Test Setup

Test System
CPU:	Intel E6300 Conroe 1.8 GHz, 2 MB Cache
Motherboard:	MSI P35 Neo2-FR, BIOS 1.8
Video Card:	PowerColor X800XL Pro 16 PCI-E
Harddisk:	Samsung P80 80 GB
Power Supply:	Ultra V-Power 450W
Software:	Windows XP SP2, Catalyst 7.6

Performance & Overclocking

The first thing done, when testing meory is simply running it at the rated speed. The ReaperX is meant to manage a very impressive CL4-4-3-12. Such a rating would already be interesting for a 2 GB kit, but we are talking about two 2 GB modules here. OCZ intends for these to run at 2.1V. This is also the setting we benched at out of the box. During our testing we noticed, that the memory easily manages the rated speed at default voltage of 1.8V on the MSI P35. The divider between CPU and memory was kept at 1:1, as the processor can handle at least 500 MHz with a multiplier of 6. The MSI board is also capable of such speed.

We suddenly hit a wall at around 920 MHz for the memory. No amout of voltage at this point seemed to change the result in our favor. Even raising the divider in combination with a lower FSB and looser timings did not yield any better speeds. Seems like our kit won't manage more than this - which is still quite impressive at 4-4-3-12.

Reducing the timings to 3-3-3 also meant reducing the FSB to 333 MHz. There are very few 2 GB kits out there that can manage CL3 at 800 MHz, so that this 4 GB kit can boot and function at 667 MHz is quite impressive. At such tight timings the memory did not manage to get far, no matter how much voltage was applied.

You will usually see CL5 benchmarks as well, but considering the memory clearly won't budge beyond ~920 MHz, which it can do at CL4 already, there is no reason to run any more benchmarks. It does however manage this speed at default voltage, which is good for boards which only offer FSB overclocking.

OCZ DDR2 Reaper X 800 MHz CL4-4-3-15 4GB Kit
CPU Clock Memory Ratio	Memory Speed	Memory Timings	Everest Read	Everest Write	Everest Latency	Quake 3 Timedemo	3DMark 2001SE	SuperPi Mod 1M
6 x 400 1:1	400 MHz	4-4-3-15 1.8V	7495 MB/s	6362 MB/s	67.5 ns	266 fps	26353	22.92 s
6 x 420 1:1	420 MHz	4-4-3-15 1.8V	7881 MB/s	6704 MB/s	64.0 ns	281 fps	27069	21.73 s
6 x 450 1:1	450 MHz	4-4-3-15 2.0V	8441 MB/s	7173 MB/s	59.9 ns	302 fps	28142	20.31 s
6 x 461 1:1	461 MHz	4-4-3-15 2.4V	8592 MB/s	7327 MB/s	58.2 ns	306 fps	28384	20.26 s
7 x 334 1:1	334 MHz	3-3-3-12 1.5V	6634 MB/s	6051 MB/s	74.8 ns	260 fps	25867	23.82 s
7 x 350 1:1	350 MHz	3-3-3-12 2.1V	6946 MB/s	6379 MB/s	71.6 ns	270 fps	26485	22.75 s
7 x 368 1:1	368 MHz	3-3-3-12 2.4V	7303 MB/s	6704 MB/s	68.0 ns	286 fps	27149	21.60 s
6 x 461 1:1	461 MHz	5-5-5-15 1.8V*	8498 MB/s	7209 MB/s	61.3 ns	304 fps	27685	20.43 s

The voltage scaling graph shows the limit of the memory, which we reach at ~2.1V already. And the fact that the memory does not yield any better results at CL5, this line has been left out. Please note that each individual module overclocks different, the results here can only be seen as an indicator of performance.