Benchmark Results: Synthetics
Our first result is already quite interesting. Mainly, we see Turbo Boost yielding minimal, if any, performance gain in the PCMark Vantage overall suite score. Meanwhile, overclocking catapults both processors forward. Turbo Boost is much more effective in the TV and Movies and Productivity tests, though overclocking is able to make an even greater impact in both cases, as we’d expect.
Interesting also is that there’s minimal benefit to Hyper-Threading—a pattern that continues throughout the suite’s sub-tests. Of course, because these are all based on features built into Windows 7, it’s possible that the operating system’s components simply aren’t as well-optimized for Hyper-Threading as Microsoft might otherwise have us believe.
Turbo Boost makes very little difference in the overall suite score of 3DMark Vantage, but it’s at least a measurable advantage in the CPU suite. It makes no difference, really, in the GPU measurement. But for that matter, neither does manual overclocking. We’re not surprised, though. Both of these CPUs are fast enough to keep from bottlenecking our single Radeon HD 5850, so we’d expect very little gaming performance increase at the hands of additional clock rate.
This synthetic is able to expose a massive performance increase via Hyper-Threading in the CPU suite, which is matched by a boost at the hands of overclocking, taking the quad-core i5-750 at 4 GHz up to the performance level of a stock i7-860. We’ll see how those results bear out in the real-world testing.
The clearest gains come from overclocking in the Whetstone iSSE3 test, where Hyper-Threading doesn’t make much of a difference. The Dhrystone iSSE4.2 metric, however, shows Intel’s Core i5-750 at 4 GHz unable to match a Core i7-860 running at its stock 2.8 GHz.
The multimedia suite similarly doesn’t show us getting much at all from Turbo Boost technology, but significant gains via overclocking both CPUs to 4 GHz. Hyper-Threading plays a significant role in both components of this one, which is interesting because we’d expect Turbo Boost to be a more impactful capability in the real-world benchmarks.
At stock speeds, memory bandwidth is largely unchanged regardless of whether Turbo Boost is enabled or disabled. This is because Turbo Boost only affects the processor’s clock multiplier, leaving BCLK settings (and thus, the related memory ratios) alone.
But when we overclock, increasing the BCLK (since these CPUs employ locked multipliers), the corresponding memory transfer rate goes up as well, yielding the higher memory bandwidth numbers we see in SiSoftware’s Sandra 2010 Bandwidth benchmark.