o2bench

Raw C-Style array storage vs C++ STL containers (std::array and std::vector), before and after compiler optimizations -O2 and cache alignment alignas(64)

Purpose

To compare the compiler's ability to optimize memory access operations based on the data structure.

Results

Key Observations

std::vector has the highest allocation time both before and after optimization, likely due to additional overhead from dynamic allocation

std::array (aligned) performs best in both scenarios, demonstrating the benefits of memory alignment

Compiler optimizations (-O2) provide substantial performance improvements for all data structures

The relative improvement from optimization is similar across all data structures (81.4%, 79.7%, 78.9%, 77.9%)

Optimization Impact

Average time reduction: 79.5%

All data structures show significant performance gains with compiler optimizations, but the fundamental performance characteristics of each container type are preserved.

Even with optimizations, std::vector remains significantly slower than array-based alternatives for this particular initialization pattern.

Compile

mkdir build
cd build 
cmake .. 
make 

Run

../run_tests.sh