Exploiting Vector Processing in Dynamic Binary Translation

Auto vectorization techniques have been adopted by compilers to exploit data-level parallelism in parallel processing for decades. However, since processor architectures have kept enhancing with new features to improve vector/SIMD performance, legacy application binaries failed to fully exploit new ... View More