Serial ATA performance
The Serial ATA disk controller is one of the most important components of a modern core logic chipset, so we threw each platform a selection of I/O-intensive storage tests using a Western Digital Raptor WD1500ADFD. Testing was conducted with the SB700 running in AHCI mode, which is necessary for Native Command Queuing, using the latest drivers supplied by AMD.
AMD's older SB600 south bridge had numerous problems running in AHCI mode. It required a Vista hotfix to work properly, and you'd need an external storage controller to get the operating system installed. The hotfix fortunately isn't required with the SB700. However, if you want to get Vista running in AHCI mode, you have to first install the OS in IDE mode, then swap the hard drive to an auxiliary storage controller running on the same system, switch the SB700 to AHCI mode and install the necessary drivers, and then move the hard drive back to the south bridge. It's ridiculous to ask users to jump through hoops just to enable AHCI. AMD should have worked closer with Microsoft to prevent a documented problem with an old south bridge chip from afflicting the new SB700.
We'll begin our storage tests with IOMeter, which subjects our systems to increasing multi-user loads. Testing was restricted to IOMeter's workstation and database test patterns, since those are more appropriate for desktop systems than the file or web server test patterns.
The SB700's Serial ATA controller really is more of the same. Like the SB600 before it, the SB700 appears to have problems scaling performance as the number of simultaneous I/O requests increases. We only tested with a single driver revision here, but it's the latest one. Clearly, our G35 Express board's ICH9R south bridge is doing a much better job handling IOMeter workloads.
As you can see, response times are much lower on the G35 than with the 780G, particularly as the number of simultaneous I/O requests scales up.
At least CPU utilization is reasonable for the 780G, although given its poor throughput, that's not saying much. We've seen driver revisions that fix the SB600's performance scaling problems have a detrimental impact on CPU utilization, and since the two share essentially the same SATA underpinnings, we'd expect similar results from the SB700.