Test Settings: Lowest Stable Latencies
Because of the previously mentioned "Boot Strap" limitations, we had to select different FSB speeds to test DDR3-1333 and DDR3-1600 data rates. But how could we do that without throwing the rest of our speeds off?
Lacking any 5:3 DRAM to FSB clock ratio required to test DDR3-1333 with an FSB-1600 processor, we instead must compare DDR3-1333 to DDR3-1066 using FSB-1333, and also compare DDR3-1600 to DDR3-1066 using FSB-1600.
Only two CPU speeds correspond to both FSB-1333 and FSB-1600: These are 2.00 GHz CPU clock and 4.00 GHz CPU clock. Since this started out as an overclocking article, the 4.00 GHz speed was selected. The CPU multipliers needed to reach 4.00 GHz at FSB-1333 and FSB-1600 is 12 x 333 MHz and 10 x 400 MHz, respectively.
|Latency Test System Hardware|
|Motherboard||Asus Maximus Extreme Rev. 2.01GIntel X38, BIOS 0501 (10/30/2007)|
|Socket 775 Processor||Intel Core 2 Extreme QX9770 "Yorkfield"(FSB-1600, 45 nm, 3.20 GHz, 12 MB L2 Cache)|
|Hard Drive||Western Digital WD1500ADFD-00NLR1, Firmware: 20.07P20150 GB, 10,000 RPM, 16 MB cache, SATA/150|
|Graphics Card||Foxconn GeForce 8800GTX, P/N: FV-N88XMAD2-ODNVIDIA GeForce 8800GTX - 768 MB|
|Power Supply||OCZ GameXStream OCZ700GXSSLI - 700W|
|System Software & Drivers|
|OS||Windows XP Professional 5.10.2600, Service Pack 2|
|DirectX Version||9.0c (4.09.0000.0904)|
|Platform Drivers||Intel INF 188.8.131.529|
|Graphics Driver||NVIDIA Forceware 163.75|
Since the Asus Maximus Extreme proved instrumental in diagnosing the boot strap issue, it was retained for memory latency testing.
Four-core processors use memory a little more effectively than dual cores and our highest latency test speed of DDR3-1600 matches the highest memory ratio afforded FSB-1600 processors. We used the only FSB-1600 processor available, which was Intel's Yorkfield-based Core 2 Extreme QX9770.
Game benchmarks are significantly limited by graphics performance, so we included a powerful GeForce 8800GTX from Foxconn.