90c+ CPUs

[freeagent]

Shawn, AMD measures TDP at the base clock frequency.

The PBO enabled at default is only a perk to the boost algorithm.

Yessir!

I was under the impression that AMD got their tdp at full load, stock setting. So on my 5900x it would be 3700MHz. Which at stock setting it will drop down to those clocks with a heavy load. Even if that heavy load doesn’t bring the cpu to 60c.

Stock settings suck.

 

[AusWolf]

You are missing the point of what AMD is declaring with their TDP number

They are not telling that power is same thing with temperature.
Nowhere near that. This is applied physics (thermodynamics).

They are telling to the cooler manufacturers that under a specific tCase (CPU IHS surface -lid- temp), a specific tAmbient (ambient air temp on inlet of the cooler's fan) and a specific thermal resistance cooler (HSF θca) an X amount of heat will be transferred from the IHS to the ambient air.
So cooler manufacturers have to design coolers that comply (to a minimum) to those parameters.

In order to understand this you have to understand first how heat transfer works. Roughly It involves a temperature (t)Delta between 2 objects/surfaces and the thermal resistance of what is in between them.
The higher the tDelta as the temp difference between 2 objects and the lower the heat resistance between them the more heat will be transferred and vise versa.

As tDelta they are meaning the CPU IHS surface temp (tCase) and the air temp on the inlet of the cooler fan (tAmbient).
Thermal resistance of cooler is HSF (HeatSink-Fan) θca (theta c a, measured in °C/W)


(tCase - tAmbient) / (HSF θca) = TDP
In their formula they are taking as granted the tAmbient will be 42C (as an inside of a case that has been wormed up by the system)

View attachment 263747

So cooler manufactures on minimum they have to come up with a HSF design that can maintain the CPU IHS surface at 61.8C under 42C ambient, when the CPU is on maximum power draw. In case of 3900X is the 142W (PPT).
So... if they slap on top of the 3900X a HSF with thermal resistance at least of 0.189 °C/W the transferred heat will be ~105W.

Reasonable question:
But the CPU is "producing" 142W of heat as all the electrical energy is transformed into thermal energy... Where is the heat between 105W and 142W?

This is not mentioned anywhere but you have to think that this system (CPU, HSF, ambient) is not a closed/sealed one from the rest of the world.
The CPU is in contact with the board through the socket and also some air is around the outer edges of the CPU.
So the rest of the heat 142-105= 37W will be dissipated by the board from the surrounding components and the back of it. Yes boards are hot around the socket of a 142W CPU on max power.
And this should be true on those parameters (tCase/tAmbient/θca) AMD specifies.

Here is AMDs claims of TDP:
"TDP is about thermal watts, not electrical watts. These are not the same.
TDP is the final product in a formula that specifies to cooler vendors what thermal resistance is acceptable for a cooler to enable the manufacturer-specified performance of a CPU.
Thermal resistance for heatsinks is rated in a unit called θca ("Theta C A"), which represents degrees Celsius per watt.
Specifically, θca represents thermal resistance between the CPU heatspreader and the ambient environment.
The lower the θca, the better the cooler is.
The θca rating is an operand in an equation that also includes optimal CPU temp and optimal case ambient temp at the "inlet" to the heatsink. That formula establishes the TDP.

Thermal design power

linuxreviews.org

I cannot break it down better than this, I'm sorry.

I understand all of this. I watched every video and read every article when they came up with this. I still think that it's overcomplicated and misleading for customers.

In an ideal world, there should be a consensus of what TDP means, and no manufacturer should be allowed to come up with their bullshit formulae to manipulate public perception.

 

[Zach_01]

I understand all of this. I watched every video and read every article when they came up with this. I still think that it's overcomplicated and misleading for customers.

In an ideal world, there should be a consensus of what TDP means, and no manufacturer should be allowed to come up with their bullshit formulae to manipulate public perception.

Cant argue with that

Yes exactly.

