Thursday, April 13th 2017
AMD Ryzen 5 1600X Overclocked to 5.90 GHz
New processor launches are closely followed by clock-speed and benchmark records, and that applies to even AMD's Ryzen 5 1600X six-core processor. Professional overclocker Der8auer succeeded in overclocking the chip to 5905.64 MHz without having to disable any cores. The feat was possible due to liquid nitrogen (LN2) cooling. The clock was made possibly by running the chip with a base-clock of 129.79 MHz, and a multiplier of 45.5X. The core-voltage is unclear. The processor was paired with an ASUS Crosshair VI Hero motherboard, and G.Skill Trident Z memory.
Sources:
CanardPC CPU-Z Validation, DigiWorthy
32 Comments on AMD Ryzen 5 1600X Overclocked to 5.90 GHz
Makes sense cause if they blew intel straight out of the water then intel would have just invested billions into developing their next "nehalem" chip that destroyed everything AMD came out with for the next 8 years and they'd be back inn the same boat as they were before. This way we get comparable intel performance whilst not beating them at everything allowing them to gain market share.... or maybe I just read into things too much :)
As with most conspiracy fantasies, the sense of realism eludes me entirely.
You guys need to stop mixing lack of knowledge with drawing conclusions. Next thing in line is the I in Intel refers to the Illuminati or something...
It is extremely obvious that Ryzen is optimized around its TDP budget and that going beyond that is very costly. Don't forget that Ryzen offers an 8c/16t within a 95W TDP budget, where Intel needs 140w budget for that. Effectively the perf/watt of Ryzen UP TO 4 GHZ is better than Intel. That is a huge feat, and obviously one that has drawbacks, one of them being the vCore increase to go beyond 4 Ghz. It's an architectural feature, AMD built Ryzen around a performance target and made the most efficient CPU on that premise.
Meanwhile, Intel has built all Core CPUs with headroom: they have aggressively improved Speedstep across the generations and they drop some extra volts on top 'to be safe' for example during AVX instructions. Intel could build it like this and 'improve from there' because when Core released, every competing CPU was much less efficient. Ryzen doesn't have that luxury, and it also doesn't have the luxury of an extremely optimized process and the best fabs.
You can't compare with daily water cooling OC, science happen at these temperatures.
As in, just how high is the quality of the CPU/SoC.
Lower quality, less overclock
Higher quality, higher overclock
Maybe this is why this particular sample clocked so high?
Watch the actual OC video.
Disclaimer: I haven't had the pleasure of trying ln2 yet, this is all what I've learned from watching others and being interested.
On topic, this is one hell of a feat to do this w/o disabling cores and SMT. Der8auer at it again! Whatever the case, Zen can do shit fully enabled!
Ryzen is limited to OC based on the internal thermal being set 20c higher then the chip is running stopping OC to 4.4-4.8Ghz. This leaves room to relese higher clocked chips later if they didn't do this everyone would buy a 1700 and OC to and beyond 1800x levels costing them profits. This is just a rumor but it makes sense also based on some facts.
I will try to find the engineering statement on the XFR senser stopping you from higher OC.
You can't expect to entirely different cpu designs to clock the same, it just happens it's not far off, which is why people are getting caught up on this I think.. They could clock wildly differently and still perform the same and in that case you wouldn't be thinking along these lines.
Edit:
Someone above said that it wasn't thermals holding these chips back. Sure 60c at load isn't what we'd consider 'too hot', but it is in fact 'too hot to go faster than Xmhz at Xv'. If you drop the temperature 100c you'd be able to gain quite a lot more mhz, drop the temp another 100c and you might (if you're a good overclocker) end up near 5.9ghz. See what I mean? So on this particular architecture 4.1ish happens to be the limit at ambient temperature. No weird limiting trickery, just quirks of this design.
My 1800x can just about manage cinebench at 4.1ghz on a decent custom water loop... With the coolant chilled to 13c it completed a cinebench run at 4.17ghz... So there isn't an artificial limit, just the design at it's upper limit for ambient temperatures.
That's a poll i'd like to see someday..
I bet you a lot of people outside the 'enthusiast' segment would be willing to part with serious amounts of money for this; myself included.
It would be major news, had it been vice versa:
'There is really no difference' is what dominates this video. It's a nice insight, but not something to draw any conclusions from...