Cross posted at XS
I'll make one thing clear before I get into results... the VX2 is obviously more capable of kicking the Gold series butt, however it was handy so I decided to use it.
The Gold PC2-8800 was announced the other day, although pricing is currently unknown. To throw a figure out there, I guesstimate that it will retail for around ~$350US. This may seem somewhat low for a top rate kit, but the Gold series signifies 'value', and not overclocking ability. That being said, I can't see this kit being released for more than $350 despite the rated frequencies being higher than anything else out there. Don't quote me please.
The PC2-8800 are based on D9GMH, so I had figured I knew what I was getting into. Turns out I didn't have a sweet clue. As I found out rather quickly, this is some fussy crap.
Overclocks
DDR2-800 4-4-4-12 2.0v 1T (2.40GHz, 1.25v)
DDR2-900 4-4-3-12 2.1v 2T (2.70GHz , 1.45v)
DDR2-1100 5-6-6-15 2.2v 2T (2.75GHz, 1.45v)
DDR2-1100 5-5-4-5-12 2T (2.75GHz, 1.45v)
DDR2-1120 5-5-4-5-12 2T (2.80GHz, 1.45v)
Oddly enough, these sticks were -very- fussy with CAS4. At DDR2-1000 4-4-4, they gave way to error in MemTest very quickly. Also, at the second listed overclock, even though MemTest proved stable in addition to other synthetic benchmarks, Super Pi would -not- compute past two loops. Even with more voltage, it didn't make a single difference. DDR2-800 CL4 worked fine, but I'm sure it's due to the loose frequency. As the freq increased, CL5 was pretty much forced.
The first three results use tRC 22, while the final two were at 12. No real reason for this, but I usually wait until I get close to the max possible speed before tweaking that, or tRAS.
Results
System:
AM2 Windsor 4600+
ASUS M2N32-SLI Deluxe (701 BIOS)
Hard Drive
Some video displaying card
Few fans
Some dust in the back
Windows XP MCE SP2
The Gold won some, the VX2 won the rest. Since both sets of modules use similar chips, it's no surprise that the results are so similar.
Stability
If these modules happen to retail for my assumptions ($350), I'd have to say these are pretty nice. They deliver great results at stock speeds, but don't really allow for -that- much tweaking or improvement upon 550MHz.
Though the 5-6-6 timings are a little retarded, we see in the second stability picture that stock frequencies with 5-5-4-5-12 timings proved 100% stable over the course of 8 - 10 hours, with 2.3vdimm.
For the curious... yes I did try to go beyond 560MHz, but at 564MHz is started becoming unstable. Nothing seemed to help it, whether it be even more loose timings, or more voltage.
I'll make one thing clear before I get into results... the VX2 is obviously more capable of kicking the Gold series butt, however it was handy so I decided to use it.
The Gold PC2-8800 was announced the other day, although pricing is currently unknown. To throw a figure out there, I guesstimate that it will retail for around ~$350US. This may seem somewhat low for a top rate kit, but the Gold series signifies 'value', and not overclocking ability. That being said, I can't see this kit being released for more than $350 despite the rated frequencies being higher than anything else out there. Don't quote me please.
The PC2-8800 are based on D9GMH, so I had figured I knew what I was getting into. Turns out I didn't have a sweet clue. As I found out rather quickly, this is some fussy crap.
Overclocks
DDR2-800 4-4-4-12 2.0v 1T (2.40GHz, 1.25v)
DDR2-900 4-4-3-12 2.1v 2T (2.70GHz , 1.45v)
DDR2-1100 5-6-6-15 2.2v 2T (2.75GHz, 1.45v)
DDR2-1100 5-5-4-5-12 2T (2.75GHz, 1.45v)
DDR2-1120 5-5-4-5-12 2T (2.80GHz, 1.45v)
Oddly enough, these sticks were -very- fussy with CAS4. At DDR2-1000 4-4-4, they gave way to error in MemTest very quickly. Also, at the second listed overclock, even though MemTest proved stable in addition to other synthetic benchmarks, Super Pi would -not- compute past two loops. Even with more voltage, it didn't make a single difference. DDR2-800 CL4 worked fine, but I'm sure it's due to the loose frequency. As the freq increased, CL5 was pretty much forced.
The first three results use tRC 22, while the final two were at 12. No real reason for this, but I usually wait until I get close to the max possible speed before tweaking that, or tRAS.
Results
System:
AM2 Windsor 4600+
ASUS M2N32-SLI Deluxe (701 BIOS)
Hard Drive
Some video displaying card
Few fans
Some dust in the back
Windows XP MCE SP2
The Gold won some, the VX2 won the rest. Since both sets of modules use similar chips, it's no surprise that the results are so similar.
Stability
If these modules happen to retail for my assumptions ($350), I'd have to say these are pretty nice. They deliver great results at stock speeds, but don't really allow for -that- much tweaking or improvement upon 550MHz.
Though the 5-6-6 timings are a little retarded, we see in the second stability picture that stock frequencies with 5-5-4-5-12 timings proved 100% stable over the course of 8 - 10 hours, with 2.3vdimm.
For the curious... yes I did try to go beyond 560MHz, but at 564MHz is started becoming unstable. Nothing seemed to help it, whether it be even more loose timings, or more voltage.