInt(cursor); int n = UNSAFE.getInt(cursor);
cursor += 4; cursor += 4;
b0 = (byte) (n & 0xFF); if ((n & 0xFF) == ';') { // ;vvv
b1 = (byte) ((n >> 8) & 0xFF);
b2 = (byte) ((n >> 16) & 0xFF);
b3 = (byte) ((n >> 24) & 0xFF);
if (b0 == ';') { // ...;1.1
UNSAFE.putInt(keyAddr, 0); // always pad with extra int to facilitate 8-byte aligned comparisons UNSAFE.putInt(keyAddr, 0); // always pad with extra int to facilitate 8-byte aligned comparisons
keyLastBytes = 4; keyLastBytes = 4;
b0 = b1; byte b0 = (byte) ((n >> 8) & 0xFF);
b1 = b2; byte b1 = (byte) ((n >> 16) & 0xFF);
b2 = b3; byte b2 = (byte) ((n >> 24) & 0xFF);
b3 = (byte) (UNSAFE.getByte(cursor++) & 0xFF); if (b0 == '-') {
if (b2 != '.') { // 6 bytes: -dd.dn
cursor++; // decimal point
byte b4 = UNSAFE.getByte(cursor);
cursor += 2; // adv beyond digit and newline
val = -(((b1 - '0') * 10 + (b2 - '0')) * 10 + (b4 - '0'));
}
else { // 5 bytes: -d.dn
byte b3 = UNSAFE.getByte(cursor);
cursor += 2; // digit and newline
val = -((b1 - '0') * 10 + (b3 - '0'));
}
}
else {
if (b1 != '.') { // 5 bytes: dd.dn
var b3 = UNSAFE.getByte(cursor);
cursor += 2; // digit and newline
val = ((b0 - '0') * 10 + (b1 - '0')) * 10 + (b3 - '0');
}
else { // 4 bytes: d.dn
cursor++; // newline
val = (b0 - '0') * 10 + (b2 - '0');
}
}
break; break;
} }
else if (b1 == ';') { // ...a;1.1 else if ((n & 0xFF00) == 0x3b00) { // k;vv
int k = n & 0xFF; int k = n & 0xFF;
UNSAFE.putLong(keyAddr, k); // pad with extra int for comparison alignment UNSAFE.putLong(keyAddr, k); // pad with extra int for comparison alignment
keyLastBytes = 1; keyLastBytes = 1;
keyArrLen++; keyArrLen++;
keyHash += k; keyHash += k;
b0 = b2; byte b0 = (byte) ((n >> 16) & 0xFF);
b1 = b3; byte b1 = (byte) ((n >> 24) & 0xFF);
b2 = (byte) (UNSAFE.getByte(cursor++) & 0xFF); byte b2 = UNSAFE.getByte(cursor++);
b3 = (byte) (UNSAFE.getByte(cursor++) & 0xFF); if (b0 == '-') {
if (b2 != '.') { // 6 bytes: -dd.dn
cursor++; // decimal point
byte b4 = UNSAFE.getByte(cursor);
cursor += 2; // adv beyond digit and newline
val = -(((b1 - '0') * 10 + (b2 - '0')) * 10 + (b4 - '0'));
}
else { // 5 bytes: -d.dn
byte b3 = UNSAFE.getByte(cursor);
cursor += 2; // digit newline
val = -((b1 - '0') * 10 + (b3 - '0'));
}
}
else {
if (b1 != '.') { // 5 bytes: dd.dn
byte b3 = UNSAFE.getByte(cursor);
cursor += 2; // newline
val = ((b0 - '0') * 10 + (b1 - '0')) * 10 + (b3 - '0');
}
else { // 4 bytes: d.dn
cursor++;
val = (b0 - '0') * 10 + (b2 - '0');
}
}
break; break;
} }
else if (b2 == ';') { // ...ab;1.1 else if ((n & 0xFF0000) == 0x3b0000) { // kk;v
int k = n & 0xFFFF; int k = n & 0xFFFF;
UNSAFE.putLong(keyAddr, k); // pad with extra int for comparison alignment UNSAFE.putLong(keyAddr, k); // pad with extra int for comparison alignment
keyLastBytes = 2; keyLastBytes = 2;
keyArrLen++; keyArrLen++;
keyHash += k; keyHash += k;
b0 = b3; byte b0 = (byte) ((n >> 24) & 0xFF);
b1 = (byte) (UNSAFE.getByte(cursor++) & 0xFF); if (b0 == '-') {
b2 = (byte) (UNSAFE.getByte(cursor++) & 0xFF); n = UNSAFE.getInt(cursor);
