2GB OCZ Plat 2-3-2-5 vs 2GB G.Skill ZX - results

Jimbo Mahoney · 12 May 2006 at 23:09

OK, I've been playing with these two sets today. I'm using an Opty 146 and DFI Ultra D on 623-3 BIOS.

Here are my findings:

G.Skill - tops out 2-3-2-5 1T @ 212 Mhz, 2.7v

OCZ - tops out 2-3-2-5 1T between 205Mhz - 210 Mhz, 2.7v.

OK, so the G.Skill seems better, BUT

I couldn't complete a POST with the CPU @ 3.05Ghz using a 133 divider with the G.Skill, but I could with the OCZ (identical settings).

So I'm sticking with the OCZ, although to keep the CPU @ 3.05Ghz, I have to drop the timings to 2.5-3-3-8 1T :rolleyes:

I can only boot @ 2.8GHz using 2-3-2-5 timings.

I may play with the G.Skill on 2.5-3-3-8 timings tomorrow and see if I can use a better divider, like 140 to give 217Mhz on the mem...

Jimbo Mahoney · 13 May 2006 at 10:07

I'll be trying max frequency today.

I found neither set liked 2.8v for some reason. 2.7v was the optimal.

Jimbo Mahoney · 13 May 2006 at 11:19

OK, been testing using 2.5-3-3-8 1T:

OCZ - between 230 - 235 Mhz (230 is OK in memtest #5, 235 gives errors)

G.Skill - around 245 Mhz (245 is OK, 250 locks memtest)

And, I can boot using 3.05 Ghz on a 150 divider to run the memory @ 234Mhz, so now I'm sticking with the G.Skill

In Summary:

The G.Skill is the better set by about 5%.

To run tight timings with 2GB, you may have to drop the CPU clock = BAD (drop the memory timings instead)

Just to give an indication of (some) real-world difference in all of this:

3.05Ghz / 1GB BH5 @ 217Mhz 2-2-2-6 1T:
FEAR = 47/107/278
1M SuperPi = 28.2

3.05Ghz / 2GB OCZ @ 203 Mhz 2.5-3-3-8 1T:
FEAR = 45/105/272
1M SuperPi = 28.6

3.05Ghz / 2GB G.Skill @ 234Mhz 2.5-3-3-8 1T:
FEAR = 47/105/275
1M SuperPi = 28.5

Jimbo Mahoney · 13 May 2006 at 16:26

trojan698 said:
Nice work man, although it probably varies from kit to kit, naturally. One question, the 623 bios, is it better for overclocking?

Agreed.

623-3 BIOS - It's always been good for me, but I've not bothered with anything newer. I only noticed a difference in the earlier BIOS because the board was relatively new and with these newer A64 CPUs, everything seems more stable (memory controller revisions etc). I suspect the latest BIOS would be no different, apart from minor fixes.