News G.Skill goes modular with Trident X Series RAM

G.Skill is planning to launch a modular RAM family dubbed the Trident X Series, apparently inspired by Corsair.

http://www.bit-tech.net/news/hardware/2012/04/19/g-skill-modular-trident-x/1

 

Looks interesting but i don't really like G.skill ram. It looks too cheap; I know lots of people like G.skill and praise it's usage, but i can't help feel that corsairs heatsinks are so much more professional.

 

I guess being able to convert your memory to lowish profile could be useful. Still... Meh.

Also:

 

If they offer a slide-on waterblock option it would be interesting.

 

Trying hard to care about this. Nope. Memory doesn't need heatsinks, cool looking heatspreader is fine.

 

Maybe, but :
1) to remove dominator heatsinks, you need hex screwdriver, which is not that common. In case of Vengeance, you don't have that option.
2) at least in my area, lately Corsair RAM modules have relatively high RMA rate.

By the way, G.Skill prices in my country for these new ones :
F3-2400C10D-8GTX (2x4GB TridentX DDR3-2400) - 115,4 euros
F3-2400C10D-16GTX (2x8GB TridentX DDR3-2400) - 226.08 euros
F3-2400C10Q-16GTX (4x4GB TridentX DDR3-2400) - 224.53 euros
F3-2400C10Q-32GTX (4x8GB TridentX DDR3-2400) - 435.13 euros

Now let's see Corsair... Oh wait, Corsair has no DDR3-2400 .

 

Agreed, I feel like sticks of RAM like this are one of the common scams in the computer world, like how Macs are more expensive therefore better. Today's RAM does get pretty hot, but not to the point that qualifies for a large heatsink.

I personally don't see how you can make a stick of memory perform better on 1 specific type of CPU. I could sort of understand chipset, but not CPU. To me, that's like saying a company designs an expensive trailer and says that this trailer works especially well with this 1 specific truck. But how? What makes it so much better? I could see how it might work better with trucks rather than SUVs, but even then, how much of a difference does that make? I'm not just ranting here, I'd really like to know with proven facts - what makes this type of RAM perform better on SB or IB rather than anything else?

Aside from the "performance RAM" that targets a specific CPU, I also don't understand how people honestly think they're so much better. Yes, they can overclock a lot, but at the cost of latency. There's a point where frequency doesn't matter, but latency always has room for improvement (until you reach 0ns). I realize today that latency is becoming less important, but this kind of memory has been around since latency was more important than frequency.

 

@schmidtbag: G.Skill actually provides a list of supported boards for the specific modules. It is a sort of reverse Memory QVL. Usually motherboard manufacturers make a list of compatible, tested memory modules. And G.Skill makes a list of compatible, tested motherboards for every specific memory module. For example check out the homepage for one of the TridentX modules, at bottom you got Qualified Motherboards List, and because it lists Z77 boards, it is "targetted" for Ivy Bridge in Press release:
http://www.gskill.com/products.php?index=511&c1=&c2=

 

Thanks for the input. But, wouldn't this performance difference be nearly insignificant? I feel like at best you'd get 1FPS higher in a game or some sort of rendering might be a few seconds faster.

 

Remember the memory controller is inside the CPU these days. To *some* extent, you can hang anything at the end of the data trace, but Intel and AMD, believe it or not, tweak their IMC for particular behaviour which may or may not match pretty well with IC type X,Y or Z.... OR.... be a complete mismatch. Sandy Bridge does NOT work well with Elpida Hypers for instance, yet they're awesome on Nehalem/Gulftown.

At the end of the day, It all comes down to not the BIOS, but the CPU. It IS possible to tweak the BIOS so the sticks work, but often there are hidden latencies that have to be increased to make it work, so while you might still have 1066MHz on paper, the performance is in the toilet.



So, to couple that to the "designed for CPU X" PR comment... the SPD can be tweaked here and there, but really, the RAM just uses an IC that's been well tested in their lab and works well on IMCs A,B and C

 

It is not about performance, but about compatibility, if we talk about these "Ivy Bridge" TridentX, "Sandy Bridge-E" RipjawZ or "Sandy Bridge" RipjawX memory modules - that is why you got supported motherboard list for every module.

Speed is another issue - yes, the effect is very small, that is why i still use DDR3-1333 CL7 .

 

Can't seem to care either. I'm a G.Skill fan but this seems to be quite the gimmick.

Can we expect a Bit tech review of this? I'm sure most of us here would like to see if there's actually a difference, or if that difference actually matters.

 

whenever I see modules like this feel deep satisfaction knowing that my 1.35v low profile, low voltage modules from Samsung could clock circles around them. on benchmark run's I've taken them to ~2200mhz on 1.5v, I can't go higher simply because i'm capped at a 2133 divider, if i weren't they could probably pull 2.5Ghz without breaking 1.6v. all without heatspreaders, a cool name, fancy packaging, or even a full sized PCB. not to mention how much cheaper they are.

 

@yougotkicked: you mean http://www.amazon.com/Samsung-Electronics-Extreme-MV-3V4G3D-US/dp/B00592002W -> http://www.techpowerup.com/reviews/Samsung/MV-3V4G3/6.html

 

I can't see the additional heatsink making any difference with all the air that must be in the gap. Would love to see some benchmarks though (hint hint). You could even compare the Corsairs!

 

Any memory can be converted to low profile. I have/had some Vengeance ram that I needed to fit under a Noctua NH-D14, so I pried the heatsink off and replaced with some copper ramsinks. Then clocked them as high as the motherboard would allow (which is only 1866 for that board and memory). It still runs perfectly, in my Llano build that I use mainly for browsing and working.

 

Wave goodbye to your RAM module warranty, because it left the building.

 

+1 for Bit-Tech doing another large memory test to see what real world difference there is between cheapo ram and OMG++SuperHYperDDRXX modules.
Also, is there any point in ram-sinks? Easy to test*, just pick up some RAM with a heat spreader, see how high it will clock (and take temeratures), then pry the heatsink off and test again.


*if you can afford to bork some RAM.

 

@debs3759 I tried that with some old, unused ram and every ram chip got torn out of the silicone

Ram only takes like 5 watts running - there is little chance that any heat sink is going to be worth the extra cost (if any).