Results: SiSoftware Sandra
Gigabyte’s small lead in Sandra's Arithmetic module appears strange, especially since we locked the board in at an exact 100 MHz base clock and made sure that its full range of Turbo Boost ratios were represented at various loads. EVGA’s stumble in Sandra's Cryptography test appears equally out of place, and retesting confirmed that this data was consistent.
Suddenly, EVGA's finish in the Cryptography test makes a lot more sense. Poor memory performance curbs the rate at which our CPU can churn through AES-related instructions. Our Z87 Stinger operates at the same DDR3-1600 CAS 9 defaults as the other boards, but perhaps EVGA is playing with other timings to boost memory stability?