4. Tuning items

This chapter describes ten of forty-four tuning items as samples. The details of each thing are described in Section 3.1.1 (Tuning items).

• 4.1. SIMDization of division operations and suppression of SIMDization for loops with small iteration counts
• 4.2. Reducing load and store operations of data by loop unrolling
• 4.3. SIMDization by loop collapse
• 4.4. Changing the access direction of arrays
• 4.5. SIMDization by SVE ACLE
• 4.6. Built-in prefetch
• 4.7. Moving division operations to outside of the loop, and applying SIMDization to the division operations
• 4.8. Moving invariant expressions to outside of the loop
• 4.9. Loop unrolling manually instead of OCLs
• 4.10. Improving the memory placement of two-dimensional arrays for sequential access