Benchmarking Wild vs. Mold

Mold recently published linker benchmarks showing Wild performing substantially slower than its own previously published results, prompting Wild's developer to investigate the discrepancies. The differences stem from varying benchmark configurations, including whether the output file is deleted between runs, the filesystem used (tmpfs vs. ext4), and whether the --no-fork flag is passed, which significantly impact linker performance. When Wild reproduced Mold's benchmarks using identical settings on an M1 machine, the results aligned closely with Mold's findings, demonstrating that the earlier performance gap was due to experimental setup differences rather than actual performance regression.