AMD Ryzen 7000 Undervolting Yields Great Results with Temperatures

[TheoneandonlyMrK]

Can we do nuclear power station and lava jokes about AM5 now? Or will it be frowned upon by the AMD crew that did it RE Intel ADL /s for @thegnome

See post no. 2, or did you comment without reading any posts? Maybe i will edit my post with a /s for sarcasm just for you.

You can't win that fight on TPU its about 85-15% in AMD users favour.

How does that us V them shit start indeed.

And that last one is rubbish.

More own Intel than AMD, some own both but your ratio is pulled from your butt.

 

[Berfs1]

Alrighty to answer everyone's questions on why the voltage is so high stock.. there is something called "yields". Basically AMD for example sets a certain clock frequency/voltage threshold for a certain SKU, any potential CPU that doesn't meet the required clock frequency at the required voltage would be considered a failure in a batch (assuming they don't downbin it). Yields are directly related to binning (and by extension overclocking). A manufacturer sets a higher than needed [for most samples] stock voltage so that their yields (% of passing samples) can go up. A high bin CPU doesn't need as much voltage to hit a certain frequency as the majority of other samples.

That being said, what appears to be going on is, that tester has a high bin chip, and therefore can undervolt pretty significantly.

 

[R0H1T]

The issue being discussed here is the extra voltage that motherboards apply to most CPU's these days, that's not exactly how AMD does it & yes it's also an issue with Intel. The stock voltages are more often than not insane!

Some mobo brands do it to make their boards look better(stable?) while others do it to enable higher clocks out of the box for corresponding CPU's which go in there.

 

[Solid State Brain]

The issue being discussed here is the extra voltage that motherboards apply to most CPU's these days, that's not exactly how AMD does it & yes it's also an issue with Intel. The stock voltages are more often than not insane!

Modern CPUs do have a built-in voltage-frequency curve that motherboards use, but there also are additional load-based voltage corrections (on Intel, "AC Loadline") that can be applied on top of this by the motherboard to factor power delivery limitations or increased requirements due to overclocking. For instance, my 12700K under default "optimized settings" with my MSI Pro Z690-A was using about 100 mV more than it should have at full load (>1.3V instead of ~1.2V or less at 4.7 GHz), causing significantly higher power consumption and temperatures close to throttling (with a high-end air cooler). I wonder how often something along these lines has affected results in the reviews, putting aside that significant differences in terms of required (and built-in) voltages between two identical CPUs may also exist and are never checked for.

 

[JustBenching]

Strange how every intel troll already saw the review of Zen4.

Do I need a review to conclude that the 7600x needs to be at least 60% faster than the 5600x to actually match the 12600k in CBR23? Really now?

What happens when the reviews come out and im correct, will you come back and apologize - admit you are the troll here?

 

[TheoneandonlyMrK]

The 12600k beats the crap out of the 7600x in mt workloads at similar power levels. What the heck are you talking about? That's the ALDERLAKE i5. I don't know why im trying to have a reasonable discussion with the amd fanbase, ***k it man.

Your casting opinions based on leaks as rock solid and hyping the shit out of it.

And think that's reasonable.

After reviews are released and some Verified facts appear, Then a reasonable discussion could be had, I say could, but given your tone and general output it's not likely.

 

[AM4isGOD]

What happens when the reviews come out and im correct, will you come back and apologize - admit you are the troll here?

As If

 

[Denver]

Do I need a review to conclude that the 7600x needs to be at least 60% faster than the 5600x to actually match the 12600k in CBR23? Really now?

What happens when the reviews come out and im correct, will you come back and apologize - admit you are the troll here?

Anyone can say anything, including that the R5 7600X can beat i5 AL in games using only 50-60w.

 

[JustBenching]

Your casting opinions based on leaks as rock solid and hyping the shit out of it.

And think that's reasonable.

After reviews are released and some Verified facts appear, Then a reasonable discussion could be had, I say could, but given your tone and general output it's not likely.

I already said - based on the leaks. But even without them, we are not going to have a 60% performance jump between a 5600x and a 7600x. It's literally impossible.Therefore it's safe to assume that the 12600k will beat the 7600x in CBR23, and not by a small margin.

