Bencmark Results: Synthetics
We wouldn’t expect to see any noticeable difference in CPU or memory performance between two systems that differ only in graphics, yet enthusiasts looking to break away from the desk will be interested to see how much performance must be sacrificed.
Synthetic benchmarks represent the “worst case” for the mobile CPU, since these benchmarks are multi-threaded and optimized for Intel’s quad-core devices. That’s because Intel Turbo Boost doesn’t work as aggressively when all four cores are occupied and the mobile processors have a lower standard frequency. As stated on the previous page, more optimistic turbo multipliers for the mobile processors allow them to frequency-match the desktop part whenever fewer processing threads are in use.
A 35% lower clock speed gives way to a performance loss of around 40% for the mobile CPU in this fully threaded benchmark. Our real-world benchmarks have shown that many applications are poorly threaded, though the industry is advancing.
The desktop platform also has a triple-channel memory controller, yet once again we remind readers that our real-world benchmarks have not previously shown a significant benefit in applications.
PCMark results are influenced by the fact that AVADirect uses high-performance solid-state disks. Our current application benchmark set has not shown more than the slightest preference for the lower seek times that these devices provide, but in the real-world, booting Windows and loading the levels in your favorite games will happen faster.
3DMark shows some inconsistency, though a look at its GPU results (not shown) confirm that a desktop Radeon HD 5850 is twice as powerful as a Mobility Radeon HD 5870.