G.Skill TridentX DDR3-2400 32GB RAM Kit Preview

[hokiealumnus]

G.Skill has sent another RAM kit for review (model # F3-2400C10Q-32GTX) and this time it's not raw speed they're after like their DDR3-2666 kit we reviewed back in May. No, this time it's about speed and capacity. Lots of capacity. 32GB worth.

It's not just big, but pretty darn fast too, at DDR3-2400 with timings of 10-12-12-31.

You can here me talk in the review though, so no more jibber jabber; have a look at some RAM goodness!









Thanks to the great IMC on the 3770K, it's as simple as setting XMP, saving and rebooting.





That's all for now...more to come in the following days/weeks.

[hokiealumnus]

Last night I managed to get the stock speed testing done and ran the in-OS HyperPi stability test, which they passed with flying colors simply using XMP. It's amazing how easy this was. 32GB at this kind of speed didn't used to be an easy thing to do. My, how things have changed.



I'll be running the UltraX R.S.T. Pro 3 probably tonight and then we'll be able to get to some overclocking (and timing experimentation; with Woomack's example, that sounds like it could be fun!).

[hokiealumnus]

Quick & dirty overclock. 1.70Vdimm, 1.14 VCCIO = DDR3-2561.



(Note I'm not even pretending this is stable; I'm just searching for frequency with default timings here.)

Also, for the curious, here are the timings as set by XMP.

[hokiealumnus]

Well, the UltraX R.S.T. Pro 3 is working its magic, torturing the RAM for all it's worth.



The kicker is that 32GB of RAM takes forever to test! 15 hours in, two out of five passes have passed without error and it just started the third.



Why, oh why, did I decide to always run five passes for reviews? Based on this I think it'll take ~36 hours to finish.

On the plus side, no errors after two tests generally means there won't be any.

[ARandomOWL]

tRFC = 300+ lol, didn't know it was possible to set so high.

[hokiealumnus]

Yep; the compromise for running high-capacity at high-speed is almost always high secondary timings. Without that, they couldn't run respectable primary timings.

[ARandomOWL]

Indeed, it makes sense from a marketing point of view. I am surprised that tWCL is that tight in XMP, but then everything else is extremely loose.

[XII]

Yep; the compromise for running high-capacity at high-speed is almost always high secondary timings. Without that, they couldn't run respectable primary timings.

Timings look almost the same as the 4x 8gb RipjawX you tested before. At ~2200 they max-out 9-11-11?

[hokiealumnus]

That kit actually made it to just over 2300 / 9-11-11-31 SP32M stable. I should try that kit on this CPU sometime; my SNB-E chip's IMC is junk and can't even make 2400.

As far as this kit, four loops passed, one to go!

[XII]

For 2400-ish, had to drop timings to 10-12-12, but IMC isn`t to stable then, temp-related it seems. If core temp stays below 50C, it goes. Cold night, sitting outside, hurricane fans.......
As for 24/7, ~2200 9-11-11, and ~1900 8-10-9 seem the limit. tRFC 298 and 268 resp.
Rex4F has trouble using XMS timings though.

[hokiealumnus]

That's not bad at all XII, you have a pretty decent IMC. Mine won't even run at 2400 regardless of temperature (and believe me, I tried as cold as it would go under LN2 without cold-bugging).

34 hours later, I can say with absolute certainty this kit is stable at its rated speed.

[XII]

That's not bad at all XII, you have a pretty decent IMC. Mine won't even run at 2400 regardless of temperature (and believe me, I tried as cold as it would go under LN2 without cold-bugging).

34 hours later, I can say with absolute certainty this kit is stable at its rated speed.

If you`re short in time, I`m happy to test it for you, in 64gb setup RipjawsX I mean

[maxxx]

That is some serious memory testing...

[hokiealumnus]

Rather than piecemeal any more benches, I just went to work and wrote the review this weekend. Of course, I'm surbscribed and more than happy to answer any questions if you have them. Cheers!

What's black, red and goes really fast? Well, fast for an inanimate object anyway. Why,G.Skill's TridentX F3-2400C10Q-32GTX, 32GB, DDR3-2400 memory kit of course!

... Return to article to continue reading.

EDIT - As a general note, I forgot to mention that XMP asks for a VCCIO voltage of 1.25V with this kit. That was not auto-set with the MVG and I set it manually. When it was running "auto" at 1.05V, it was not stable enough to pass the R.S.T. Pro 3 testing. When bumped to the requested 1.25V, everything was right as rain.

[hokiealumnus]

Woomack over at OCF has a set of these and found out that they like some funky timings. At their rated DDR3-2400 and 1.65V, these are perfectly happy operating with timings of 10-10-13-26.



Strange, but true!

[Don_Dan]

That's typical of the newer Samsung 4Gb ICs, hiwa has already shown that with a single-sided 2400 kit a while ago. Unfortunately the pic seems to be down.

[hokiealumnus]

I'd 'thank' you Don_Dan, but the thanks button disappeared.

[Don_Dan]

You're welcome!

[hokiealumnus]

FYI, the 8GB (2x4G) kit rated for the same speed/timings/voltage as this kit is currently on sale for an insane $69.99 when you use promo code: EMCNCGJ44.