Architectural Support for Server-Side PHP Processing

Abstract

PHP is the dominant server-side scripting language used to implement dynamic web content. Just-in-time compilation, as implemented in Facebook's state-of-the-art HipHopVM, helps mitigate the poor performance of PHP, but substantial overheads remain, especially for realistic, large-scale PHP applications. This paper analyzes such applications and shows that there is little opportunity for conventional microarchitectural enhancements. Furthermore, prior approaches for function-level hardware acceleration present many challenges due to the extremely flat distribution of execution time across a large number of functions in these complex applications. In-depth analysis reveals a more promising alternative: targeted acceleration of four fine-grained PHP activities: hash table accesses, heap management, string manipulation, and regular expression handling. We highlight a set of guiding principles and then propose and evaluate inexpensive hardware accelerators for these activities that accrue substantial performance and energy gains across dozens of functions. Our results reflect an average 17.93% improvement in performance and 21.01% reduction in energy while executing these complex PHP workloads on a state-of-the-art software and hardware platform.