Convert Kilowatts To Horsepower (kW to HP Calculator)

Kilowatt to horsepower conversion is one of the most common ones. Here’s how you can convert kW to HP: 1 kW = 1.341 HP The easiest way to remember the kW to HP conversion formula is to remember how many kilowatts is 1 horsepower of energy: 1 HP = 3/4 kW We’ll convert kW to ... Read more

learnmetrics.com

Thermal Design Point?

I think wattage converted to BTU is directly on point.

I'm using the ICE as an energy usage, which is the argument that a cpu is 100% energy efficient. Which in fact is the complete opposite.

We waste a lot of energy to store data code.
Then to process it.

Perhaps the example does give a twist, but if YOU think a cpu is 100% efficient like the other gentleman, you're both just plain dead wrong.

I'm done here then. It's no problem Zach. Wasn't trying to draw out a debate, I'm being quoted, which means you guys want replies.

So there's my links to back my statements above.

Enjoy, peace.

This is a different discussion really from what TDP is or it should be or what every company calls it to be to suit its marketing.

We measure VRM efficiency/inefficiency by how much heat they output from the input/output power at any given amount and we draw curve about it.
On the other hand CPU power and every chip in general that output heat is a little more complicated than this simple approach I think.

 

[Deleted member 185158]

Cant argue with that


This is a different discussion really from what TDP is or it should be or what every company calls it to be to suit its marketing.

We measure VRM efficiency/inefficiency by how much heat they output from the input/output power at any given amount and we draw curve about it.
On the other hand CPU power and every chip in general that output heat is a little more complicated than this simple approach I think.

The TDP of a cpu is the cpu only from anything I've ever read.

Typically for processors, I try to view the white sheets.

 

[Toothless]

Ok. interesting. I found before overriding defaults for PPT, TDC, and EDC I can independently do an all core -20 so I have some headroom to play with. 4.4ghz all core and 4950mhz is pretty similar to what I was getting but I need to improve how I'm documenting and observing frequency to know for sure. Depending on how I balance things can get 4.525ghz to 4.575ghz all core or 5.025ghz on a few cores depending on how I balance PPT, TDC, and EDC before curve optimization.


Yea it seems depending on the combination I try I either stay around 60c or spike above approaching 80c. I have to admit I was getting impatient last night during testing various combinations but also it seems if I set PPT too high in relation to EDC it blows past my set EDC limit (ignoring it ) making testing more problematic. I'm not on the latest BIOS/UEFI at the moment and maybe I need to be in order to do this properly.

I let it hit 80c. I've come to terms that minus a custom loop, it'll hit that.

I followed a guide on Reddit for my overclock stuff, I can DM it if you'd like.

 

[A Computer Guy]

I let it hit 80c. I've come to terms that minus a custom loop, it'll hit that.

I followed a guide on Reddit for my overclock stuff, I can DM it if you'd like.

I might have come across that before but sure send it to me.

 

[RJARRRPCGP]

5800X3D feels like a locked Intel CPU lol

That's why I'm hesitant to get one of those, albeit it's not off the table yet, at least for a future AM4 build.

 

[tabascosauz]

That's why I'm hesitant to get one of those, albeit it's not off the table yet, at least for a future AM4 build.

I should clarify, it's only like a non-K Intel CPU in that multiplier and Fmax are locked (and Curve Optimizer/PBO are not accessible unless you use PBO2 Tuner). It doesn't run anywhere near as cool as a 12400 would it takes a lot of work to tame, even more than 5800X for those above reasons.

If you haven't already gotten one, just wait for the real Vcache products on AM5 when they've worked all the refinements in. This one is just a technology demonstrator.

 

[freeagent]

This one is just a technology demonstrator.

Might be a collector's item someday

 

[Zach_01]

Might be a collector's item someday

Might be the first to go off the market also next year

The TDP of a cpu is the cpu only from anything I've ever read.

I’m not saying it’s not.

We are on different page?

VRMs was en easier example of what efficiency is on the subject:
Is the CPU 0~100% efficient?

Same as:
Is an incandescent light bulb 0~100% efficient? What about LED light bulbs?