Anyways, the discussion started with his claims that Intel is pushing power consumption, when it's the exact opposite, amd is the one pushing it, and in a big way. So, deal with it.

Anyone can say anything, including that the R5 7600X can beat i5 AL in games using only 50-60w.

Anyone can say anything, but unless supported by some data - he can equally be just laughed out of the room. Any data for your claim? Cause up till now, ald > zen 3 in gaming efficiency and performance.

 

[Denver]

I already said - based on the leaks. But even without them, we are not going to have a 60% performance jump between a 5600x and a 7600x. It's literally impossible.Therefore it's safe to assume that the 12600k will beat the 7600x in CBR23, and not by a small margin.

Anyways, the discussion started with his claims that Intel is pushing power consumption, when it's the exact opposite, amd is the one pushing it, and in a big way. So, deal with it.


Anyone can say anything, but unless supported by some data - he can equally be just laughed out of the room. Any data for your claim? Cause up till now, ald > zen 3 in gaming efficiency and performance.

The subject Isn't an ES CPU running at 67w @ 5Ghz over there? In games the power consumption should be even lower.

If you really believe that intel can beat 5nm efficiency, you should laugh at yourself.

 

[Darmok N Jalad]

Q: Why are power levels going up when chip features sizes are going down, I would have thought that less electrons/power would be needed to do the work on smaller transistors?

You got a few reasons—yes, new nodes do bring better efficiency to do the same amount of work. You can strictly leverage this and consume less power than before. However, with those gains, you can add performance to your design and use the same power as before. I bet if you clock the 7600 to perform like the 5600, you’d see a power savings. Or, you can clock the 7600 to consume the same power as the 5600 and you’ll get a performance increase. Or, you can clock the 7600 to an even higher power limit, should the design allow, and get even more performance. Zen4 has been designed for both IPC and clock speed gains over Zen3, at the expense of additional max power consumption.

Another variable is that when your competitor uses more power consumption to advance performance, which Intel has done over the last few generations, you eventually have no choice but to raise your own, or pull off a miracle design that can perform without that power. It’s just not out there—this is leading edge engineering, and there’s no low-fruit left to pick. If we were still talking 95W caps, the landscape would look quite different.
I’m sure that, should you wish, you can cap Zen4 to 95W and still see performance gains over previous generations, but it’s not the game anyone is playing in the industry anymore. Considering the existing and looming energy struggles, it’s a peculiar thing these companies are chasing. In California is asking folks not to charge EVs in the evening during a heat wave, it seems silly to have a 1kW gaming rig running instead. I’m not necessarily against consumption, mind you, but the grid doesn’t seem like it’s becoming more accommodating of high-consumption electronics—it’s going the other way!

 

[usiname]

Do I need a review to conclude that the 7600x needs to be at least 60% faster than the 5600x to actually match the 12600k in CBR23? Really now?

What happens when the reviews come out and im correct, will you come back and apologize - admit you are the troll here?

Now point please where I said something for performance of Zen4 over AL, just one single word. Oh wait, your arguments how 7600x will be 30% faster than 5600x(YES 5600x) with double consumption went in the trash and you changed the topic from efficience of 5600x vs 7600x to 12600k performance over 7600x.
By the way, how 12600k will be more efficient than 7600x when it is not even more efficient than 5600x

 

[JustBenching]

Now point please where I said something for performance of Zen4 over AL, just one single word. Oh wait, your arguments how 7600x will be 30% faster than 5600x(YES 5600x) with double consumption went in the trash and you changed the topic from efficience of 5600x vs 7600x to 12600k performance over 7600x.
By the way, how 12600k will be more efficient than 7600x when it is not even more efficient than 5600x
View attachment 260514

Your own question answers itself. According to the leaks, the 7600x scores 15k @ 130watts

But arguing with fans is so tiring. Ill leave you to it man

 

[usiname]

Your own question answers itself. According to the leaks, the 7600x scores 15k @ 130watts

But arguing with fans is so tiring. Ill leave you to it man

Better fan, than troll who claims that while 5600x is doing 11500 with 80w, 7600x will score 15k with 130w (65% more watts for 30% more performance)
Congratulations, you are the most clueless person in the thread!

