VideoCardz: AMD Ryzen 9 3950X to become world’s first 16-core gaming CPU

Status
Not open for further replies.
This is where it gets a bit interesting for AMD. AMD’s version of “TDP” isn’t comparable to Intel’s. You can’t point to a 95W TDP on the 8700K and compare it to a 95W TDP on the 1700X. The numbers are reached in different ways.

For AMD, TDP is calculated by subtracting 42 from 61.8 and dividing by 0.189. Those numbers are derived from the following: AMD claims that the “optimal” tCase temperature is 61.8 degrees, hence 61.8. AMD also says the optimal ambient temperature at entrance to the heatsink fan is 42 degrees at the inlet. They also say that the minimum degrees per Watt rating of a heatsink to achieve rated performance should be 0.189 thermal resistance.
Click to expand...

https://www.gamersnexus.net/hwrevie...w-game-streaming-cpu-benchmarks-memory/page-3
 
Gideon said:
View attachment 167752

Hmm odd that looks nothing like 140 watts or even close to it. You keep trying to peddle that lie but no one is buying it.
Click to expand...

Tomshardware isn't measuring power. They simply reporting the reading of the internal sensors of the CPU. Not only their numbers not agree with real measurements or real cooling requirements (e.g. the rating of official coolers), but the numbers published by Tomshardware not even agree with readings at the wall. :rolleyes:

OrangeKhrush said:
https://www.gamersnexus.net/hwrevie...w-game-streaming-cpu-benchmarks-memory/page-3
Click to expand...

We already know that AMD marketing TDP values are derived from an invented formula that doesn't represent real TDP: "AMD marketing formula for TDPs is lies." Read the thread.
 
juanrga said:
Tomshardware isn't measuring power. They simply reporting the reading of the internal sensors of the CPU. Not only their numbers not agree with real measurements or real cooling requirements (e.g. the rating of official coolers), but the numbers published by Tomshardware not even agree with readings at the wall. :rolleyes:



We already know that AMD marketing TDP values are derived from an invented formula that doesn't represent real TDP: "AMD marketing formula for TDPs is lies." Read the thread.
Click to expand...

You don't get to decide what is real and what isn't. Call me when you become somebody who does, until then if I need to know how to cook a paella, you best be online so because I might drop you a PM.
 
juanrga said:
Tomshardware isn't measuring power. They simply reporting the reading of the internal sensors of the CPU. Not only their numbers not agree with real measurements or real cooling requirements (e.g. the rating of official coolers), but the numbers published by Tomshardware not even agree with readings at the wall. :rolleyes:



We already know that AMD marketing TDP values are derived from an invented formula that doesn't represent real TDP: "AMD marketing formula for TDPs is lies." Read the thread.
Click to expand...

Oh so now Toms Hardware is no good.

power.png
Screenshot_2019-06-14 Intel Core i9-7900X (Skylake-X) Review bit-tech net.png

blender-pwr-consumption.png
 
...why are you comparing the TDP for a bunch of HEDT parts....

There's no 9900K, no 2700X, and no links provided.
 
Gideon said:
Oh so now Toms Hardware is no good.

View attachment 167791 View attachment 167789
View attachment 167792
Click to expand...

Tomshardware isn't measuring power with a probe so their numbers are irrelevant. The 168W and 202W at the wall measured by BitTech and KitGuru are compatible with 110--120W at the socket level for the 1800X (whose real TDP is 128W). GamerNexus doesn't measure the 1800X, but measures the 1700 and obtains 80W in the 12V channel using Blender, which is a correct value because the real TDP is 90W.

OrangeKhrush said:
You don't get to decide what is real and what isn't.
Click to expand...

The laws of physics do. That is the reason of why the the technical docs and the coolers are rated for the real TDPs, not for the marketing values.

OrangeKhrush said:
https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

This article about ends the dispute, which is why it has become such an irrelevant topic.
Click to expand...

Ian's article is very misguided.
 
juanrga said:
Tomshardware isn't measuring power. They simply reporting the reading of the internal sensors of the CPU. Not only their numbers not agree with real measurements or real cooling requirements (e.g. the rating of official coolers), but the numbers published by Tomshardware not even agree with readings at the wall. :rolleyes:



We already know that AMD marketing TDP values are derived from an invented formula that doesn't represent real TDP: "AMD marketing formula for TDPs is lies." Read the thread.
Click to expand...

You are mistaken. Trust their own words:

https://www.tomshardware.com/review...ement-cpu-gpu-components-powenetics,5481.html

The power-consumption measurement hardware that our German colleague relies upon is quite the setup. It includes two multi-channel scopes, current clamps, slimline probes, and a 5 ¾-digit multimeter. In all, that's more than $10,000 worth of equipment.
Click to expand...

or your goto guy for disinformation (The Stilt) that gave up has no real equipment or site and is way out of date.

