• 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.

Your new CPU not as fast as shown in the review you watched? Silicon lottery affects stock Ryzen performance

Grim5 said:
That is correct; the alarming part is not performance variance between the 13 CPUs - what is alarming is the large variance in power draw and voltages while gaming
Click to expand...

I've not watched the video but its stated in this thread that all 13 CPUs perform within spec. I assume that also refers to power draw. If thats the case then whats the problem? If I buy a 7600X and its draws 105W at stock and that within stated specs then thats fine, if I get one and it only draws 90w or less even better. Aslong as its within specs I don't see the problem and its ultimately just boils down to silicon lottery
 
He said they were Stock but with no power limits, I don't know if that just means PBO on or something else he's done. My 7600 stock is limited to about 90w, with PBO 120w and core tuned 105w for example
 
Last edited:
What he's said is true and what he's showing is clickbait.

To justify a non-zeroed graph you have to be able to say that what you cut off wasn't valuable data. Voltage can make sense to use a non-zeroed graph because cpus operate in a band of voltage so it's genuinely irrelevant to show all the way down to a voltage the cpu can't even turn on at.

But when displaying work done and hiding 2/3 of the work done to distort the tip, well that's a visual lie about performance.

There is one single graph that shows work done, its the first one for maximum bait. This is it when you display the other 2/3 of the work and give it more appropriate dimensions:

tmp2.png


Ah yes Roman, lets put it into perspective and look at the bigger picture as you said at the end of the video :p
 
Yeah, I'm still not really sure what the point is, does he mean literally out of the box ? Because this isn't what he's tested. Is it with some overclocking, well hasn't that been true forever with any chip they all will hit different speeds over the rated spec.

There's a difference in Cinebench r23 on my chip of 1600 points between stock and just being a noob playing with Ryzen master. What dictates the quality of the chip ?
 
Ha, OK, not literally. I mean't chips used to be simple, they would have a speed, and you would get that speed, or it was defective. You might then overclock it and get more speed, but that was considered an extra. Now with chips boosting, one click auto overclocking even power and eco-modes, what's the expectation golden sample or defective and nothing in-between ? That's the trouble with this video.
 
Cyber-Mav said:
thinking about this more could there be variance in temps due to how the ihs is bonded to the die?
Click to expand...

I'm not sure about that one.

My understanding is that higher quality silicon/bin means that there's less leakage and that means lower voltage and lower temperatures, but I think AMD's algorithms would compensate for this by pushing more volts to achieve higher clocks and better performance, which complicates the results. If we take the best performing CPU in fps per watt (the one on the far right), it had one of the lowest temps, joint highest clock, 2nd lowest voltage and highest score (in Cinebench).

On the face of it, the 4th from the left seems to do quite well too, it has low temperatures and power consumption (in Cinebench), but it appears not to be a high quality silicon/bin, because it needs mega voltage and doesn't achieve a high clock speed, so maybe the algorithm is struggling with keeping it under control and the lower temps and power are actually a result of throttling.
 
Cyber-Mav said:
Thats a massive difference in binning quality control.
Click to expand...

There is no "binning" :)

There is an AMD warehouse manager in the comments making a good point, the 7600X sells in massive numbers, to keep up with those sales AMD are just grabbing whatever roles off the production line, they are not allocating bins for higher or lower end they are just using them all as they are for the low end.

The thing with that is if they did bin them you would get a low end one (Binned as not good) its a low end CPU, you don't get the best bins in the cheapest CPU's do you? He does mention this, so he almost had it.... but no it went over his head. (*Critical thinking) because they don't you have a chance of getting a really good one.

My 5800X was destined to be a 5950X, it has 2 CCD's, as a result it clocks really quite nicely for a 5800X, its better than most i have seen.

Once again Roland, or whatever his name is thinks he has another AMD has problems video but fails to realise most people are capable of something he isn't, *critical thinking. Because AMD sell so many of them you could be in for a bonus with AMD not binning them....

I'll say it again as i find myself doing these days with so many of his videos, the man is an idiot.
 
Last edited:
Hotwired said:
What he's said is true and what he's showing is clickbait.

To justify a non-zeroed graph you have to be able to say that what you cut off wasn't valuable data. Voltage can make sense to use a non-zeroed graph because cpus operate in a band of voltage so it's genuinely irrelevant to show all the way down to a voltage the cpu can't even turn on at.

But when displaying work done and hiding 2/3 of the work done to distort the tip, well that's a visual lie about performance.

There is one single graph that shows work done, its the first one for maximum bait. This is it when you display the other 2/3 of the work and give it more appropriate dimensions:

tmp2.png


Ah yes Roman, lets put it into perspective and look at the bigger picture as you said at the end of the video :p
Click to expand...
Well spotted....

"Oh its easier to digest" Well yes but everyone knows the ONLY reason to magnify graphs is to make it look worse than it actually is, its "easier to digest" because the difference are so small you can't "easily" see them if you don't magnify the graph, in other words if you have to magnify the graph so that people can see what your trying to convey then what you're trying to convey isn't really there.
 
Last edited:
humbug said:
There is no "binning" :)



My 5800X was destined to be a 5950X, it has 2 CCD's, as a result it clocks really quite nicely for a 5800X, its better than most i have seen.
Click to expand...

Yea I I got lucly with mine as well.

Although I do believe that in general the quality of the 5800x's were better than the 5900x's, the latter of which tended (but not always) to have the odd "ropey" core ot two, where as the 5800x's tended to have a better more consistent set of 8.
 
After reading this thread and watching the video I’m struggling to orientate myself. My new AMD CPU is faster than what was shown in the reviews.

Should I send AMD a check or something?
 
You must log in or register to reply here.
Back
Top Bottom