• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

FEATURED G.Skill Trident Z 2x8GB DDR4-3200 CL15 - F4-3200C15D-16GTZ

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Woomack

Benching Team Leader
Joined
Jan 2, 2005
This thread will be about well known and really popular G.Skill Trident Z 2x8GB DDR4-3200 CL15-15-15 memory ( F4-3200C15D-16GTZ ).
Since it's about the same as DDR4-3000 14-14-14, 3200 13-13-13/14-14-14, 3600 15-15-15/16-16-16 and some other kits then I won't make all tests. Many OCF members already tested this memory so I will add some of my overclocking results and I count that other users will add some of their results too.

If you can't find settings interesting for you then check DDR4-3000 CL14 or DDR4-3600 CL16 threads where are more settings that should work on DDR4-3200 CL15.

The same as most other new G.Skill memory kits based on single sided ( single rank ) 8GB modules also Trident Z 3200 has Samsung "B-die" under the heatsinks. It simply means it overclocks really high and it's compatible with most available motherboards.

Here is link to G.Skill product page.

4332.png

No more photos as all already know how Trident Z memory modules look like :)


Here is list on what you can count during overclocking:

3000 12-12-12/13-13-13 1.35-1.45V
3200 12-12-12/13-13-13 1.35-1.45V
3466 13-13-13/14-14-14/15-15-15 1.35-1.45V
3600 14-14-14/15-15-15/16-16-16 1.35-1.45V
3733 15-15-15/16-16-16/17-17-17 1.35-1.45V
4000 15-15-15/16-16-16/17-17-17 1.35-1.45V
4133 16-16-16/17-17-17/18-18-18 1.35-1.45V

Above settings depend from luck and used motherboard. For example one kit can make 4000 CL16 1.35V while other will need 1.45V for the same.
In next post I will show example results as I was showing many others around the forums for last months ( like in 3000/3600 thread ).

Trident Z 3200 memory has only one XMP profile which is working without issues on ASRock and MSI Z170 motherboards. It's also working on ASRock X99M Killer 3.1.
XMP: 3200 15-15-15-35 2N 1.35V

3200_trz32.jpg
 
Last edited:
OC Results

3200 13-13-13 1.35V

3200_trz32oc.jpg


3600 16-16-16 1.35V

3600_trz32.jpg


4000 16-16-16 1.40V

4000_trz32.jpg


3000 C14 and 3600 C16 kits were a bit better ( read both needed slightly lower voltage for the same settings ).
 
Other results

4333 17-21-21 1.5V

msi2.jpg

Maximum frequency was a bit over 4400 but I couldn't save it. Maximum CPU-Z validation 4358 what is not much higher than the result above. I assume that on better motherboard it can go a bit higher.
 
Hi Woomack. This is my first post. :)

First, let me thank you for all the work you do as I've browsed several of your posts regarding your overclocked memory profiles. :thup:

I've recently built a PC with this G.Skill kit. I'm using an 8600K on a GB AG7 Z370 mb with the F5h BIOS revision (12/12/2017). When setting the RAM to XMP (only one profile is provided), the AG7 sets the VCCSA to 1.3 and VCCIO to 1.2. The VCCSA seems rather high. I understand some mobos do that (MSI),

With the 8600K at default, I've oc'ed the G.Skill to 4000 MHz. I just tried 18-18-18-36/700/2T, but it failed after 38mins of OCCT AVX Linpack. VCCIO is 1.28 and VCCSA 1.30, DDRv is 1.55. I had DDRv at 1.5 before but OCCT failed just over 30mins on that run.

Can you recommend some timings, both primary & tertiary for 4000? Would tweaking tWCL affect stability or performance much?

Thanks for your help!
 
:welcome:

I will start from the end ... tWCL doesn't matter much but you can check something below CL. In general it can't be higher than CL, some motherboards correct it automatically, some not.

In general secondary+ timings depend on memory chips and motherboard so hard to recommend something ( and I don't have all settings in my mind to just recommend it without testing ). Start from stabilizing main timings and later play with additional. Try lowering them step by step + short test like HyperPi32M or something else that uses memory. Drop timings a bit and check if performance is changing. Many timings are not worth to touch.

