Asrock Outs G.Skill DDR4-4333 Memory Kit With Z170M Formula OC Compatibility Announcement

[kewlguy239]

Asrock is touting the merits of its Z170M OC Formula motherboard, which the company claimed is the only Z170 platform capable of handling an upcoming G.Skill Trident Z memory kit clocked at 4333 MHz.

Asrock Outs G.Skill DDR4-4333 Memory Kit With Z170M Formula OC Compatibility Announcement : Read more

 

[RCFProd]

DAYUM!

 

[nitrium]

Isn't that screen capture doctored? I humbly suggest it is, because it shows 4338.4 as the DRAM frequency, but since this is DDR (double data rate) RAM, it should be showing exactly HALF of that value, i.e. 2169.2 MHz, right? I know my DDR3 1600 RAM shows as 800MHz, which is exactly what you would expect. See this review for DDR4 3200:
http://www.overclockers.com/g-skill-trident-z-16gb-ddr4-3200-c16-memory-review/ specifically this image and you'll see what I mean: http://www.overclockers.com/wp-content/uploads/2015/10/TRZ_3200_OC.jpg

 

[AlistairAB]

Isn't that screen capture doctored? I humbly suggest it is, because it shows 4338.4 as the DRAM frequency, but since this is DDR (double data rate) RAM, it should be showing exactly HALF of that value, i.e. 2169.2 MHz, right? I know my DDR3 1600 RAM shows as 800MHz, which is exactly what you would expect. See this review for DDR4 3200:
http://www.overclockers.com/g-skill-trident-z-16gb-ddr4-3200-c16-memory-review/ specifically this image and you'll see what I mean: http://www.overclockers.com/wp-content/uploads/2015/10/TRZ_3200_OC.jpg

Maybe it is just a bug in CPU-Z? It's not recognizing the twice per clock nature of that ram. It could be since it hasn't even been announced yet.

 

[AlistairAB]

XMP profile is reported correctly in the top right, 2169 MHz

 

[nitrium]

nizing the twice per clock nature of that ram. It could be since it hasn't even been announced yet.

Not imo, because CPU-Z is directly measuring the bus which would be defined by the needs of the RAM put in there (this number actually constantly fluctuates in realtime). This is NOT a value read from the RAM profile (unlike the XMP reading) but from the system itself. In any case, we know DDR4 system RAM is correctly read by CPU-Z from the link I provided above for other DDR4 modules (it is standardised after all).

 

[okanimy]

Isn't that screen capture doctored? I humbly suggest it is, because it shows 4338.4 as the DRAM frequency, but since this is DDR (double data rate) RAM, it should be showing exactly HALF of that value, i.e. 2169.2 MHz, right? I know my DDR3 1600 RAM shows as 800MHz, which is exactly what you would expect. See this review for DDR4 3200:
http://www.overclockers.com/g-skill-trident-z-16gb-ddr4-3200-c16-memory-review/ specifically this image and you'll see what I mean: http://www.overclockers.com/wp-content/uploads/2015/10/TRZ_3200_OC.jpg

Actually, I think there's a specific pairing of CPU-Z version (should be 1.72 by the looks of this screen) and the Skylake platform that causes the memory speed to show up as the full MHz instead of the half MHz that we're all so used to seeing. The 3200 link you pointed out is using CPU-Z 1.74, so it's back to showing half of the value.

 

[AlistairAB]

nizing the twice per clock nature of that ram. It could be since it hasn't even been announced yet.

Not imo, because CPU-Z is directly measuring the bus which would be defined by the needs of the RAM put in there (this number actually constantly fluctuates in realtime). This is NOT a value read from the RAM profile (unlike the XMP reading) but from the system itself. In any case, we know DDR4 system RAM is correctly read by CPU-Z from the link I provided above for other DDR4 modules (it is standardised after all).

Yes I agree it is not a profile value but in real time, however it is still dependent upon the BIOS readings for example, so it is understandable that it might not read correctly.

