|
A Closer
Look
The OCZ XTC memory sports a
shiny new heatspreader. XTC stands for Xtreme Thermal Convection
and has added to the debate on whether heatspreaders help or hinder.
We aren't going to add to the debate one way or the other, but I like
the new look.
Compared to the older style
heatspreader, the new XTC is much lighter. The honeycomb mesh
takes a bunch of the weight away while allowing for greater airflow to
the memory chips. From a purely aesthetic point of view they are a
very nice upgrade.
Testing
and Overclocking
Currently if you want
to run DDR2, you have to have an Intel processor. Of course that
will all change in a few months, but for now Intel will have to do.
Intel's current chipsets are ideal for testing high speed DDR2.
I have chosen the P5WD2 Premium board that sports a 955 chipset. I
know this board is capable of 280MHz+ front side bus. The CPU I
have is limited to 260~270MHz FSB, but it should be plenty to get the
most out of the memory. Let's see what this memory can really do.
| Motherboard |
Asus P5WD2 Premium |
| CPU |
P4 640 LGA775 |
| Video |
ATI X800XT |
| Memory |
1GB PC2-6400 Platinum
Edition XTC |
| Power Supply |
AeroCool ZerodBA 620w |
| Storage |
400GB Seagate HDD |
| Optical |
Lite-On 16X DVD+/-RW +DL |
| OS |
Windows XP SP2 |
This is what we are
starting with. It's pretty kewl to see 400MHz in the frequency
field and know that you didn't even overclock to achieve such a
speed. With that as a starting point, I jumped into the BIOS
to see what we could get. The highest overclock I was able to
get was 460MHz that's DDR920. It took a front side bus speed
of 230MHz to get that. At 240Mhz front side bus the divider
reset back to 400Mhz, but provided a nice performance gain.
|
Bus Speed
(MHz) |
DDR Speed |
Memory
Timings |
Sandra
INT |
Sandra FPU |
Everest Read |
Everest Write |
|
200 |
800 |
4-5-4-15 |
5116 |
5100 |
6148 |
2360 |
|
230 |
920 |
4-5-4-15 |
5859 |
5861 |
7053 |
2659 |
|
240 |
800 |
4-5-4-15 |
6036 |
6006 |
7188 |
2662 |
This really illustrates
that 800MHz does not always equal 800MHz. Running 800MHz at a
200MHz bus speed provided much weaker scores than 800MHz at 240MHz
bus.
 |
 |
|
