Conclusion

While we have yet to see any device capable of defining the performance limits of USB 3.0, benchmarks prove it’s a huge step up from USB 2.0, even when using a average desktop hard drive. USB 3.0 is able to match the eSATA controller against which it initially competes, and its 5.0 Gb/s limit will continue to remain competitive, even after 6.0 Gb/s transfers are applied to eSATA.

Offering around 50% more amperage over the same power wires as the USB 2.0 interface it shares, USB 3.0 looks to become the de-facto standard for high-speed portable devices. The marketing power of the USB name, along with its shared connector and compatibility with non-ATA devices, will likely relegate the competing eSATA standard to stationary backup devices.

Asus’ solution appears to be the most elegant option because it doesn’t steal pathways from the x16 graphics card slot, but instead relies on a PLX bridge to convert four of the chipset’s 2.5 Gb/s pathways to two 5.0 Gb/s pathways. Yet Gigabyte managed to edge out Asus in write performance by taking its 5.0 Gb/s pathway directly from the CPU, eliminating any middle parts (like the DMI interface connecting Core i7 to P55) that could slow the interface down, while also limiting the PCIe slot to x8 mode. This tradeoff can be blamed directly on LGA 1156 platform limitations, and high-end buyers who want the best of everything should instead consider X58-based solutions such as Gigabyte’s X58A-UD7.

We knew Intel's incorporation of PCI Express into its Lynnfield design would bite us in the butt somewhere, but we always thought it'd be the performance of next-generation graphics cards split over a pair of x8 slots. Such is the double-edged sword of integration.

We reserve criticism of USB 3.0’s overall performance until someone is able to supply a component with significantly greater bandwidth, but we still have other concerns. Chief among these is the lack of standardization for USB 3.0 front-panel breakout cables, since the convenience of front-panel access is key to its marketability. We have little doubt that major system manufacturers such as HP and Dell are designing new cases right now with proprietary case-to-motherboard connections, and that could spell disaster for independent builders, unless retail-component manufacturers can quickly agree on a breakout-header standard. It took several years before retail motherboards came with standardized front-panel USB connections, and a repeat performance by the industry is something the custom-built market might not be able to tolerate.