Testing GSkill RipjawsZ F3-19200CL9Q-16GBZMD 4x4GB [DDR3-2600+ action]

[Leeghoofd]

Good luck, don't push the voltages, PJ already killed two of them precious sticks :p

[Alex-Ro]

who is pj? I always fight with him on god forgotten old memory on ebay lol =))

[Shiranui Gen-An]

Good luck, don't push the voltages, PJ already killed two of them precious sticks :p

That...doesn't sound good. I'd assumed these would be essentially the same as the Trident X 2400C9 but possibly with subs better tuned to SB-E/X79. Are they just not as robust? What was the voltage threshold not to exceed?

[Shiranui Gen-An]

As I mentioned, my kit came in yesterday afternoon. I was expecting it to be new "old" stock but clearly from the serial numbers, G.Skill is still producing this model. I have no idea why NewEgg, et al. don't carry it.











So far, I've had more luck with these in terms of stability at their rated speed/timings on X79 than any of my other 2400 kits. I can't see anything significant about the XMP timings that would account for this, except perhaps for the fact that it sets VCCSA to 1.25v and leaves VTT at 1.05v whereas I normally set VTT manually 1.10~1.20v.



Whatever the case, 1 hour OCCT Linpack test with AVX enabled passes with the 3820 at stock and overclocked to 4.5GHz with 128BCLK. Usually it fails after 15 minutes.




I have yet to test on my main system (RIVE+3930k). I couldn't get 2400 stable for 24/7 operation with that combo, though I can pass things like 12x Hyper Pi 32M.

[Reefa_Madness]

So far, I've had more luck with these in terms of stability at their rated speed/timings on X79 than any of my other 2400 kits. I can't see anything significant about the XMP timings that would account for this, except perhaps for the fact that it sets VCCSA to 1.25v and leaves VTT at 1.05v whereas I normally set VTT manually 1.10~1.20v.

If these are numbered like the Trident X Series, then they would appear to be HYK0 based on the "2500" S/N (vs the "1500" kits which seem to be HCH9). Perhaps that factors in to why you are having better luck.

[Shiranui Gen-An]

If these are numbered like the Trident X Series, then they would appear to be HYK0 based on the "2500" S/N (vs the "1500" kits which seem to be HCH9). Perhaps that factors in to why you are having better luck.

They definitely are HYK0:



However, my Trident X 2400C9 are also HYK0, and I couldn't get them stable 24/7 although I think the XMP for these sticks clued me in a bit; I may have actually been setting VTT too high. At any rate, I wouldn't think HCH9 vs HYK0 would make much of a difference for 24/7 stability. OCCT passed fine with these on my main system at XMP settings and CPU at stock.

[Shiranui Gen-An]

This is more of a test of my 3820's IMC than the capabilities of the sticks, but I managed DDR3-2600 C10 Super Pi 32M stable.

[websmile]

[Alex-Ro]

That is one strong imc

[Shiranui Gen-An]

I think these 3820 chips can take a beating better than the C2 39xx CPUs. I'm not that worried about pumping up VCCSA and VTT

[TaPaKaH]

Did the mem really need 1.725 or you were overshooting it with the voltage just to stay on the safe side?

[Shiranui Gen-An]

Did the mem really need 1.725 or you were overshooting it with the voltage just to stay on the safe side?

1.70 couldn't pass Super Pi 32M so yeah. And this board undervolts a little anyway.

[Shiranui Gen-An]

Ivy Bridge may be about to go out of style with Haswell's imminent release, but man does this platform make memory clocking easy. As usual, less volts to pass Super Pi 32M than SB-E for 1300MHz 10-12-12-31-2T, only 1.65v!



Aaand...I didn't have to change a blessed thing for 4x Hyper Pi 32M. I'm not sure why they needed more VDIMM on X79, unless it's something like how Elpida Hyper clocked better on X58/P55 than on the newer platforms.



Didn't have to bump VDIMM up any for DDR3-2666 C10 Super Pi 32M. This is getting ridiculous.



Needed to go up to almost 1.72v for Hyper Pi 32M stability, but still not too shabby.



I'm not sure how far I want to push these in terms of VDIMM given the warning from Leeghoofd, but I wouldn't doubt 2800 C10 or 2600 C9 is possible.

[Shiranui Gen-An]

Testing on Haswell, 2x4GB at DDR3-2800 9-12-12-28-2T Super Pi 32M 2.05v

[TaPaKaH]

wow, I couldn't get mine to work at 1300 on multiple boards

[Shiranui Gen-An]

This kit is back in stock@Newegg, but is $150 more than the TridentX kit: http://www.newegg.com/Product/Produc...82E16820231521

[websmile]

Talking about the Trident X, being bored I ordered two new 2400c9 Kits which were suddenly available again in germany - and they are SN1500 instead of the 2500 they used to be, so seems that these have hch9 or hck0 instead of the HYKO that were used before, alike on the kit I had in late 2012

[Shiranui Gen-An]

Or maybe BCH9?

Sent from my Nexus 4 using Tapatalk 4

[websmile]

Maybe, but they are Samsung D for sure according to behaviour, both kits do 2800 10-12-12 32M and 2600 9-11-11, and boot and do 1M at 2800 9-12-12-21 at 1,9v, which is boards Vdimm limit