No one said that every setup will run at 4000+. In general it's not as easy as you read around the web and most results that other users post are at unstable settings. If you want stability then check a bit more relaxed timings like 18-19-19 or 19-19-19 1.35-1.45V. I don't know if you have any profiles in your BIOS but on ASUS ROG motherboards are usually profiles that make all easier ( even if won't work then you can check what timings are on the list and compare them ).

I'm not sure what about high clock on Z370 but I don't think it will need more than 1.25V SA and IO. tRFC can be lower, something like 520 should work stable and it affects performance a bit so you may play with that.
Most tertiary timings are to keep stability and are not affecting performance much. Usually what motherboard sets is good enough and don't need changes.

2 memory slot motherboards are easier to stabilize at higher clock. I see that I was testing this memory kit on MSI Z170I which is ITX mobo. I have Z370 ITX for couple of days and new DDR4-4266 kit which runs at XMP but I only set it to XMP and it runs 24/7 since then ( needed new BIOS to do that ). I had no time to check additional settings but I will probably post a new thread about this memory so there is a chance on more details. 4266 kits are using similar IC to what you have in your memory kit.
 
I will attest to Woomack's abilities with these ICs. I can hit 4000 MHz with mine, and get 3866 MHz stable, just takes more voltage than I like on this motherboard. Did it all with a DDR4 3000 TridentZ set and reading Woomack's reviews. Good stuff there.
 
Strangely enough, 19-18-18-36/700 passed OCCT Linpack with AVX but as soon as I ran RealBench my system rebooted.

5GHz, m4000.jpg

Subsequently I bumped the timings down to 19-19-19-41. I tried setting tRAS to 38 but it drastically reduces memory performance in Aida. I've noticed this with several different tRAS settings in the BIOS. My guess is that it's a BIOS issue. I've had this sort of things happen to me with my old Max V Formula as well.

Woomack: Is 520 tRFC safe for 24/7 use? I understand this setting doesn't necessarily reveal errors outright and the wrong setting could cause system corruption "silently." I'm by no means doubting you, just asking for clarity. :)

Alaric: What is your tRFC set to and your DDR voltage?
 
Last edited:
lower tRFC is safe but too low may cause some stability issues, usually something near 500 will be stable on Samsung modules @4000 and dual channel motherboards
 
lower tRFC is safe but too low may cause some stability issues, usually something near 500 will be stable on Samsung modules @4000 and dual channel motherboards

Now look what you did, Woomack. I have to bump my tRFC back up to 500 and try 3866 MHz again, because I just ran it on 275 tRFC before. LOL
 
275-300 is good for benching at high voltages or lower frequency but on some platforms it's not stable
 
Hmm. But now I'm wondering if I can drop my voltage if my tRFC comes down. I seem to recall hitting some good speed (3866 MHz) but not getting stable and it may have been after I dropped my tRFC to 275. It wouldn't take much to check. Just bump the multi up one and set the tRFC to 530 or something. I should still get enough performance bump from the speed increase to make it worthwhile on this platform, right? Even if I have to keep my current DRAMv I'd be happy.
 
From what I know the tRFC is memory refresh cycle from the memory controller and with faster memory it is set higher equaling the same memory refresh time as slower memory.
 
Last edited:
Did some more tests. Ran 19-19-19-41/520 at 1.45v and passed OCCT AVX. Now on to RealBench for 8hrs. If it passes then I'm done tweaking, at least for now. :D

5GHz mem4k, 19-19-19-41-520.jpg
 
I'm wondering if your rig rebooted with RealBench do to memory overclock or processor overclock?


My rig wasn't overclocked at the time. I was playing some videos via MPC-BE just before trying to run RB. There may be a driver conflict somewhere. I'm using the latest 18.3.4 Adrenalin drivers and since they're beta drivers, it may just be due to that. I've seen this behavior before but it was a different platform (Z77), and indeed it was the AMD driver at the time.

I think I've finally found a stable overclock, and will be sticking with it. It passed both OCCT AVX and an 8hr RealBench test. Here are the results:


5GHz mem4k 9-19-19-41-520, OCCT AVX finished.jpg 5GHz mem4k 9-19-19-41-520, OCCT AVX running.jpg 5GHz mem4k 9-19-19-41-520, RealBench 8hrs.jpg CPU-Z Bench.jpg


CPU-Z validation:

https://valid.x86.fr/skdyug
 
G. Skill has kits that are specced to run at 1.5v with the same lifetime warranty as the 1.35v stuff. I would say the Sammy B dies are good at 1.45v.
 
Back