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

Store station name as byte slice and only convert to string once at the end, but using u64 hash as key is still faster

Browse Source
This commit is contained in:
Fabian Schmidt 2024-12-31 11:24:31 +01:00
parent 98cd6e930c
commit 45b3014cbb
1 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/main/rust/src/implementations/libraries.rs
View File

@ -17,7 +17,8 @@ pub fn run() {
let mmap_ptr = mmap.as_ptr(); let mmap_ptr = mmap.as_ptr();
let file_length = mmap.len(); let file_length = mmap.len();
let hasher = FxBuildHasher; let hasher = FxBuildHasher;
let mut stations: HashMap<u64, (String, StationMeasurements)> = // Even if I could now just use the byte slice as a key, doing the hash is still faster
let mut stations: HashMap<u64, (&[u8], StationMeasurements)> =
HashMap::with_capacity_and_hasher(DEFAULT_HASHMAP_LENGTH, hasher); HashMap::with_capacity_and_hasher(DEFAULT_HASHMAP_LENGTH, hasher);
let (tx, rx) = mpsc::channel(); let (tx, rx) = mpsc::channel();
let cores = thread::available_parallelism().unwrap().into(); let cores = thread::available_parallelism().unwrap().into();
@ -43,7 +44,7 @@ pub fn run() {
let (start, end) = *bounds.get(i).unwrap(); let (start, end) = *bounds.get(i).unwrap();
let mmap_slice = unsafe { from_raw_parts(mmap_ptr.add(start), end - start) }; let mmap_slice = unsafe { from_raw_parts(mmap_ptr.add(start), end - start) };
s.spawn(move || { s.spawn(move || {
let mut t_stations: HashMap<u64, (String, StationMeasurements)> = let mut t_stations: HashMap<u64, (&[u8], StationMeasurements)> =
HashMap::with_capacity_and_hasher(DEFAULT_HASHMAP_LENGTH, hasher); HashMap::with_capacity_and_hasher(DEFAULT_HASHMAP_LENGTH, hasher);
for line in mmap_slice.split(|&byte| byte == b'\n') { for line in mmap_slice.split(|&byte| byte == b'\n') {
if line.is_empty() { if line.is_empty() {
@ -51,7 +52,6 @@ pub fn run() {
} }
let (station, temp) = line.rsplit_once(|&byte| byte == b';').unwrap(); let (station, temp) = line.rsplit_once(|&byte| byte == b';').unwrap();
let hash = hash::bytes(station); let hash = hash::bytes(station);
let station = unsafe { std::str::from_utf8_unchecked(station) };
let temp = parse::temp(temp); let temp = parse::temp(temp);
let measurements_option = t_stations.get_mut(&hash); let measurements_option = t_stations.get_mut(&hash);
if let Some((_, measurements)) = measurements_option { if let Some((_, measurements)) = measurements_option {
@ -63,7 +63,7 @@ pub fn run() {
count: 1, count: 1,
sum: temp, sum: temp,
}; };
t_stations.insert(hash, (station.to_string(), measurements)); t_stations.insert(hash, (station, measurements));
} }
} }
let _ = tx.send(t_stations); let _ = tx.send(t_stations);
@ -76,13 +76,14 @@ pub fn run() {
if let Some((_, joined_measurements)) = joined_measurements_options { if let Some((_, joined_measurements)) = joined_measurements_options {
joined_measurements.merge(measurements); joined_measurements.merge(measurements);
} else { } else {
stations.insert(*hash, (station.to_owned(), *measurements)); stations.insert(*hash, (station, *measurements));
} }
} }
} }
let mut stations: Vec<String> = stations let mut stations: Vec<String> = stations
.iter() .iter()
.map(|(_, (station, measurements))| { .map(|(_, (station, measurements))| {
let station = unsafe { std::str::from_utf8_unchecked(station) };
let measurements = measurements.to_string(); let measurements = measurements.to_string();
#[cfg(feature = "json")] #[cfg(feature = "json")]
{ {