b3 = (byte) (UNSAFE.getByte(cursor++) & 0xFF); cursor += 4;
byte b1 = (byte) (n & 0xFF);
byte b2 = (byte) ((n >> 8) & 0xFF);
byte b3 = (byte) ((n >> 16) & 0xFF);
if (b2 != '.') { // 6 bytes: -dd.dn
byte b4 = (byte) ((n >> 24) & 0xFF);
cursor++; // newline
val = -(((b1 - '0') * 10 + (b2 - '0')) * 10 + (b4 - '0'));
}
else { // 5 bytes: -d.dn
val = -((b1 - '0') * 10 + (b3 - '0'));
}
}
else {
byte b1 = UNSAFE.getByte(cursor++);
byte b2 = UNSAFE.getByte(cursor++);
byte b3 = UNSAFE.getByte(cursor++);
if (b1 != '.') { // 5 bytes: dd.dn
cursor++; // newline
val = ((b0 - '0') * 10 + (b1 - '0')) * 10 + (b3 - '0');
}
else { // 4 bytes: d.dn
val = (b0 - '0') * 10 + (b2 - '0');
}
}
break; break;
} }
else if (b3 == ';') { // ...abc;1.1 else if ((n & 0xFF000000) == 0x3b000000) { // kkk;
int k = n & 0xFFFFFF; int k = n & 0xFFFFFF;
UNSAFE.putLong(keyAddr, k); // pad with extra int for comparison alignment UNSAFE.putLong(keyAddr, k); // pad with extra int for comparison alignment
keyLastBytes = 3; keyLastBytes = 3;
@ -234,13 +300,33 @@ public class CalculateAverage_ebarlas {
keyHash += k; keyHash += k;
n = UNSAFE.getInt(cursor); n = UNSAFE.getInt(cursor);
cursor += 4; cursor += 4;
b0 = (byte) (n & 0xFF); byte b0 = (byte) (n & 0xFF);
b1 = (byte) ((n >> 8) & 0xFF); byte b1 = (byte) ((n >> 8) & 0xFF);
b2 = (byte) ((n >> 16) & 0xFF); byte b2 = (byte) ((n >> 16) & 0xFF);
b3 = (byte) ((n >> 24) & 0xFF); byte b3 = (byte) ((n >> 24) & 0xFF);
if (b0 == '-') {
if (b2 != '.') { // 6 bytes: -dd.dn
byte b4 = UNSAFE.getByte(cursor);
cursor += 2; // adv beyond digit and newline
val = -(((b1 - '0') * 10 + (b2 - '0')) * 10 + (b4 - '0'));
}
else { // 5 bytes: -d.dn
cursor++; // newline
val = -((b1 - '0') * 10 + (b3 - '0'));
}
}
else {
if (b1 != '.') { // 5 bytes: dd.dn
cursor++; // newline
val = ((b0 - '0') * 10 + (b1 - '0')) * 10 + (b3 - '0');
}
else { // 4 bytes: d.dn
val = (b0 - '0') * 10 + (b2 - '0');
}
}
break; break;
} }
else { else { // kkkk
UNSAFE.putInt(keyAddr, n); UNSAFE.putInt(keyAddr, n);
keyArrLen++; keyArrLen++;
keyAddr += 4; keyAddr += 4;
@ -256,27 +342,6 @@ public class CalculateAverage_ebarlas {
else if (!equals(st.keyAddr, st.keyLen, keyBaseAddr, keyArrLen)) { else if (!equals(st.keyAddr, st.keyLen, keyBaseAddr, keyArrLen)) {
st = findInTable(stats, keyHash, keyBaseAddr, keyArrLen, keyLastBytes); st = findInTable(stats, keyHash, keyBaseAddr, keyArrLen, keyLastBytes);
} }
int val;
if (b0 == '-') {
if (b2 != '.') { // 6 bytes: -dd.dn
var b = UNSAFE.getByte(cursor);
cursor += 2; // adv beyond digit and newline
val = -(((b1 - '0') * 10 + (b2 - '0')) * 10 + (b - '0'));
}
else { // 5 bytes: -d.dn
cursor++; // newline
val = -((b1 - '0') * 10 + (b3 - '0'));
}
}
else {
if (b1 != '.') { // 5 bytes: dd.dn
cursor++; // newline
val = ((b0 - '0') * 10 + (b1 - '0')) * 10 + (b3 - '0');
}
else { // 4 bytes: d.dn
val = (b0 - '0') * 10 + (b2 - '0');
}
}
st.min = Math.min(st.min, val); st.min = Math.min(st.min, val);
st.max = Math.max(st.max, val); st.max = Math.max(st.max, val);
st.sum += val; st.sum += val;