 

[freeagent]

Might be the first to go off the market also next year

That’s ok, I have one for my collection

 

[Zach_01]

No way a cpu is 100% efficient. You wouldn't dissipate heat because the electrical energy ....

Is watts

Is converted

To BTU

Thus

TDP

THERMAL design point.

Not power consumption because it uses very little of that power.

In the example, the energy used in ICE is converted as you said into kinetic energy. The rest dissipated as a heat. And why engines are also rated by the KW not by horse power. But only 30% of that energy is kinetic. The rest just wasted and dissipated as heat.


Yes you can.

It's called electrical wattage converted to BTU.

Next time, at full load, observe the wattage of the cpu. Then use the table below to see how much thermals you need to dissipate.

It's really that simple.

Watts to BTU conversion calculator

Watts (W) to BTU per hour, power conversion: calculator and how to convert formula.

www.rapidtables.com

To further add....

So if a cpu ( or anything that uses an energy source) was 100% efficient, you would have NOT the need to dissipate heat.

On the other hand CPU power and every chip in general that output heat is a little more complicated than this simple approach I think.

You cant compare how a CPU uses power like a PSU efficiency or a light bulb and how much energy is convert to light and heat. You can’t put a percentage on a CPU in that way.
I couldn’t explain it my self but I found an answer that tells all I wanted to say.

And I quote:

“When a power supply is rated as 85% efficient, what they mean is that 85% of the energy that goes into the power supply makes it to the components that it is powering, and ~15% is released as heat.

CPUs/GPUs/etc don't work by directing the energy elsewhere, so it's not clear how you would rate their efficiency. They use all of the energy they receive, and it is all turned into heat.

You could do something like number of operations per second per watt of power (and this is in fact an important factor, especially in mobile processors), but that's not a percentage.

It's a common misconception that CPUs receive energy, use some of it for computation, and release the rest as heat. This isn't true. It's the act of performing the computations that releases the heat. 100 watts into a CPU will produce 100 watts of heat.”

Source of quote:

https://www.reddit.com/r/askscience/comments/35be48
So you can’t really say that a CPU has 0% efficiency because all the power received is transformed into heat.

 

[Deleted member 185158]

You cant compare how a CPU uses power like a PSU efficiency or a light bulb and how much energy is convert to light and heat. You can’t put a percentage on a CPU in that way.
I couldn’t explain it my self but I found an answer that tells all I wanted to say.

And I quote:

“When a power supply is rated as 85% efficient, what they mean is that 85% of the energy that goes into the power supply makes it to the components that it is powering, and ~15% is released as heat.

CPUs/GPUs/etc don't work by directing the energy elsewhere, so it's not clear how you would rate their efficiency. They use all of the energy they receive, and it is all turned into heat.

You could do something like number of operations per second per watt of power (and this is in fact an important factor, especially in mobile processors), but that's not a percentage.

It's a common misconception that CPUs receive energy, use some of it for computation, and release the rest as heat. This isn't true. It's the act of performing the computations that releases the heat. 100 watts into a CPU will produce 100 watts of heat.”

Source of quote:

https://www.reddit.com/r/askscience/comments/35be48
So you can’t really say that a CPU has 0% efficiency because all the power received is transformed into heat.

So from that, putting In 100w is dissipated as 100w of heat means ZERO efficiency, because it's just simply not really used, at least in a conventional sense of the way most would think of it.

I would say that perhaps 100w in closer to 95w out because some of the power is used in some sense to have the processor simply to turn on.

Processor efficiency is generally calculated by instructions per watt, and we would compare today's processor to yesterday's processor.

Obviously at the same 100w dissipation of an 8c16t cpu is much better than a single core.....

But it can't be 100% efficient cause then a single core would be 100%, a dual core a 100% and an 8 core 100%..... how could all of these be as efficient as the other??

 

[Zach_01]

So from that, putting In 100w is dissipated as 100w of heat means ZERO efficiency, because it's just simply not really used, at least in a conventional sense of the way most would think of it.

I would say that perhaps 100w in closer to 95w out because some of the power is used in some sense to have the processor simply to turn on.

