Efficient SIMD Code Generation for Runtime Alignment and Length Conversion

When generating codes for today’s multimedia extensions, one of the major challenges is to deal with memory alignment is-sues. While hand programming still yields best performing SIMD codes, it is both time consuming and error prone. Compiler tech-nology has greatly improved, including techniques that simdize loops with misaligned accesses by automatically rearranging mis-aligned memory streams in registers. Current techniques are ap-plicable to runtime alignments, but they aggressively reduce the alignment overhead only when all alignments are known at com-pile time. This paper presents two major enhancements to the state of the art, improving both performance and coverage. First, we propose a novel technique to simdize loops with runtime alignment nearly as efficiently as those with compile-time misalignment. Runtime alignment is pervasive in real applications because it is either part of the algorithms, or it is an artifact of the compiler’s inabil-ity to extract accurate alignment information from complex ap-plications. Second, we incorporate length conversion operations, e.g., conversions between data of different sizes, into the align-ment handling framework. Length conversions are pervasive in multimedia applications where mixed integer types are often used. Supporting length conversion can greatly improve the coverage of simdizable loops. Experimental results indicate that our runtime alignment technique achieves a 19 % to 32 % speedup increase over prior art for a benchmark stressing the impact of misaligned data. We also demonstrate speedup factors of up to 8.11 for real bench-marks over sequential execution.