Failing Prime 95 1344k instantly @ 1600MHz

[boldenc]

I'm having problem with 4 sticks 4GB G-Skill Ripjaws 1600MHz 9-9-9 1.5v http://www.newegg.com/Product/Produc...82E16820231313
The prime 95 custom test 1344k with 90% ram usage, will fail instantly once I start the test, tried to play with all the voltage but I just increased 2 notch each value, didn't try more but it still fails instantly this specific test with 0.5 Round Error. Tried with XMP profile 1.2 and manually.
If I try 1333MHz the test runs fine without errors. The same ram configuration were running perfectly on ASUS P8Z77-V at 1600MHz.
My guess the motherboard is setting a ram timing the modules doesn't like.
Timing used while testing and failing @ 1600


i7 3770k
ASRock Z77 OC Formula P1.60
G.SKILL F3-12800CL9S-4GBRL x 4

[Tradesman]

Try raising the DRAM voltage + 0.06 and set the CPUVTT to about 1.12

[boldenc]

I tried Dram voltage up to 1.60v, and CPU VTT to 1.150 without any difference, still fails after sec.
I tested each 2 sticks @ 1600MHz in all slots and it didn't fail. But once I put the 4 sticks back, it fails.

[Tradesman]

Is this 4 individual sticks, 2 sets of 2 sticks or a single set of 4 - if mixed sets, may have ones that don't want to play together. Might check each stick by itself with Memtest (4-5 passes) if all pass and this is a single set, might be a bad mobo. Do you have or can your borrow other sticks to see if they too have a problem? If known good sticks fail then that points to mobo or CPU.

[boldenc]

They are 4 individual single sticks but all are the same model, bought in different times.
Memtest doesn't show errors.

Update: I got new set f3-1600c7d-16gtx and testing them.
Btw the new sticks come with the new trident heat spreader is very hard to install into the motherboard, the heat spreader is very sharp and you can't push them securely without hurting your hands .....

[Tradesman]

With the mixed sticks, as mentioned above, even with the same exact model can be problematic. While testing, if you are having problems with the trident heat sinks, the upper fin is removable (small screw on either side). Might have overly tight DRAM slots, I've built about 30+ rigs with Tridents and haven' had any problems

[boldenc]

so far the new sticks are working fine and passed all stability tests @ 1866 8-9-9 1.55v. My motherboard DRAM slots requires hella force. Now I just wonder if I get a new pack of the same ram for 32GB. Will I have the same issue?

[Tradesman]

I've had mixed luck with mixing sets of Tridents, for the most part they have worked together with minimal voltage adjustments, but I did have one set that was fine by itself, but wouldn't play with other sets, guessing the set used ICs from a different production lot than the the other sets I tried it with

[boldenc]

May you give me the settings I should use for 32GB configuration for the Trident X for the timings and voltage to be prepared

btw this ram is supposed to be XMP 1.3 but I noticed the bios is showing them as XMP 1.2, also CPU-Z says SPD XMP 1.2. Is that normal?

[Tradesman]

Believe there should be two profiles 1 for XMP 1.2 the other for 1.3. I haven't run them (the 1600s) with 32GB Have run the 2133 and 2400 sets of 32...but would suggest tWR 15, tRFC 224, tRRD 6, tWTR 7, tRTP 7, tFAW 23 and tCWL 6 as a starting point

[boldenc]

Both profiles are marked as XMP 1.2

[Tradesman]

Is that in the BIOS? No option for a XMP 1.3 profile? Will see if I can find anything and get back this evening, like I mentioned haven't used the 1600 sticks

[boldenc]

[Tradesman]

Odd, it's great though that they OC well, 1868 is looking good

[GSKILL TECH]

Sounds like the four separate modules do not match well enough to run at DDR3-1600

This is why we must recommend purchasing and using complete kits of memory.

Thank you
GSKILL TECH