| OCZ Technology PC3-10666 ReaperX HPC 2GB DDR3 Kit
|
Configuration:
The OCZ Technology 2GB PC3-10666 ReaperX HPC Enhanced Bandwidth DDR3 Dual Channel Memory Kit was installed in a system with the following components for this review:
» Intel Core 2 Duo E6600 dual core processor
» ASUS P5KC P35 ATX motherboard
» Seagate Barracuda SATA II 7200 RPM 80GB hard drive
» Zippy GSM-6600P(G1) 600W power supply
» eVGA GeForce 8800GTS G92 512MB video card
» Microsoft Windows XP Pro
The module's specifications indicate default timings of 6-5-5-18 at 1333MHz, with a voltage of 1.85V. Upon first boot the test system's BIOS identified them as actually being 9-9-9-25 @ 1333MHz and I left them like that for the first boot into Windows. This type of discrepancy is not an issue, and is put in place for compatibility purposes.
 
The screenshots above were taken from CPU-Z, with the image on the left showing the default settings described above. The image on the right shows the SPD tab which reveals three different profiles have been programmed. The three profiles include speeds of 1066MHz, 1244MHz, and 1422MHz, and some very loose timings. While the speeds seem rather arbitrary and the timings aren't very aggressive, this data will be serve as a good point of reference, rather than guessing when I try to overclock the kit past the rated speed of 1333MHz.
The BIOS settings were now adjusted to be as close to specification possible; DDR3 1333MHZ, with timings of 6-5-5-18, and 1.80V (due to the limitation on the motherboard only allowing increments of 0.10V). Back in Windows, CPU-Z displayed the settings I just entered in the BIOS. To ensure the system was stable I ran a 30 minute stress test of OCCT, which completed successfully.
Overclocking and Optimizing:
Extreme enthusiasts may be able to take their high end systems well beyond any overclocking feats accomplished here, but I see this as a kit that might appeal to gamers. The low latencies offered at reasonably good speeds should provide a plug-and-play boost to those who don't necessarily want to spend the time or energy with a serious overclock. That said, I will take the base performance and see just how much better it can get at higher frequencies!
The processor used in this review wasn't a member of the 1333MHz bus speed family, so to reach that plateau I had to overclock it to 333x9 to give it a speed of 3.0GHz and FSB of 1333MHz. The CPU has no problems with this speed and is 100% stable.
Even though this memory kit is rated for DDR3 1333MHz, I hope to get it stable at elevated speeds with relaxed timings (if necessary). Before I got to high speed testing, I ran this kit at the specified speed of another 2GB DDR3 kit for comparison. The Aeneon XTUNE 2GB PC3-10600 DDR3 Dual Channel Memory Kit is rated at 1333MHz and timings of 8-8-8-15. That sounded like something the OCZ Technology kit should be able to do, but I couldn't get it to pass the OCCT test (but I ran benchmarks anyway).
Testing how low the timings could be set while at default speed with the OCZ Technology 2GB PC3-10666 ReaperX HPC DDR3 Dual Channel Memory Kit didn't meet with much luck. Lowering any of the first three numbers [CAS-TRCD-TRP] resulted in an "overclocking failure" with the motherboard. I was able to put TRAS lower, but was met with lots of instability in testing (SuperPI wouldn't pass a 1M test). The timings on this kit are fairly tight to begin with, so it may be safe to guess that there isn't any room to improve them.
Enough time was spent at the default speed for this kit, and it was time to speed things up. I went back into the BIOS and went straight to a bus speed of 375MHz and put the CPU multiplier on 8, which resulted in a frontside bus speed of 1500MHz while still being around 3GHz for total CPU speed. For the timings, I used the data found in the SPD tab of CPU-Z, 8-7-7-27. Creeping much above 1500MHz was met with instability, which could easily be the maximum the processor or motherboard might have been able to handle, and not a limit with the memory. Regardless, at 1500MHz these modules put up some nice results in benchmarks, which we will take a look at in the next section.
|
|
|
|
