2x1GB G.Skill PC3200ZX *AND* 2x1GB OCZ EB PC4000 Platinum OCing Report!

[bachus_anonym]

G.Skill 2x1GB PC3200ZX










SPi32M i Sandra @ 260x11=2860MHz 5-2-3-3.0-1T

SPi32M i Sandra @ 270x11=2971MHz 5-2-3-3.0-1T

SPi32M i Sandra @ 273x11=3003MHz 5-2-3-3.0-1T

273Mhz was absolutely maximum I could get this pair to pass even SPi32M calculation. Also, 2.9v (2.95v with DMM) seems to be the point where sticks stop returning any gains. Anything above that, it will only produce errors, even at known stable speeds.
Once I found out max 32M clock, it was time to stress them a little bit more. I dropped clock to 270Mhz and gave them a bit more torture...

1000% Memtest for Windows 1817MB used @ 270x11=2971MHz 5-2-3-3.0-1T

4 hours of SP2004 LargeFFT 1731MB used @ 270x11=2971MHz 5-2-3-3.0-1T

As you can see, all was fine and dandy... To finish off my test, I jacked up HTT, put on a 183 divider, got memory clock to 273MHz again and tried 32M one last time to ensure divider stability.

SPi32M i Sandra @ 297x11=3267MHz div183/DDR546 5-2-3-3.0-1T

Running a divider was totally plug-n-play. Just change it to 183 and viola.

*** Updated on Dec 18th, 2005 ***

I finally had a chance to bench those modules at 3D. As much as I believed I could tweak them to run 3D as high as 2D, it was illusive There must be something odd with early batches of Infineon CE-6 (???) because they are stable, but not stable at the same time I was able to pass 3D tests at 255MHz. Even 1MHz above caused lockups in ANY 3D test I threw at it. So, 27ish is max 2D stable, but 3D stability could be found lower - at 255Mhz.

3DMark2001 and 05 @ 255x11=2805MHz 5-2-3-3.0-1T

3DMark03 @ 255x11=2805MHz 5-2-3-3.0-1T

Aquamark3 @ 255x11=2805MHz 5-2-3-3.0-1T