Three PCI Express-Based SSDs: When SATA 6 Gb/s Is Too Slow

Test Procedure And Access Time Results

Test Setup Table

Swipe to scroll horizontally
System Hardware
CPUIntel Core i7-920 (Bloomfield), 45 nm, 2.66 GHz, 8 MB Shared L3 Cache
Motherboard (LGA 1366)Supermicro X8SAX, Revision: 1.0, Chipset Intel X58 + ICH10R, BIOS: 1.0B
ControllerIntel ICH10, SATA 6 Gb/s
RAM3 x 1 GB DDR3-1333, Corsair CM3X1024-1333C9DHX
HDDSeagate NL35 400 GB, ST3400832NS, 7200 RPM, SATA 1.5Gb/s, 8 MB Cache
Power SupplyOCZ EliteXstream 800 W, OCZ800EXS-EU
Performance MeasurementsAS-SSD 1.6.4067 PCMark Vantage
I/O PerformanceIOMeter 2008.08.18 Fileserver-Benchmark Webserver-Benchmark Database-Benchmark Workstation-Benchmark Streaming Reads Streaming Writes 4k Random Reads 4k Random Writes
System Software & Drivers
Operating SystemWindows 7 Ultimate Edition

Test Procedure

We added AS SSD as a new benchmark, and we decided to test all devices right out of the box in a fresh state, as well as preconditioned. Here is the procedure:

  1. AS SSD test run 1
  2. Pre conditioning: Iometer random write run (128 KB block size, two hours)
  3. Run full Iometer test suite
  4. Run PCMark Vantage
  5. AS SSD test run 2

Results: Access Time

We provide access time results for the sake of completeness, as these are hardly relevant in most environments. However, it is clear that the Fusion-io drives have advantages, as they do not have to work through SATA and hence show lower latencies. It is also obvious that heavily-used drives show a clear degradation in performance. And lastly, there is not a relevant difference in access time between a conventional SSD like the OCZ Vertex 2 and the PCI Express-based solutions.