• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** AMD "Zen" thread (inc AM4/APU discussion) ***

Caracus2k said:
2666Mhz oc and dual channel might be ok for now but its a little too close to the point where performance starts to seriously suffer across a wide range of apps and games. Anyone running a 6700k l, especially if overclocked, with 2133Mhz ram is most definitely holding back their system. (or course its well know that running mem much above 3000Mhz shows very little gain outside of benchmarks)
Click to expand...

This is wrong, I'm afraid. By all means try to back that up with some actual numbers. I mean, I'd like Quad Channel and faster RAM. You can certainly show some benefits sometimes with these things, but you can't support the impression you're creating above that it's getting dangerously close to some sort of tipping point of failing performance.
 
Harlequin said:
http://cpugrade.com/amd-ryzen-thoughts/

that article says ryzen is capable of using DDR4 3200
Click to expand...

I'd probably read the article before quoting it if I was you ;

Notes Concerning the Tested AM4 Platform and Ryzen Sample

As per the publication, there are some important notes that need to be taken into consideration with these results:

The sample used to accertain these performance metrics, was an engineering sample with stepping A0, on a platform with unfinalized features. The sample part number is 2D3151A2M88E4.

The A0 stepping contains two IPC-inhibiting bugs; one relating to the micro-ops cache, and one relating to the SMT implementation.

Despite these present bugs, the features were enabled. Evidently, this is going to hinder performance by an unknown percentage.

The AM4 platform currently has issues detecting SSDs and also failed to detect an NVIDIA GeForce GTX 1080.

The Ryzen sample's memory controller suffers from instabilities with system memory at data rates above DDR4-2400, despite the fact that this specific sample supported DDR4-2666.
Click to expand...
 
Martini1991 said:
I'd probably read the article before quoting it if I was you ;
Click to expand...

do you just skim read then post before reading properly? let me also quote from that article:

AMD has already publicly revealed, as of the New Horizon live event, that stock clock frequencies for production silicon Summit Ridge SKUs, will start at 3.40 GHz. That represents a 7.9% increase over the sample used here to gather the results. Zen's memory controller will also be capable of detecting DDR4-3200 modules, which will give the final product 33% more memory bandwidth than the sample here could take advantage of.
Click to expand...
 
Silent_Scone said:
Ultimately started as an observation is now degraded into treading on egg shells, trying to explain how memory frequency and IMC strength are intrinsically related.

Joy. Lessons over :p
Click to expand...

ignoring the small fact that I have linked 2 Z170 based motherboards , 1 can clock the ram higher than the above - based upon quality of the parts used; ergo IMC is only 1 part of the overall system.

apparently you have limited understanding in that.
 
Memory OC ability is really not that big a deal

Well, here's another set of gaming benchmarks using 2133 / 2666 / 3200 speeds on the X99 platform and also for the Z170 platform.

It seems to me that in practice, each CPU has a 'minimum sweetspot' above which there are no substantial gains. Basically you need to be able to fill the CPU with instructions/data for it to operate at maximum speed and that's it. How much you can OC beyond that is of little practical relevance. In other words, while you can bottleneck a fast chip with very low-speed memory (e.g. 1333MHz) you really can't starve it once you hit speed X. Maybe in Zen's case a golden sample fully OCed needs X=2133 and there's no point going above that...

To quote from this article's conclusions on the last page:

Averaged across our seven games, we see a 2% boost in average FPS and a 3% boost in minimums going from DDR4-2133 to DDR4-3200, which is relatively impressive given the settings we were playing at (maxed out at 2560x1440, typically).
Click to expand...

So really, numbers so low that there's not much to see there...

More importantly, the article concludes (and I think this is spot on):

Overall, we've found that DDR4-2666 memory is definitely the optimal choice in terms of performance and value on both the X99 and Z170 platforms, and while faster memory will benefit the Z170 platform, it comes at a cost that likely isn't worth it for gamers when the money could go to other components like CPUs and video cards.
Click to expand...

There's so many other things where you'll hit a bottleneck before you need to go after memory OC... This ability should really be near the very bottom in your list of priorities.
 
Harlequin said:
And your `facts` in regards to Ryzen are only facts if they are negative.

ryzen is doomed , says martini , as it doesn't have quad channel - forgetting that kaby lake is dual channel as well
Click to expand...

I never said any of that though.
Now we've evolved to out and out lying?

And again, if Zens IMC is worse, who cares? If it's a fact, it's a fact. It doesn't make it negative, you're the one who's making it a negative.
 
Martini1991 said:
And I linked to the cheapest Z2170 boards that both put 3866MHZ DDR4 support there.

So...
Click to expand...



Most of that is down to ASUS T-Topology trace routing as to the differences. It's a totally different argument.

What use is optimising signal integrity if the memory controller can't handle the speeds you're working towards.

The obvious answer to that question is...none :D lol.
 
Harlequin said:
http://cpugrade.com/amd-ryzen-thoughts/

that article says ryzen is capable of using DDR4 3200
Click to expand...

mm... a cording to that its the same specification as broadwell, other than dual channel.

I think for people to lay down a number and insist it can't go beyond that is not a smart thing to do, to be polite about it.

Even this argument that "AMD hysterically have weak IMC's" is a bit off context, Phenom II CPU's had an IMC bottleneck, which was cured by overclocking the IMC, which everyone did, including me, i don't see how that's relevant here.
Piledriver was rated at 1866Mhz, maybe thats more relevant here, i had two of those, i ran them both with 4 sticks of RAM at 2400Mhz, thats not weak at all.
 
humbug said:
mm... a cording to that its the same specification as broadwell, other than dual channel.

I think for people to lay down a number and insist it can't go beyond that is not a smart thing to do, to be polite about it.

Even this argument that "AMD hysterically have weak IMC's" is a bit off context, Phenom II CPU's had an IMC bottleneck, which was cured by overclocking the IMC, which everyone did, including me, i don't see how that's relevant here.
Piledriver was rated at 1866Mhz, maybe thats more relevant here, i had two of those, i ran them both with 4 sticks of RAM at 2400Mhz, thats not weak at all.
Click to expand...



Phenom II X4's struggled to run 1600MHZ stable because the IMC was so weak, it's not out of context at all.

Thubans IMC was a lot better though.

Nothing to do with the CPU NB clockspeed.
 
Silent_Scone said:
Most of that is down to ASUS T-Topology trace routing as to the differences. It's a totally different argument.

What use is optimising signal integrity if the memory controller can't handle the speeds you're working towards.

The obvious answer to that question is...none :D lol.
Click to expand...

its not a different argument , as you are now beginning to understand - IMC can go so far , but with a high quality motherboard and chipset , can take the ram that bit further
 
You must log in or register to reply here.
Back
Top Bottom