Skip to content

Optimize out some read operations from the quick deflate algorithm #374

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
May 28, 2025
Merged

Optimize out some read operations from the quick deflate algorithm #374

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 17 additions & 12 deletions zlib-rs/src/deflate/algorithm/quick.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -93,23 +93,27 @@ pub fn deflate_quick(stream: &mut DeflateStream, flush: DeflateFlush) -> BlockSt
}
}

let lc: u8;
if state.lookahead >= WANT_MIN_MATCH {
let hash_head = StandardHashCalc::quick_insert_string(state, state.strstart);
macro_rules! first_four_bytes {
($slice:expr, $offset:expr) => {
u32::from_le_bytes($slice[$offset..$offset + 4].try_into().unwrap())
};
}

let str_val = {
let str_start = &state.window.filled()[state.strstart..];
first_four_bytes!(str_start, 0)
};
let hash_head = StandardHashCalc::quick_insert_value(state, state.strstart, str_val);
let dist = state.strstart as isize - hash_head as isize;

if dist <= state.max_dist() as isize && dist > 0 {
let str_start = &state.window.filled()[state.strstart..];
let match_start = &state.window.filled()[hash_head as usize..];

macro_rules! first_two_bytes {
($slice:expr, $offset:expr) => {
u16::from_le_bytes($slice[$offset..$offset + 2].try_into().unwrap())
};
}

if first_two_bytes!(str_start, 0) == first_two_bytes!(match_start, 0) {
if str_val == first_four_bytes!(match_start, 0) {
let mut match_len = crate::deflate::compare256::compare256_slice(
&str_start[2..],
&state.window.filled()[state.strstart + 2..],
&match_start[2..],
) + 2;

Expand All @@ -133,9 +137,10 @@ pub fn deflate_quick(stream: &mut DeflateStream, flush: DeflateFlush) -> BlockSt
}
}
}
lc = str_val as u8;
} else {
lc = state.window.filled()[state.strstart];
}

let lc = state.window.filled()[state.strstart];
state.bit_writer.emit_lit(StaticTreeDesc::L.static_tree, lc);
state.strstart += 1;
state.lookahead -= 1;
Expand Down
5 changes: 5 additions & 0 deletions zlib-rs/src/deflate/hash_calc.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,6 +38,11 @@ impl StandardHashCalc {
let slice = &state.window.filled()[string + Self::HASH_CALC_OFFSET..];
let val = u32::from_le_bytes(slice[..4].try_into().unwrap());

Self::quick_insert_value(state, string, val)
}

#[inline]
pub fn quick_insert_value(state: &mut State, string: usize, val: u32) -> u16 {
Copy link
Member

@folkertdev folkertdev May 28, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

for me, this needs an #[inline], otherwise level 2 regresses by a couple of percent. Interestingly I then see some improvements for the other compression levels too (mostly in instructions), maybe because different (better) inlining decisions are made.

let hm = Self::update_hash(0, val) as usize;

let head = state.head.as_slice()[hm];
Expand Down
Loading