Software prefetching for indirect memory accesses
- 1 February 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 305-317
Abstract
Many modern data processing and HPC workloads are heavily memory-latency bound. A tempting proposition to solve this is software prefetching, where special non-blocking loads are used to bring data into the cache hierarchy just before being required. However, these are difficult to insert to effectively improve performance, and techniques for automatic insertion are currently limited. This paper develops a novel compiler pass to automatically generate software prefetches for indirect memory accesses, a special class of irregular memory accesses often seen in high-performance workloads. We evaluate this across a wide set of systems, all of which gain benefit from the technique. We then evaluate the extent to which good prefetch instructions are architecture dependent. Across a set of memory-bound benchmarks, our automated pass achieves average speedups of 1.3x and 1.1x for an Intel Haswell processor and an ARM Cortex-A57, both out-of-order cores, and performance improvements of 2.1x and 3.7x for the in-order ARM Cortex-A53 and Intel Xeon Phi.Keywords
This publication has 20 references indexed in Scilit:
- SQRL2014
- Meet the walkers2013
- Compiler-Based Data Prefetching and Streaming Non-temporal Store Generation for the Intel(R) Xeon Phi(TM) Coprocessor2013
- When Prefetching Works, When It Doesn’t, and WhyACM Transactions on Architecture and Code Optimization, 2012
- Improving hash join performance through prefetchingACM Transactions on Database Systems, 2007
- CHALLENGES IN PARALLEL GRAPH PROCESSINGParallel Processing Letters, 2007
- S12---The HPC Challenge (HPCC) benchmark suite2006
- LLVM: A compilation framework for lifelong program analysis & transformation2004
- Compiler-based prefetching for recursive data structures1996
- SPAID: software prefetching in pointer- and call-intensive environments1995