Identify causes of Wasm-R3 failures on seven applications

Identify and diagnose the underlying causes that prevented Wasm-R3 from producing accurate replay benchmarks for seven specific real-world WebAssembly web applications, and classify the failure modes to inform fixes or extensions to the approach.

Background

In the evaluation of Wasm-R3 across 43 real-world WebAssembly web applications, the authors successfully produced accurate replay benchmarks for 27 applications. However, they encountered failures in 16 cases, five due to implementation limitations and four due to dependencies.

For the remaining seven failures, the authors explicitly state that they could not determine the cause and are investigating. Understanding these failure causes is necessary to improve Wasm-R3’s robustness and coverage across diverse applications.