SIMD processor units have become ubiquitous. Using SIMD instructions is the key for performance for many applications. Modern compilers have made immense progress in generating ecient SIMD code. However, they still may fail or SIMDize poorly, due to conservativeness, source complexity or missing capabilities. When SIMDization fails, programmers are left with little clues about the root causes and actions to be taken.
Our proposed guided SIMDization framework builds on the assembly-code quality assessment toolkit MAQAO to analyzes binaries for possible SIMDization hindrances. It proposes improvement strategies and readily quanti es their impact, using in vivo evaluations of suggested transformation. Thanks to our framework, the programmer gets clear directions and quanti ed expectations on how to improve his/her code SIMDizability. We show results of our technique on TSVC benchmark.