Processor efficiency is generally calculated by instructions per watt, and we would compare today's processor to yesterday's processor.

Obviously at the same 100w dissipation of an 8c16t cpu is much better than a single core.....

But it can't be 100% efficient cause then a single core would be 100%, a dual core a 100% and an 8 core 100%..... how could all of these be as efficient as the other??

What I'm saying is that you cant put a percentage of efficiency on a CPU/GPU or any other like wise chip like you do on other devices.
Its neither 0% nor 100%. Its not anything in middle either. There is not a possible way to measure this.

Like you said efficiency of a CPU is measured eventually in performance/watt and can only be used as metric to compare it to any another CPU, but this kind of efficiency is completely and utterly different of the one of a PSU or a light bulb.
CPU is no incandescent light bulb to to feed it 100W and take 2W as light and 98W as heat or even like the opposite if we are talking about LED lights.
Doesn't work like this.
You can't say I feed the CPU with 100W and I take 2W or 98W as transistor operations and whats left is heat.

Heat in a CPU is produced by its transistors operations and the whatever resistance opposed to current and maybe other stuff too that elude me right now.
So eventually all of the electric power is conversed to heat. If there was a number like the way you mean it, it would be known by now for every chip ever made.
Such a metric does not exist on a CPU, GPU etc...

I have nothing else to say about the subject and I will not make another circle which I'm feeling we are doing.

 

[Deleted member 185158]

What I'm saying is that you cant put a percentage of efficiency on a CPU/GPU or any other like wise chip like you do on other devices.
Its neither 0% nor 100%. Its not anything in middle either. There is not a possible way to measure this.

Like you said efficiency of a CPU is measured eventually in performance/watt and can only be used as metric to compare it to any another CPU, but this kind of efficiency is completely and utterly different of the one of a PSU or a light bulb.
CPU is no incandescent light bulb to to feed it 100W and take 2W as light and 98W as heat or even like the opposite if we are talking about LED lights.
Doesn't work like this.
You can't say I feed the CPU with 100W and I take 2W or 98W as transistor operations and whats left is heat.

Heat in a CPU is produced by its transistors operations and the whatever resistance opposed to current and maybe other stuff too that elude me right now.
So eventually all of the electric power is conversed to heat. If there was a number like the way you mean it, it would be known by now for every chip ever made.
Such a metric does not exist on a CPU, GPU etc...

I have nothing else to say about the subject and I will not make another circle which I'm feeling we are doing.

I feel we are agreeing to disagree actually. Fully understand the point of view....

But the wattage is going somewhere. It's not going into a CPU and then nothing happens. The wattage is dissipated as heat. Or we wouldn't need heat sinks.

The efficiency measurement is just the difficult portion of it all.

We can say today's chips are much more efficient.

So saying 5 billion transistors in operation dissipating 100w VS 500 thousand dissipating 100w.
Obviously the one with more transistors would be more efficient with the heat dissipated.

The amount of wattage used, I think is the difficult part really. It's gotta be something becuase the CPU is on and transistors are in operation.
So it would be valid to say the CPU actually consumes (uses) maybe 5% of that wattage. Because it's on and transistors are operational.

What you have in bold there.... is exactly what I'm saying. What's left is heat. Same thing with that 30% effecient IC engine. What is not converted to kinetic energy is dissipated as heat.

For technical purposes, the electrical current is really just "passed through" transistors.

So I'll quote some information, I'd love to hear your take on it. Is the wattage USED or is the wattage DISSIPATED?????

Not sure how reputable. But it's a read of basics really.

Transistor - Basics, Working Principle, Definition, Applications

Transistors have revolutionised the electronic industries since they were first invented. It is a semiconductor device made of silicon, a chemical compound commonly found in sand. Visit to learn about its working and applications.

byjus.com

A transistor consists of two PN diodes connected back to back. It has three terminals namely emitter, base and collector. The basic idea behind a transistor is that it lets you control the flow of current through one channel by varying the intensity of a much smaller current that's flowing through a second channel.

Oh hey, check this out. Found it on google pictures. Thought to look for some visuals.