Adding measuring at the wall should never be used as a metric. Especially with modern VRMs.
 
juanrga said:
Tomshardware isn't measuring power with a probe so their numbers are irrelevant. The 168W and 202W at the wall measured by BitTech and KitGuru are compatible with 110--120W at the socket level for the 1800X (whose real TDP is 128W). GamerNexus doesn't measure the 1800X, but measures the 1700 and obtains 80W in the 12V channel using Blender, which is a correct value because the real TDP is 90W.



The laws of physics do. That is the reason of why the the technical docs and the coolers are rated for the real TDPs, not for the marketing values.



Ian's article is very misguided.
Click to expand...

oh i am sorry, I am pretty sure nobody thought thermal physics was considered when they designed these CPU's, the world is a better place now that you have arrived to tell us what is not already obvious.

So we can now sum up the bs of the last 2 pages, TDP is not total system power, derp and TDP is actually rated for thermal dissipation and related more to cooling derp.

I trust Ian Cutress more than you.
 
OrangeKhrush said:
https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

This article about ends the dispute, which is why it has become such an irrelevant topic.
Click to expand...

Ian Cutress at Anandtech said:
It should be noted that up until this point, almost every stock test in every CPU review in existance has been run at 'out of the box', and NOT 'Intel Recommended'.
Click to expand...

That pretty much rebuts your point, as again, it's up to the reviewers to verify settings and performance.
 
OrangeKhrush said:
PL1 state is at rated, PL2 is like 100w higher, the fact that intel doesn't even ship a stock cooler is enough to suggest they are playing bullshit with your TDP dreams.
Click to expand...

AMD doesn't ship coolers on many enthusiast CPUs either...?

I don't see the issue here with Intel. They've rated their parts, board makers have done something different.
 
IdiotInCharge said:
AMD doesn't ship coolers on many enthusiast CPUs either...?

I don't see the issue here with Intel. They've rated their parts, board makers have done something different.
Click to expand...
The 2700 ships with a Wraith Spire, the only ones that don't have coolers is the Thread ripper SKU's
 
schmide said:
You are mistaken. Trust their own words:

https://www.tomshardware.com/review...ement-cpu-gpu-components-powenetics,5481.html



or your goto guy for disinformation (The Stilt) that gave up has no real equipment or site and is way out of date.

Adding measuring at the wall should never be used as a metric. Especially with modern VRMs.
Click to expand...

That article mentions a new power measurement system that will be installed in their US labs and will be used for future reviews. The review of the 1800X was made before that and it didn't measure power but simply reported reading of CPU sensors. So your post is irrelevant.

Not only the Stilt did bring us one of the more rigorous and extensive reviews of Zen and Zen+, but his power measurements agree with other reviews, with cooler ratings, and with technical docs.
 
juanrga said:
That article mentions a new power measurement system that will be installed in their US labs and will be used for future reviews. The review of the 1800X was made before that and it didn't measure power but simply reported reading of CPU sensors. So your post is irrelevant.

Not only the Stilt did bring us one of the more rigorous and extensive reviews of Zen and Zen+, but his power measurements agree with other reviews, with cooler ratings, and with technical docs.
Click to expand...

To what length will you go. LOL. Do you really think they wrote the article before they started using the equipment. They even mention that the equipment was used to measure the 480s pcie overdraw. circa 2016.

In fact, we were the first to break the news of Polaris’ tendency to draw too much current from the PCI Express slot in our AMD Radeon RX 480 8GB review, thanks to this hardware.
Click to expand...

They have the equipment, they use it.
 
schmide said:
To what length will you go. LOL. Do you really think they wrote the article before they started using the equipment. They even mention that the equipment was used to measure the 480s pcie overdraw. circa 2016.



They have the equipment, they use it.
Click to expand...

Yep. They expected two years to write the article and their 1800X review reported "package power consumption" because that is what one measures in the PSU :rolleyes: Let us also ignore this:

The only way to see 104.7W ish "Package Power" quoted by Tom's Hardware is to actually manually limit the PPT to 105W, instead of the default 141.75W (which is used when PBO is disabled).
Click to expand...

Let us further ignore the other four reviews mentioned here, common sense, the technical docs, and the cooler ratings.
 
AMDs death was greatly exaggerated. 3950x would be cool too own but, I am content with my 1600, 2600 and 1700.

Intel has problems and this time, they cannot buy their way out of it or hope for an Isreali miracle. Also, Hector Ruiniz is not in control this time.
 
Well I see he's gone into fantasy land now and complete denial. I can see why your single again and now back here annoying us with your excuse train why some chart or benchmark that shows Intel in a bad light is all wrong.
 