 

[JustBenching]

Better fan, than troll who claims that while 5600x is doing 11500 with 80w, 7600x will score 15k with 130w (65% more watts for 30% more performance)
Congratulations, you are the most clueless person in the thread!

I don't claim anything, that's what the leaks suggest.

And yeah, it makes sense that a 60% increase in power will lead to a <60% increase in performance. That's a given.

 

[usiname]

I don't claim anything, that's what the leaks suggest.

And yeah, it makes sense that a 60% increase in power will lead to a <60% increase in performance. That's a given.

You are beyond clueless...
Do you realise that you comparing two different architectures on different nodes? Imagine even on same power the new architecture will have better performance/watt.
This is the last time when I lose time with you. Welcome to my block list

 

[TheoneandonlyMrK]

I already said - based on the leaks. But even without them, we are not going to have a 60% performance jump between a 5600x and a 7600x. It's literally impossible.Therefore it's safe to assume that the 12600k will beat the 7600x in CBR23, and not by a small margin.

Anyways, the discussion started with his claims that Intel is pushing power consumption, when it's the exact opposite, amd is the one pushing it, and in a big way. So, deal with it.


Anyone can say anything, but unless supported by some data - he can equally be just laughed out of the room. Any data for your claim? Cause up till now, ald > zen 3 in gaming efficiency and performance.

I'm sure I can deal with whatever happens, this is a hobby chill your strudels.

And again wrong, I disagree on who started pushing power first this decade, or are we looking back all the way to FX.

 

[nexxusty]

This is BS.

There is no actual VCore adjustment shown, just the "wattage difference", which obviously equates to a temperature difference.

There's no way a Ryzen 7000, any of them, will get to 93c at stock, as long as the cooler used is appropriate.

Period.

 

[lexluthermiester]

That's a ton of voltage at stock for no reason.

I have to agree. AMD has been doing this since first gen Ryzen and still have not explained why..

There's no way a Ryzen 7000, any of them, will get to 93c at stock, as long as the cooler used is appropriate.

Period.

You don't and can't know that about an unreleased product, so clam up.

 

[ARF]

Oh, dear

New reports regarding the thermal performance of AMD's Ryzen 7000 CPUs have been shared by Enthusiast Citizen over at Bilibili & it looks like the upcoming Zen 4 chips will require serious cooling hardware to keep them tame.

The leaker who has been very reliable with his past information and leaks has stated that AMD's Ryzen 7000 Desktop CPUs based on the Zen 4 core architecture will be some of the hottest chips produced to date. The leaker talks about two specific chips, an AMD Ryzen 9 7950X & the Ryzen 5 7600X. Do note that the information shared is based on ES/QS samples so final results may vary.

AMD Ryzen 7000 CPUs Reportedly Run Hot, Ryzen 9 7950X Hits Up To 95C Thermal Threshold at 230W, Ryzen 5 7600X Up To 90C at 120W (wccftech.com)

 

[AM4isGOD]

Looks like i will be keeping my 12700k or dropping in a 13700k.

 

[mahirzukic2]

Looks like i will be keeping my 12700k or dropping in a 13700k.

I still think we need to wait for the thorough reviews before making any decisions.

 

[TheoneandonlyMrK]

"The leaker talks about two specific chips, an AMD Ryzen 9 7950X & the Ryzen 5 7600X. Do note that the information shared is based on ES/QS samples so final results may vary."

Makes sense not, then I read ES/QS.

Oh.

 

[Mussels]

If you don't like the fTPM bug fixed, or the usb disconnects fixed, or the botched sata speeds sure, you don't need to bios update


I thought the 7950x runs at 240w, didn't know it's 125w. That's great news. Any source?

Those bugs were introduced in updates, so you can easily stay on older ones

You can easily stay on BIOS updates before the windows 11 changes, manually enable TPM and secureboot and be problem free

 

[BRNNN]

Okay okay, I'll ask you, where is the logic in selling you unlocked processors at different prices? Exactly the same processors at a different price?
You have to see this, and everything will be clear to you. Check out AMD fx-8370, fx-8350, fx-8320 at Cpu World. "Boosted p States", "Low power P states". After that, look for the fx-8370E. Same processor but TDP 95Watt vs 125Watt. Anyone who doesn't understand what this is about is crazy