Performance Improvement in Kernels by Guiding Compiler Auto-Vectorization Heuristics

Vectorization support in hardware continues to expand and grow as well we still continue on superscalar architectures. Unfortunately, compilers are not always able to generate optimal code for the hardware;detecting and generating vectorized code is extremely complex. Programmers can use a number of tools to aid in development and tuning, but most of these tools require expert or domain-specific knowledge to use. In this work we aim to provide techniques for determining the best way to optimize certain codes, with an end goal of guiding the compiler into generating optimized code without requiring expert knowledge from the developer. Initally, we study how to combine vectorization reports with iterative comilation and code generation and summarize our insights and patterns on how the compiler vectorizes code. Our utilities for iterative compiliation and code generation can be further used by non-experts in the generation and analysis of programs. Finally, we leverage the obtained knowledge to design a Support Vector Machine classifier to predict the speedup of a program given a sequence of optimization underprediction, with 82% of these accurate within 15 % both ways