G.Skill 16 GB TridentX DDR3-1600 CL7 Memory Kit Review

Add Your Comments

Well, we brought you two of the higher speed kits in the past, one from G.Skill and the other from ADATA. Neither of them did well in testing due to their astronomical prices. Sure, they reached mind bending frequencies, but that alone is not enough to make a reasonable person consider them.

With the G.Skill TridentX kit we’re looking at today, now we’ve gone to the other extreme. Not everybody possesses a Haswell or Ivy Bridge CPU that can reach astronomical frequencies. More importantly, most people don’t have any interest in or need to push ultra high frequencies.

This 16 GB kit, G.Skill’s F3-1600C7D-16GTX, runs at a much lower frequency – DDR3-1600 – but with extra low latency. Its four main timings (tCL-tRCD-tRP-tRAS) are 7-8-8-24, very tight by today’s standards.

Specifications & Product Tour

G.Skill didn’t pack a fan with this kit. It’s well packaged in simple molded plastic. It’s good for hanging in retail stores, in case anyone reading this actually buys computer parts at brick & mortar stores any more.

TridentX DDR3-1600 Box

TridentX DDR3-1600 Box

TridentX DDR3-1600 Box Rear

TridentX DDR3-1600 Box Rear

Model # F3-1600C7D-16GX

Model # F3-1600C7D-16GX

As usual, not much is included – just two 8 GB sticks of memory and a case badge.

RAM & Case Badge

RAM & Case Badge

The RAM itself is good looking as TridentX typically is. I’ve been a fan of this styling since it first came out.

TridentX DDR3-1600 CL7

TridentX DDR3-1600 CL7

TridentX DDR3-1600 CL7

TridentX DDR3-1600 CL7

TridentX DDR3-1600 CL7

TridentX DDR3-1600 CL7

These are double sided sticks with 4 GB banks of memory on each side of the PCB.

Double-sided modules!

Double-sided modules!

Model # on RAM

Model # on RAM

TridentX memory comes with removable top fins. This serves a dual purpose. For those that push their memory to the absolute max using extreme cooling, the flat top is perfect for mounting a RAM pot. It’s also perfect for those with larger heatsink/fan combinations that overhang the RAM. Removing that top fin instantly makes them low profile sticks.

TridentX Removable Top Fin

TridentX Removable Top Fin

Ok, a couple more photos and we’ll stop.

G.Skill

G.Skill

TridentX

TridentX

TridentX...X!

TridentX…X!

So far so good, let’s get these on the test bench.

Test System

We’ll be testing this kit, like the other recent memory kits, on an Intel Haswell-based system featuring the ASUS Maximus VI Extreme motherboard and i7 4770K CPU. Competition comes from G.Skill (two different kits) and ADATA.

CPU i7 4770K
MB ASUS Maximus VI Extreme
RAM G.Skill TridentX DDR3-2600 10-12-12-31
G.Skill TridentX DDR3-2933 12-14-14-35
ADATA XPG V2 DDR3-2800 12-14-14-35
G.Skill TridentX DDR3-1600 7-8-8-24
OS Windows 7 Professional x64

As mentioned, I like the TridentX look, especially on a red & black ROG motherboard.

TridentX DDR3-1600 CL7 Installed

TridentX DDR3-1600 CL7 Installed

Stability at Rated Speed

The first thing we always test is stability at rated speed. To get the proper speed, timings & voltage was as simple as setting the XMP profile and saving it to BIOS.  Our chosen stability tester was kindly provided by UltraX – the UltraX R.S.T. Pro USB RAM tester. It performs the same intense testing as its PCIe predecessor, but in a more convenient package. As you can see, five passes passed after almost eight hours of hammering on the RAM. With their lower voltage (1.5 V), the sticks weren’t quite as hot as their 1.65 V rated brethren.

R.S.T Pro Passed

R.S.T Pro Passed

It’s nowhere near as intense as the R.S.T. Pro, but for in-OS testing, we went with HyperPi, using all eight threads available on the CPU. For 99.9% of people, this will be the only stability test you need. If it passes HyperPi, it’s stable.

HyperPi Passed

HyperPi Passed

Not only are these sticks stable at their rated DDR3-1600 with very tight CL 7, they do it at a meager rated 1.5 V. Nothing at all to complain about here.

Performance

For performance testing, we run everything with the CPU at ASUS multi-core enhanced stock (3.9 GHz on all cores) and the RAM at its rated speed, timings & voltage. Everything was run three times, with the results averaged.

One thing to keep in mind as you look through this testing is that latency can only make up for so much speed. The closest kit we have compared here is a full GHz (500 MHz actual, times two for double data rate) faster than this kit. So while latency will help matters, no DDR3 latency on the market today (or probably in the past) can make up for a GHz difference.

Synthetic Testing

Memory bandwidth was actually not that bad. Writes suffered the most relative to the other kits, but reads were reasonably close. Latency suffered due to the massive frequency hit, but was still less than ten nanoseconds from the fastest kit to this one due to the tighter timings.

AIDA64 Memory Bandwidth

AIDA64 Memory Bandwidth

AIDA64 Memory Latency

AIDA64 Memory Latency

Compression, Video Conversion and Rendering

You can read the graphs as well as I can type out descriptions for them, but suffice to say this kit is very nearly as good at every day tasks as all of the other kits tested here. For day to day operation, this kit is great – and with lower voltage, it’s better for your memory controller and system temperatures as a whole.

