Review: AMD Ryzen 7 2700 processor (65W)

Here you have it. Those chips do stock 3.7GHz at 1.1V. Manually setting voltage while having XFR active will lead to crash, because chip will still try to go for higher than stable clock. XFR is actually not limited by voltage, but by Power, Thermal and Amperage limits.
If you use XFR, you have to keep CPU voltage on auto too. Once you start setting something manually, you have to set quite a few related values too.

Your 1.4V would get you with XFR maybe up to 4.15GHz if you had strong enough LLC, otherwise this 1.4V drops to like 1.37V And that's like 4.1GHz. Yet XFR will try to go even to 4.35GHz.

 

Well there's your problem... 1.4v is pretty much the upper limit as to how high the voltage should go; AMD claims any higher than that will start shortening the CPU lifespan.

 

By that logic, you might as well buy a 2600 and save even more money.

 

A question please, in these reviews and charts, when a 7700k mentioned in stock form, does it mean it operates at 4.5ghz all cores synced at full load?
What about 8700k, does it "usually" mean all cores at 4.7ghz at load?

 

Note to Hilbert: If you could include it in future CPU game benches, Fallout 4 is probably the most CPU bound game there is. Even on 1440p it can show sizable differences between CPUs.

 

Actually X have more sense for those that do not want to bother with OC as you can get close to max oc clocks on default settings.

 

Not sure about the 7700K but the 8700K has a max all core clock of 4.3GHz. In practice though it can't keep that up for long and it'll throttle back to around 3.8GHz. If you look at the Cinebench MT scores that would seem about right. My 1600X scores 3450 in MT (CPU-Z) which has an all core clock of 3.7GHz and is beaten by the 8700Ks score of 3702.

 

Had nothing to do with the cpu--2700x draws 105W TDP, remember. So if you are using a mobo with a 95W TDP or lower--you are going to have problems. Or, it could be the increased power draw is taxing your PSU--which most certainly could cause BSODs galore if it's failing or simply inadequate to the task. Many things it could be--but it surely isn't a problem with the 2700x...

 

Note on HH's comments in the reviews about not being able to repeat a problem with memory bandwidth degrading in tandem with overclocking the cpu: I haven't seen it either, ever. The only thing I've seen is a drop in AIDA64's memory read bandwidth scores when the Spectre2 microcode is applied to my 1600, via the MSI x370 Gaming Pro Carbon AC motherboard bios. That is easily demonstrated and repeatable--but has nothing at all to do with overclocking the cpu. I wonder if that's where the initial reports came from, and possibly the people reporting it simply didn't equate the S2 cpu patch application with the mem bandwidth slowdown. Could be, I would guess. Or perhaps it is simply a bug in some of the earlier bios implementations in the x370 boards. At any rate, again, I have yet to see that so far.

 

Can the power figures for the overclocked CPU (overall system) please be included in the Power Consumption tables. It will be great to see how overclocking affects power consumption in different workloads.