DrMrLordX
Lifer
- Apr 27, 2000
- 22,886
- 12,943
- 136
-Resolution Gaming[/B]. I have presented numerous slides and review links to present and backup my claims. But, nobody and I repeat nobody have posted higher than 1080p slides expect me.
The problem here is that:
a). the benches you're posting do not feature the fastest CPUs produced by AMD or Intel. There are no 9590s or 5960X/5930k benchmarks there. I see the 8350 and 4930k . . .
b). the framerates you're showing in those benchmarks are almost universally bad, certainly not within the range of what a "high-end" user would tolerate.
All you've done is made the case against 4k gaming, period, by showing it in a poor light. Do you really think that a guy with an overclocked 5930k and TitanX SLI is going to have those problems? How about quad SLI?
I have looked at some 4k TitanX SLI benches, and I see a mixed bag. First check out this one:
http://hexus.net/tech/reviews/graphics/81892-nvidia-geforce-gtx-titan-x-sli/?page=4
65 min FPS with TitanX SLI at 4k with all the bells and whistles. And that's on a 4770k. Then there's this one, admittedly with a test system with specs that are not listed (boo):
http://www.pcgamer.com/benchmarks-gtx-titan-x-in-sli/
Here we see TitanX SLI getting 16 average fps in the same game at 4k (settings unknown, though they probably couldn't have used higher settings than what hexus used in their reviews).
What can we conclude from this?
1). If you want to make 4k look bad on everything, there are sites out there that have the data you want, like PC Gamer. If you are (somehow) putting up genuinely-awful framerates such as 16 minimum/56 average with TitanX SLI in a game that has been around for a little while, chances are you can put up those less-than-stellar numbers with any number of different CPUs. That doesn't make the CPUs equal in any sense - it just means the benchmark settings are off, or there was a problem with the driver at the time, or something.
If you do whatever it was that Hexus did to get a minimum fps of 65 fps with a 4770k, then you have to open up to the possibility that changing CPU in that situation might actually affect your framerates. The GPUs were obviously able to handle the work of running Bioshock Infinite @4k for Hexus with all the bells and whistles, including AA which many gamers turn off at that res anyway (doesn't help much, or so they say). Do you really think at 4.4 GHz is gonna break 60 fps minimum in that situation, when many benchmarks show 9590s not breaking 60 fps minimum in games at 1440P? I think not.
He doesn't know what he's talking about.
This is what is commonly known as projection. See here for further info: https://en.wikipedia.org/wiki/Psychological_projection
I'd at least like to give him a chance to explain his intimations rather than just leaving them as vague comments. I can think of nothing that points to fundamental design flaws in the X99 platform that makes systems based on it inherently unstable. Most, if not all users of the rather-expensive tech have reported that it's the bee's knees.
One of the SKU has low voltage margin, 9.8% actualy, wich is at the limit of the 10% minimal requirement.
That s not enough to guarantee stability at high loading, this was "solved" by increasing the MBs voltages, wich has been pointed by many users as an error from manufacturers, while it is not...
![]()
http://www.hardware.fr/articles/924-6/overclocking.html
So I read the article via Google translate, and found nothing in there where the article's author claims:
a). that the platform is inherently unstable
b). that "wrong" voltages are being set by the UEFI or in CPU microcode
All he did was overclock some Haswell-Es and hit clockspeed walls at 1.3v that weren't all that great. In other news, people on this forum have gotten better results out of their Haswell-Es. The author of the article DID have this to say:
Although it is always difficult to draw conclusions on the basis of a limited number of processor, we still found low limits on our 3 processors with 4.1 GHz to 4.2. If the 4.5 GHz were available with Sandy Bridge-E with reasonable power, he had to settle for 200 to 300 MHz less with Ivy Bridge and Haswell-E-E does not seem to do better or a little worse. For consolation we can say that for the relatively small base frequency of i7-5960X, the gain is still significant
In other words, the benchmark author did not win the silicon lottery. Oh well.
Last edited: