Of course I can, it's not that hard. I don't get why you are asking me to do it, but I'll do it anyways.
because you made that claim, there for, you need to prove it, not me.
Nah, thank God it's not on me actually. Hub himself admitted on his Twitter that the graph is wrong because he was using locked voltages.
ok, then screen shot it, and post it, i dont have twitter, so i cant see it
As I've said, you don't understand the voltage / power scaling, that's why you can't tell why that graph is wrong.
Did you link the wrong thing? because the TPU link is just Cinebench MT scores.
No, it's the right link. It's the power scaling of a 12900k. The numbers are off by up to 70% in some of them measurements.
Given they use the same "bracket" mounting holes as LGA1700, I bet beside cost it could also be related to the socket footprint to not interference with the existing coolers, especially given it potentially be sort of a short lived socket. IMO using the RL version should be good enough to not making stuff like HBC noctua cooler being needed for best performance
So you're just saying the scores are lower than they should be then?
Well I'm not suggesting that they are lying, but yeah the numbers are completely off.
At the linked screen cap, at 65W it is 18265 and at 125W 22818 which isn't far off and isn't unreasonable for run to run difference with maybe some minor stuffs running at background, and that isn't 70% difference..
I'm talking about the first link from TPU. You are probably talking about HUB's graph. That one is also wrong. A 13900k at 125w scores 31-32k, not 22. It scores 22k at 65w.
where are the reference of these no. comes from? don't say from someone's single chip with undervolting
There is no undervolting involved. I've tested my chips and lots of reviewers have verified the numbers I'm giving you.
I have the 14900k, which have a lower VF curve and boost limits, and I don't bother to test with the old days VF curve that a 13900k might have to verify. Just to say, with the latest fixed bios, without any thinkering and undervolting, the 14900k of mine got 36k at 253W, much like what Buildzoid did, and it is logical to think that a stock 13900k shouldn't get that scores the old days reviews gets with the new, intel spec LLC
that is a useless analogy, using what essentially a 13900KS with much lower VF curve to lock to 125W don't mean a thing.
Dude what the heck are you talking about? Your "13900ks" isn't 50% better than a 13900k, and yet youll score a 50% higher score cause hub's scores are just flat out wrong. I mean he admitted himself they are wrong, what the hell are you on about?
Do you even know what he says he think he was wrong about? It’s that his MSI board seems to be aggressive on voltage with power limit, and in layman’s term, it’s having a different. V/F behaviour. So a sustained higher voltage will roll off the score as it hits the power limit with less current, hence less performance.
That's what he thinks it is, but it's not. It's the same issue TPU had on their own graph. MSI board on auto used to have fixed voltage for the CPU. Power limiting doesn't drop the voltage even though it's running lower clocks. That's why their efficiency remains steady across different power limits.
You also need to remember that on auto is what reviewers should do, as they normally use the same brand of boards over different platforms, if MSI locked for Intel, they also is locked for AMD, so still kinda valid for cross platform comparison and for ppl using those boards. As long as they are publishing what boards/memory they are using I don't see the problem there. You won't know if using Asus will have some extra MCE or so which say, burns the X3D chips.
Facebook
Twitter
Instagram