 

[AusWolf]

Might be the first to go off the market also next year

That would make sense considering that that CPU is the biggest enemy of Zen 4 in gaming (not the 12900K).

 

[shadowvault]

I dunno, I have APU AMD 10-7850K and even with Cooler Master hyper gives me around 40-ish idle and around 70 full load. Is a metal, it gets warm, OP might wanna look into a better thermal paste or maybe the reading sensor got broken but doesn't explain temperature by itself unless it has a temp barometer, they have a camera nowadays I saw, things. Been away for a while from PC's but I still got my dexterity bar raised up, I've been around them since I know

 

[AusWolf]

I dunno, I have APU AMD 10-7850K and even with Cooler Master hyper gives me around 40-ish idle and around 70 full load. Is a metal, it gets warm, OP might wanna look into a better thermal paste or maybe the reading sensor got broken but doesn't explain temperature by itself unless it has a temp barometer, they have a camera nowadays I saw, things. Been away for a while from PC's but I still got my dexterity bar raised up, I've been around them since I know

Just read reviews of Ryzen 7000 series CPUs. They are nothing like the 7850K. Just because you know your 7850K, it doesn't mean you know all CPUs.

 

[A Computer Guy]

I let it hit 80c. I've come to terms that minus a custom loop, it'll hit that.

I followed a guide on Reddit for my overclock stuff, I can DM it if you'd like.

Here is what I got if your interested. I decided to give Ryzen Master a try to do the curve optimization automatically and did a quick run with CPU-z.
I'm not entirely convinced the curves are stable yet but it's enough to get through some experimenting.

Managed to hit All Core 4.674GHz at about 70c with CPU-z.

Cinebench 23 is a bit different and hits at a lower all core clock when it's running.

 

[Zach_01]

I feel we are agreeing to disagree actually. Fully understand the point of view....

But the wattage is going somewhere. It's not going into a CPU and then nothing happens. The wattage is dissipated as heat. Or we wouldn't need heat sinks.

The efficiency measurement is just the difficult portion of it all.

We can say today's chips are much more efficient.

So saying 5 billion transistors in operation dissipating 100w VS 500 thousand dissipating 100w.
Obviously the one with more transistors would be more efficient with the heat dissipated.

The amount of wattage used, I think is the difficult part really. It's gotta be something becuase the CPU is on and transistors are in operation.
So it would be valid to say the CPU actually consumes (uses) maybe 5% of that wattage. Because it's on and transistors are operational.

What you have in bold there.... is exactly what I'm saying. What's left is heat. Same thing with that 30% effecient IC engine. What is not converted to kinetic energy is dissipated as heat.

For technical purposes, the electrical current is really just "passed through" transistors.

So I'll quote some information, I'd love to hear your take on it. Is the wattage USED or is the wattage DISSIPATED?????

Not sure how reputable. But it's a read of basics really.

Transistor - Basics, Working Principle, Definition, Applications

Transistors have revolutionised the electronic industries since they were first invented. It is a semiconductor device made of silicon, a chemical compound commonly found in sand. Visit to learn about its working and applications.

byjus.com

Oh hey, check this out. Found it on google pictures. Thought to look for some visuals.

View attachment 263823

The link you provide its inaccessible to EU.
And this picture only shows that a CPU has an input of electric power and outputs the power as a heat, most of it through IHS and some of it though the socket and mainboard which is what AMD is saying indirectly with their definition of TDP against PPT and with their formula. This picture does not saying anything about efficiency.

----------------------------------

For the last time...
According to conservation of energy:
Energy cannot be created or destroyed, only can change form or be transferred to a different location or object.


In order to calculate the electric/power efficiency of a device you have to have 3 different values. 1 input and 2 output (1 input=2 output)
Out of the 2 output 1 is useful the other is considered lost.

Incandescent light bulb:
1. Input electric power (100W)
2. Light (2W) = useful
3. Heat (98W) = lost
This device has an efficiency of 2% since its used for light. The rest 98% is "lost" as heat from the conversion. Lost as not transformed into light, not really lost.