ManofGod said:
AMDs death was greatly exaggerated. 3950x would be cool too own but, I am content with my 1600, 2600 and 1700.

Intel has problems and this time, they cannot buy their way out of it or hope for an Isreali miracle. Also, Hector Ruiniz is not in control this time.
Click to expand...

Intel will be fine, just as AMD has been fine to this point as you indicate. Sometimes one steps ahead of the other, and it would be a mistake to ever count one or the other as out of it. Especially when as you point out, a competent CEO can help drive things in a much healthier direction. I really like Dr. Su's no-nonsense approach, and I think she is much better at making high level go/no-go decisions than Ruiz.

Isn't it awesome to even be able to have a real discussion on what parts might be better again?

There, I'm done with positivity for today. ;)
 
Before we go further down this rabbit hole.

PSUs are typically 70-85% efficient, so power measured at the wall is 15-30% more than actual board usage.

VRMs are typically 90% efficient, so power measured there is typically 10% more than socket usage.

A good motherboard VRM talk. Gamers Nexus' Buildzoid



So no matter how well you measure things, the processor gets less than what you're able to measure.

Moreover when dealing with TDP you have a system with both thermal capacity and resistance. Both companies allow the systems to run over power to achieve better performance.

None of this should be shocking news.

IMO TDP should be seen as a minimum not a maximum.
 
TDP being THERMAL has been decoupled from power for years.
AMD TDP makes sense in its own right, so does Intels. Since their definitions are not a secret, then in context they make sense.
Not even Intels white papers dispute AMDs as 'lies' or otherwise... The issue does get complicated... Long gone are the days of a 1 ghz cpu was 1ghz all the time and used x amount of watts, and dissipated x amount of heat.
To call one lies and not the other is just dishonest. BOTH represent less POWER than what the cpu can soak up. I don't know how good stock coolers are, but if for both the cpus can performed as sold with the stock cooler, then its a proper pairing. If you buy an equally rated cooler as the cpu and the cpu works as advertised, then the tdp value is true. Power its a different issue, if you buy a 95w cpu you are not going to buy a ps with 95w to the cpu.. if you don't know that, you shouldn't put together your own pc.
 
What a fascinating discussion about TDP on Intel CPUs in an AMD thread. Gee what caused that?

Couldn't possibly be the Intel shill that even Linus Torvalds called BS on. Could it?


On topic: I am really looking forward to seeing what kind of power efficiency can be wrung or of these parts, as Zen+ was already pretty good when undervolted.
 
I like my 1400 at 65 watts .. but it can not push CX 570 like X58 can when overclocked even with older PCI 2.0 but I have learned that ReLive does affect my overall scores as the older cpu does not like the foot print ..I have a Ready Cache SSD drive I was thinking of adding to help the old cpu a little in that area ..

http://www.3dmark.com/fs/17512146

http://www.3dmark.com/fs/19125137

debating on adding 3rd RX 570

 
Last edited:
schmide said:
Before we go further down this rabbit hole.

PSUs are typically 70-85% efficient, so power measured at the wall is 15-30% more than actual board usage.

VRMs are typically 90% efficient, so power measured there is typically 10% more than socket usage.

A good motherboard VRM talk. Gamers Nexus' Buildzoid



So no matter how well you measure things, the processor gets less than what you're able to measure.
Click to expand...


We know, and the looses are already considered when giving the power actually draw by the CPU alone. That is why we read more than 128W in the 12V channel for the 128W 1800X, for instance.
 
Talking about TDP in 2019, this is funny. I can imagine running Intel recommended TDP benches, there will be 3.6ghz tests with Cinebench 1T scores around 150 and the earth will be a black void of depression.
 
DooKey said:
If it's $600+ then very few gamers and/or mainstream users are going to buy that processor and motherboard. It will sell as well as Intel HEDT systems...not very well.
Click to expand...
What? Intel's HEDT sold very well. And AMD has taken a big chunk of those numbers. Now they are offering a 600+ chip and less then half the price of slower/hotter chip's? AMD has just taken a majority of Intels HEDT sales and it is HUGE for them. With the new ThreadRipper's coming, Intel has just lost ALL OF their highest margin non enterprise segments.
 
Azrak said:
The "problem" is that the 8C CPU chiplets won't be the same high level binned chiplets found in the 12C and 16C CPUs, so getting the 8C CPU to overclock to 4.7 GHz on 1-2 cores cores like the 16C does out of the box won't be easy or might be impossible due to silicon lottery.
Click to expand...
There is little to no "silicon lottery" with any of the Zen's. Cool adequately, and they are all within a few 100 MHz. Your paying for out of the box speed/OC. and that is fine.
No magic here!
 
Status
Not open for further replies.
Back
Top