 

[nitrium]

Actually, I think there's a specific pairing of CPU-Z version (should be 1.72 by the looks of this screen) and the Skylake platform that causes the memory speed to show up as the full MHz instead of the half MHz that we're all so used to seeing. The 3200 link you pointed out is using CPU-Z 1.74, so it's back to showing half of the value.

I don't see anything about this issue in the CPU-Z version changelogs. Since we can trivially DL previous versions, can anyone here verify that CPU-Z gives incorrect readings on version 1.72 for DDR4? I guess we won't find out for certain until this RAM officially launches (I do think it's a real product), but I maintain this particular screen capture was probably doctored.

 

[DarrenLays]

Actually, I think there's a specific pairing of CPU-Z version (should be 1.72 by the looks of this screen) and the Skylake platform that causes the memory speed to show up as the full MHz instead of the half MHz that we're all so used to seeing. The 3200 link you pointed out is using CPU-Z 1.74, so it's back to showing half of the value.

I don't see anything about this issue in the CPU-Z version changelogs. Since we can trivially DL previous versions, can anyone here verify that CPU-Z gives incorrect readings on version 1.72 for DDR4? I guess we won't find out for certain until this RAM officially launches (I do think it's a real product), but I maintain this particular screen capture was probably doctored.

On some of the CPU-Z versions, with my old ASRock OC Formula Z170, it'd show the full MHZ of the kit, rather than half. It's not "doctored" or edited, it's just how CPU-Z reads it for some reason.

I have a Z170M OC Formula that just showed up yesterday that I just installed in my system before heading off to work, and it shows up.
On CPU-Z V 1.75, it shows my 3600 c16 kit as 1800, but I believe on version 1.72 it shows 3600. I'll check later if I get a chance to.

 

[jasseh]

Actually, I think there's a specific pairing of CPU-Z version (should be 1.72 by the looks of this screen) and the Skylake platform that causes the memory speed to show up as the full MHz instead of the half MHz that we're all so used to seeing. The 3200 link you pointed out is using CPU-Z 1.74, so it's back to showing half of the value.

I don't see anything about this issue in the CPU-Z version changelogs. Since we can trivially DL previous versions, can anyone here verify that CPU-Z gives incorrect readings on version 1.72 for DDR4? I guess we won't find out for certain until this RAM officially launches (I do think it's a real product), but I maintain this particular screen capture was probably doctored.

I still have version 1.72.1 (ROG) installed and can confirm that my DDR4 shows the full value rather than the half value: http://i.imgur.com/Ezpa9eI.png

 

[nitrium]

I still have version 1.72.1 (ROG) installed and can confirm that my DDR4 shows the full value rather than the half value: http://i.imgur.com/Ezpa9eI.png

Well OK then! I take back everything I said.

 

[WarWolverineWarrior]

Is that speed any good with that latency?

 

[turkey3_scratch]

Is that speed any good with that latency?

Yes it is. You see 15CL with 2133Mhz. 19CL with double of 2133Mhz is extremely nice.

 

[DarrenLays]

Is that speed any good with that latency?

Not for 247 use, but with a good kit of B die ddr4, 4000 12-11-11-28 1T is possible for benchmarking, which is a huge point increase on CPU Intensive applications.

 

[RedJaron]

Is that speed any good with that latency?

Many people don't understand that latency on DDR3 and DDR4 is measured in cycle count, not in strict time. Thus 19-26-26-46 is referring to 19, 26, and 46 clock cycles, not nanoseconds. DDR4-4333 equates to a 2166.667 frequency. Inverting that means each clock cycle is 0.462 ns in length. Multiplying that by the latency counts means the actual time latency on these modules is 8.769 ns, 12 ns, and 21.231 ns.

For comparison, that's a little faster than a DDR3-1600 module with 8-10-10-17 timings. However, the higher frequency means you can get much more data transferred from the modules, so the memory bandwidth is much higher.