Originally Posted by TheBlademaster01
This is true. Cinebench is probably the best benchmark out there for gauging parallel floating point performance. Unless you test single core performance it won't say much about gaming performance, but it is a very accurate representation for what you can expect in multithreaded workloads.
That's why personally I would like to see this benchmark ran on this chip. Should give us a great idea of its actual performance.
Originally Posted by Blameless
Design/architecture has a huge impact on clock potential, making it nearly impossible to draw conclusions from process alone.
Cinebench is based on a real-world app, so it's not a synthetic benchmark, just a canned one.
I understand that there are other variables to consider in regards to clockspeed however I feel like saying around 4.2GHz for a true 8 core CPU (about 200MHz less than a 5960X average) is better than people expecting Zen to do "5GHz OC on Air!" We know in November we are going to start seeing leaked overclocking results for Zen and I can basically picture the headlines now.
It does appear to be a real world test, but that doesn't change how I feel about being able to easily compare different products and being able to do the same with other benchmarks like 3DMark that a lot of people tend to have issues with regarding legitimacy of results.
It could very well be that it clocks great for all we know, maybe 4.5GHz under water, except it just doesn't have a lot in its favor. It's not like I don't want it to clock well, I just don't want people to go into it thinking its going to clock amazingly and then get let down later.
Originally Posted by Particle
Given that we already have 8C chips from AMD that have a stock frequency of 4.7 GHz while being fabricated on a now-ancient process, your assertion would appear to be in conflict with itself. That isn't to say Zen will clock highly or not, just that there is no evidence with which to assume one way or the other. We don't know what attention was given to its design with regards to pipelining, clock distribution, and everything else important for frequency scaling. We do know that GF's 14 nm process is probably not exceedingly impressive given the clock speeds of Polaris products. I assume they're both on GF's 14 nm LPP. All we have is speculation though.
It would appear to be a conflict if the 8350 was actually a true 8 Core CPU. There is some evidence, the ES sample is starting off really low to begin with. It is a true 8 core 16 thread CPU unlike the 8350. So again, will have some difficulty with amazing clockspeeds. And of course GF's 14nm LPP that they are using. But surely it's all speculation, but it's not speculation based off of nothing.Edited by BiG StroOnZ - 8/12/16 at 11:22am