Test Setup

Test System
CPU:	AMD Athlon64 3000+ Venice
Motherboard:	DFI LanParty NF4, Bios 5.10-2 Fix
Memory:	2x 1024 MB G.SKILL F1-4000BIU2-2GBHV PC4000
Video Card:	ATI Radeon X850 Pro PCI-E
Harddisk:	Maxtor Diamondmax 160GB
Power Supply:	HEC PurePower 475
Software:	Windows XP SP2, Catalyst 6.1

Performance

Our first benchmark shows how the memory performs at standard timings running 200 MHz.
The next test checks how far you can actually overclock at a standard voltage of 2.6V which may be what you have on motherboards that do not support DDR voltage increase.
In the third and fourth test we tried tightening the timings. It is not possible to run x-2-x-x. The fastest CL2.5 setting is 2.5-3-2-5 and for CL2 it is 2-3-2-5. G.SKILL's memory performs quite well here, I find it very important that memory can run CL2 with good timings at least at 200 MHz + some headroom.
The next two tests check if there is any increase in overclocking with increased voltage, this does not seem to be the case with this memory. Well, there is an increase but a very small one.
The last test "JEDEC DDR-400A" is for comparison with a generic DDR module running at JEDEC standard timings.

What I noticed during overclocking at CL3 was that memory testing and stressing applications ran fine easily up to 270 MHz, but 3DMark was always unstable and crashed with VPU Recover from the ATI drivers. So I had to reduce the clocks step by step and ended up at the 260 MHz range you see in the results. I have asked G.SKILL to check with another memory kit if they can replicate it or this is an isolated issue with my memory.

G.SKILL F1-3200PHU2-2GBHV PC3200
CPU Clock & Memory Ratio	Memory Speed	Memory Timings	Everest Read	Everest Write	Everest Latency	Quake 3 Timedemo	3DMark 2001SE	SuperPi Mod 1M
9 x 200 1:1	200 MHz	3-3-2-8 2.6V	5577 MB/s	1955 MB/s	55.3 ns	249.5 fps	19634	47.16 s
9 x 258 1:1	258 MHz	3-3-2-8 2.6V	7194 MB/s	2662 MB/s	43.0 ns	321.9 fps	23960	36.50 s
9 x 240 1:1	240 MHz	2.5-3-2-5 2.8V	6378 MB/s	2547 MB/s	43.2 ns	305.9 fps	23484	38.91 s
9 x 218 1:1	218 MHz	2-3-2-5 2.8V	5847 MB/s	2336 MB/s	46.5 ns	279.7 fps	21869	42.64 s
9 x 258 1:1	258 MHz	3-3-2-8 2.8V	7194 MB/s	2662 MB/s	43.0 ns	321.9 fps	23960	36.50 s
9 x 260 1:1	260 MHz	3-3-2-8 3.0V	7256 MB/s	2645 MB/s	42.8 ns	322.0 fps	24094	36.47 s
JEDEC DDR-400A	200 MHz	2.5-3-3-8 2.6V	5353 MB/s	2044 MB/s	52.6 ns	252.3 fps	19933	46.74 s

For easier comparison with other modules, we set the maximum voltage required for the best result (here 3.0V; max 3.1V) and tested until we found the highest clock frequency and fastest timings for this memory. The benchmarks Everest Read, Everest Write and Quake 3 were run. We then calculated the performance increase in percent compared to some standard DDR-400 memory running at JEDEC standard timings (2.5-3-3-8). The average percentage of the three benchmarks is listed in following table: