avery/bilrow
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Branchless temperature parsing

Browse source
This commit is contained in:
Avery Winters 2024-03-18 11:39:32 -05:00
parent 9d597c8e39
commit e934cec398
Signed by: avery
SSH key fingerprint: SHA256:eesvLB5MMqHLZrAMFt6kEhqJWnASMLcET6Sgmw0FqZI
1 changed files with 17 additions and 21 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

38
calculate/src/optimized.rs
View file

@ -116,15 +116,12 @@ pub(crate) fn run(args: Args) -> Result<()> {
Ok(()) Ok(())
} }
fn parse_temp(temp: &[u8]) -> i64 { fn parse_temp(input: &[u8], sep_pos: usize, nl_pos: usize) -> i64 {
let dec = |b| (b - b'0') as i64; let sign = 2 * (unsafe { *input.get_unchecked(sep_pos + 1) } != b'-') as i64 - 1;
let (sign, a, b, c) = match temp { let a_byte = unsafe { *input.get_unchecked(nl_pos - 4) };
[b'-', a, b, b'.', c] => (-1, dec(a), dec(b), dec(c)), let a = (b'0' <= a_byte && a_byte <= b'9') as i64 * a_byte.wrapping_sub(b'0') as i64;
[b'-', b, b'.', c] => (-1, 0, dec(b), dec(c)), let b = (unsafe { *input.get_unchecked(nl_pos - 3) } - b'0') as i64;
[a, b, b'.', c] => (1, dec(a), dec(b), dec(c)), let c = (unsafe { *input.get_unchecked(nl_pos - 1) } - b'0') as i64;
[b, b'.', c] => (1, 0, dec(b), dec(c)),
_ => panic!("bad input"),
};
sign * (a * 100 + b * 10 + c) sign * (a * 100 + b * 10 + c)
} }
@ -134,16 +131,15 @@ fn chunk(input: &[u8], pb: ProgressBar) -> Result<HashMap<&[u8], Data>> {
let mut start = 0; let mut start = 0;
let mut last_pb_update: usize = 0; let mut last_pb_update: usize = 0;
let mut iter = memchr2_iter(b'\n', b';', input); let mut iter = memchr2_iter(b'\n', b';', input);
while let Some(split) = iter.next() { while let Some(sep_pos) = iter.next() {
if input[split] != b';' { if unsafe { *input.get_unchecked(sep_pos) } != b';' {
start = split + 1; start = sep_pos + 1;
continue; continue;
} }
let pos = iter.next().context("bad input")?; let nl_pos = unsafe { iter.next().unwrap_unchecked() };
let city = &input[start..split]; let city = unsafe { input.get_unchecked(start..sep_pos) };
let temperature = &input[split + 1..pos]; let temperature = parse_temp(input, sep_pos, nl_pos);
start = pos + 1; start = nl_pos + 1;
let temperature = parse_temp(temperature);
let entry = data.entry(city); let entry = data.entry(city);
entry entry
.and_modify(|data| { .and_modify(|data| {
@ -158,12 +154,12 @@ fn chunk(input: &[u8], pb: ProgressBar) -> Result<HashMap<&[u8], Data>> {
sum: temperature, sum: temperature,
count: 1, count: 1,
}); });
let last_pb_update_delta = pos - last_pb_update; let last_pb_update_delta = nl_pos - last_pb_update;
if last_pb_update_delta >= 10_000_000 { if last_pb_update_delta >= 10_000_000 {
pb.inc(last_pb_update_delta.try_into()?); pb.inc(last_pb_update_delta as u64);
last_pb_update = pos; last_pb_update = nl_pos;
} }
} }
pb.inc((input.len() - last_pb_update).try_into()?); pb.inc((input.len() - last_pb_update) as u64);
Ok(data) Ok(data)
} }