PSU:
1. Input electric power (100W)
2. Output electric power (90W) = useful
2. Heat (10W) = lost
This device has an efficiency of 90% since its used for supplying power of different voltage from the input. The rest 10% is "lost" as heat from the conversion to a different voltage. Lost as not transformed into the new voltage, not really lost.

Combustion engine with 35% efficiency:
1. Fuel+Air (Input mass)
2. Kinetic energy = useful 35%
3. Hot exhaust gasses (mass + heat) = lost 65%

Please fill the gaps below if you can
I can't...

CPU:
1. Input electric power (100W)
2. ____________ = useful)
3. ____________ = lost
This device has an efficiency of __% since its used for __________________________________. The rest __% is "lost" as heat from the conversion to _____________. Lost as not transformed into ______________, not really lost.

What 2 forms of energy are you going to put in there on 2 and 3?

That would make sense considering that that CPU is the biggest enemy of Zen 4 in gaming (not the 12900K).

I'm not convinced yet that it is really or not.



Though this is a very narrow, maybe cherry picked, selection of games.
I am waiting for a more wider game benchmarks. Same one I think its going to post a comparison of 50 games at some point.

I dunno, I have APU AMD 10-7850K and even with Cooler Master hyper gives me around 40-ish idle and around 70 full load. Is a metal, it gets warm, OP might wanna look into a better thermal paste or maybe the reading sensor got broken but doesn't explain temperature by itself unless it has a temp barometer, they have a camera nowadays I saw, things. Been away for a while from PC's but I still got my dexterity bar raised up, I've been around them since I know

OP does not have a problem with CPU thermals. Initially on this thread we were discussing the new Ryzen7000 series and its "new normal" standards of constant operation at 95C.
Please read out first 10 pages if you're interested.

 

[shadowvault]

Just read reviews of Ryzen 7000 series CPUs. They are nothing like the 7850K.

Is a powerful processor regardless, is not a athlon or intel core duo, is somewhere around a i5 mark in Intel terms. But I didn't mean you to compare anything, I just said CPU in general the cores are hot even if you watercool it, they gotta sit at 30ish and say if you got a 2008 setup, say 2008 because if you're young you might not know. People used to do iirc some freon system and even there it went from -0 to full load 20c.

@Zach_01

I was just chit chating, I have all pleasure to talk about it as long as is logical! Thank you for joining our conversation.

 

[AusWolf]

Is a powerful processor regardless, is not a athlon or intel core duo, is somewhere around a i5 mark in Intel terms. But I didn't mean you to compare anything, I just said CPU in general the cores are hot even if you watercool it, they gotta sit at 30ish and say if you got a 2008 setup, say 2008 because if you're young you might not know. People used to do iirc some freon system and even there it went from -0 to full load 20c.

I've been building PCs since the early 2000s so I know what you're talking about. But trust me, modern, chiplet-based AMD CPUs are nothing like anything we've seen before in terms of thermal behaviour. Just read the reviews, OK?

 

[shadowvault]

AMD always was the "ocultish" side of hardware culture. I don't need to read them, I don't as you might know "shadowvault", I think of self-love 24/7, hope as we all do.

 

[Toothless]

Here is what I got if your interested. I decided to give Ryzen Master a try to do the curve optimization automatically and did a quick run with CPU-z.
I'm not entirely convinced the curves are stable yet but it's enough to get through some experimenting.

Managed to hit All Core 4.674GHz at about 70c with CPU-z.

View attachment 263827

Cinebench 23 is a bit different and hits at a lower all core clock when it's running.

View attachment 263830
View attachment 263832

You have given me hope on overclocking again. I thank thee.

 

[AusWolf]

AMD always was the "ocultish" side of hardware culture. I don't need to read them, I don't as you might know "shadowvault", I think of self-love 24/7, hope as we all do.

If you can't be bothered to check out a single review to see what the discussion is actually about, then why join the discussion at all?

"Educating" people on the old A10-7950K without first educating yourself about the topic at hand serves no purpose. Just saying.