Skip to content

Make FramePadding.countLeadingZeroBytes faster #190

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

Open
MahdiBM wants to merge 3 commits into
base: main
Choose a base branch
from
+21 −4
Open

Make FramePadding.countLeadingZeroBytes faster #190

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
753e37a
Make `FramePadding.countLeadingZeroBytes` faster
MahdiBM Oct 1, 2026
0666cbb
Address review comments
MahdiBM Oct 1, 2026
7301108
Increase byte-count in tests to something more than 32 (86/87)
MahdiBM Oct 1, 2026
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
17 changes: 17 additions & 0 deletions Sources/SwiftNetwork/QUIC/QUICFrame.swift
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -773,8 +773,24 @@ struct FramePadding: ~Copyable, QUICFrameProtocol {
private static func countLeadingZeroBytes(_ bytes: RawSpan) -> Int {
let count = bytes.byteCount
let wordSize = MemoryLayout<UInt64>.size
let vectorSize = 32

@MahdiBM MahdiBM Oct 1, 2026

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

I tried 16/64 as well, 32 seems to be the sweet spot.

var offset = 0

// 1. Speed through the zeros via vectorization.
while offset &+ vectorSize <= count {
var orResult: UInt8 = 0
// This loop is auto-vectorized.
for idx in 0..<vectorSize {
orResult |= bytes.unsafeLoad(fromUncheckedByteOffset: offset &+ idx, as: UInt8.self)
}
if orResult == 0 {
offset &+= vectorSize
} else {
break
}
}

// 2. Read the rest of the bytes in chunks of words.
while offset &+ wordSize <= count {
let word = bytes.unsafeLoadUnaligned(fromByteOffset: offset, as: UInt64.self)

Expand All @@ -785,6 +801,7 @@ struct FramePadding: ~Copyable, QUICFrameProtocol {
}
}

// 3. Scan the remaining bytes one by one.
while offset < count, bytes[offset] == 0x00 {
offset &+= 1
}
Expand Down
8 changes: 4 additions & 4 deletions Tests/QUICTests/QUICFrameTests.swift
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -106,8 +106,8 @@ class QUICFrameTests: XCTestCase {

func testPaddingInitStopsAtFirstNonZeroByte() throws {
// Offset 0 is the frame type byte, so start at 1
for sentinelOffset in 1..<23 {
var bytes = [UInt8](repeating: 0x00, count: 23)
for sentinelOffset in 1..<87 {
var bytes = [UInt8](repeating: 0x00, count: 87)
bytes[sentinelOffset] = 0xff

var frame = Frame(copyBuffer: bytes)
Expand All @@ -118,10 +118,10 @@ class QUICFrameTests: XCTestCase {
}

func testPaddingInitAllZeroes() throws {
let bytes = [UInt8](repeating: 0x00, count: 23)
let bytes = [UInt8](repeating: 0x00, count: 87)
var frame = Frame(copyBuffer: bytes)
let paddingFrame = try FramePadding(frame: &frame, packetNumberSpace: .applicationData)
XCTAssertEqual(paddingFrame.extraPadding, 22)
XCTAssertEqual(paddingFrame.extraPadding, 86)
frame.finalize(success: true)
}

Expand Down
Loading