Oxicloud/doc/LTO-OPTIMIZATIONS.md

Link Time Optimization (LTO) in OxiCloud

Overview

OxiCloud uses Link Time Optimization (LTO) to significantly improve runtime performance. LTO is a technique that allows the compiler to perform optimizations across module boundaries during the linking phase, which can lead to better inlining, dead code elimination, and overall more efficient binaries.

Implemented Optimizations

This project uses the following optimization settings:

Release Profile

[profile.release]
lto = "fat"         # Full cross-module optimization
codegen-units = 1   # Maximum optimization but slower compile time
opt-level = 3       # Maximum optimization level
panic = "abort"     # Smaller binary size by removing panic unwinding
strip = true        # Removes debug symbols for smaller binary

Development Profile

[profile.dev]
opt-level = 1       # Light optimization for faster build time
debug = true        # Keep debug information for development

Benchmark Profile

[profile.bench]
lto = "fat"         # Full optimization for benchmarks
codegen-units = 1   # Maximum optimization
opt-level = 3       # Maximum optimization level

Performance Improvements

The optimizations typically result in:

1. Smaller binary size: Removing unused code and metadata
2. Faster execution: Better inlining and code optimizations
3. Reduced memory usage: More efficient code layout and execution

LTO Options Explained

• fat: Also known as "full" LTO, performs optimizations across all crate boundaries. Maximum optimization but longest compile time.
• thin: A faster version of LTO that trades some optimization for compile speed. Good for development.
• off: No cross-module optimization.

Build Time Impact

While LTO provides runtime performance benefits, it increases compilation time. For OxiCloud, we chose:

• Development builds: Minimal LTO (opt-level = 1) for faster iteration
• Release builds: Full LTO for maximum end-user performance
• Benchmark builds: Full LTO to measure actual optimized performance

Measuring the Impact

To measure the impact of these optimizations, run our benchmarks:

# Run benchmarks with all optimizations
cargo bench

# Compare with non-optimized build (remove for comparison only)
RUSTFLAGS="-C lto=off" cargo bench

When to Adjust Settings

Consider adjusting these settings if:

1. You need faster compile times during development
2. You're experiencing unexpected runtime behavior
3. You want to experiment with optimization/binary size tradeoffs

For most users, the default settings should provide a good balance of performance and usability.