7zip

7zip

x264 Benchmark

x264 Benchmark

Cinebench R10

Cinebench R10

Cinebench R11.5

Cinebench R11.5

PoV Ray

PoV Ray

Benchmarking

Timed benchnmarks can be difficult for this kind of kit. SuperPi always wants speed and timings, not one or the other. That’s why the 2600 kit is the best at SuperPi 32M – its combination of relatively tight timings and fast speed are the best of the three kits here. The DDR3-1600 kit didn’t do poorly, but it’s not going to be the go-to choice for benchmarking SuperPi.

(The go-to choice for competitive SuperPi would be PSC-based, no-longer-manufactured kits. They’ll always be the best for benchmarking until the move to DDR4 is complete and nobody uses DDR3 any more. It isn’t often you hear ‘they don’t make them like they used to’ in a technology market, but in the case of RAM, that’s entirely true. )

SuperPi 1M

SuperPi 1M

SuperPi 32M

SuperPi 32M

WPrime is much closer than SuperPi, mostly because WPrime is pretty ambivalent about what RAM you use. These sticks aren’t the best for it, but they won’t hurt you much either.

WPrime 32M

WPrime 32M

WPrime 1024M

WPrime 1024M

Overclocking

I went with a different tactic when overclocking this kit. The previous two kits were all about frequency, and that’s fine for what they are, but mostly useless for…well, for pretty much everybody with regard to actually using your RAM. This took a little longer of course, but it’s well worth it to show the versatility of this kit.

What I did was first set the voltage to your standard 1.65 V. There is no problem doing that, the kit will take it just fine. Then I started raising frequency in your typical steps (1600, 1866, 2000, etc) and adjusting timings so that they were stable through SuperPi 32M. This isn’t HyperPi stability, but it’s close and a good starting point.

First up came DDR3-1800, 7-9-9-24. At the increased voltage of 1.65V, it passed there without issue. Note CL is the same, but tRCD & tRP are both raised from 8 to 9.

Next I kept those timings and raised it to DDR3-1866. Fail. Voltage had to be bumped to 1.7 V for that to pass. CL had to be loosened to 8 (for full timings of 8-9-9-24) to go back down to 1.65 V.

DDR3-2000 and DDR3-2133 were both stable at slightly loosened 9-10-10-28, still at 1.65V.

The last jump was larger, but timings didn’t suffer too badly. DDR3-2400 with timings of 10-12-12-30 passed with flying colors at 1.65 V. The full 1 GHz overclock wasn’t possible, at least at 1.65V and reasonable timings, so I stopped at DDR3-2400. Even so, that’s 800 MHz, a full 50% overclock with very reasonable timings.

DDR3-1800 / 7-9-9-24

DDR3-1800 / 7-9-9-24

DDR3-1866 / 7-9-9-24 (@ 1.7V)

DDR3-1866 / 7-9-9-24 (@ 1.7V)

DDR3-1866 / 8-9-9-24

DDR3-1866 / 8-9-9-24

DDR3-2000 / 9-10-10-28

DDR3-2000 / 9-10-10-28

DDR3-2133 / 9-10-10-28

DDR3-2133 / 9-10-10-28

DDR3-2400 / 10-12-12-30

DDR3-2400 / 10-12-12-30

Because I like graphing things for you SO MUCH, I went ahead and made a quick graph showing timings vs. frequency at 1.65V.

Frequency vs. Latency

Frequency vs. Latency

Final Thoughts & Conclusion

Wow, look at this. A memory kit that isn’t all about frequency! You could knock me over with a feather. I’d like to thank G.Skill for sending this kit for testing. Since Haswell’s launch (and to a smaller extent, Ivy Bridge before it), every manufacturer seems to be pushing frequency, frequency, frequency and it has led to a loss in performance at the expense of ‘ZOMG look at the frequency!’ If you’ve read my reviews, you’d have to be extremely thick to miss that being a very frustrating experience.

Well, I’m happy to report there are kits out there that aren’t single-sided. Kits that don’t say DDR3-3000 or bust. Kits that can be used by anyone, with any relatively modern platform – AMD or Intel. Do you need and/or want lower speed, tighter timings and lower voltage? Yes, it does still exist, and G.Skill has an excellent offering in their 16 GB TridentX DDR3-1600, 7-8-8-24, 1.5 V kit.

It’s not limited to just that speed, timings and voltage either. After having so much fun overclocking this kit, I’m happy to report it is also extremely versatile. It managed a 50% overclock. Fifty percent! It also did that at only 1.65 V, what you would expect any reasonable kit on the market to run at that speed, with timings that you’d expect a reasonable 16 GB kit to include.

The G.Skill F3-1600C7D-16GTX kit can be purchased at Newegg for $169.99. Remember, I never said it was cheap; this is still an enthusiast kit after all. Even though it’s not the cheapest, it’s still a whole lot more reasonable then those $350+ 8 GB frequency pushing kits. Being a 16 GB kit, with such tight timings and low voltage, that’s priced right where it should be.

There is a lot to love about this G.Skill TridentX kit, with nary a dent in its armor anywhere. This versatile, extremely overclockable and rock solid stable kit is well worth a look.

Overclockers_clear_approved

– Jeremy Vaughan (hokiealumnus)

Leave a Reply

Your email address will not be published. Required fields are marked *

Discussion
  1. Very nice review as usual Hookie! :thup:

    I was secretely hoping for someone reviewing this kit and I have to admit than I am very pleased by the results!

    Did you try the opposite way for 2600MHz kit ie. lowering frequency and thightening the timings to compare both kits?

    Thx anyway :)