Knowing the peak power dissipation at max turbo frequency wouldn't be quite as informative as you might think. The problem is power consumption varies widely depending on the workload. Presumably Intel would be expected to report power for the worst-case workload, but that would be significantly higher than what most people will ever see. This will only get worse once mainstream CPUs get 512 bit AVX units, further increasing the spread between integer only and AVX heavy workloads.
I think the basic problem is that power management on modern CPUs has gotten complex enough that you can't really summarize a CPU's power consumption in one or two numbers anymore. To really get a clear picture takes a full blown datasheet with tables showing power consumption under a range of conditions. Frankly, most enthusiasts aren't equipped evaluate such detailed information. Instead people rely on a few sparse data points provided by tech review sites and generally just throw a really big cooler on their CPU and hope for the best.
I would also be interested in manufacturers diclosing their idle consumption, because that's another area where one can save a lot of energy given the amount of time most desktop and servers spend in that state.
The Problem is that both the cooling and the power supply both work with one simple number, and we have to know what the CPU needs in order to put it in a system.
It's easy to make an Intel CPU stay within the limits of a particular power supply and cooling system by setting PL1 and PL2 appropriately. It's a Heisenberg's uncertainty sort of situation. You can fix frequency and get variable power dissipation, or fix power and get variable frequency, but you can't really fix both at the same time.
Enthusiast users building desktop systems tend to massively over-provision both power supply and cooling anyways so it's usually not an issue.
Motherboard manufacturers prefer the 'Unlimited PL2' route, because it puts their results at the top of benchmark lists.
...and so does every desktop user, no? If the cooling system can "take the heat", then the CPU should run as fast as it can. It seems like the whole point of this premature throttling is only to meet some stupid marketing spec, and the motherboard manufacturers are wise enough to ignore it.
I think the issue is that you might equate the TDP with power consumption and think that a particular chip is more efficient than another. Or that a smaller PSU would be sufficient. I would have expected a motherboard set to "auto" would use the Intel spec to hit the advertised TDP, not max everything out and end up at 3x that until it had to thermally throttle.
I think both Intel and the motherboard manufacturers are to blame: Intel for using a marketing number that won't be hit under most configurations, and the motherboard manufacturers for having over-agressive and unexpected defaults.
> If the cooling system can "take the heat", then the CPU should run as fast as it can.
The issue is what happens to benchmarks, because removing the power limit makes the limit thermal, which makes performance proportional to the efficacy of the cooling solution and other variables like the ambient temperature in the room.
There will also be variation between individual processors of the same model, since it's always been the case that some would run warmer or cooler, but now that affects out-of-the-box performance. (And with manufacturers sending chips to reviewers for testing, which ones are they likely to choose?)
This is a pretty big deal if you're a business wanting to buy the kind of small form factor desktops with cooling solutions that hew close to the official TDP numbers but you're looking at benchmarks done in larger cases with gamer-typical cooling solutions. Or people buying always-on home devices with more stringent cooling or efficiency requirements. Really anyone who wants to know how the processor would perform in an environment that it isn't permitted to continuously dissipate 180W.
The issue is what happens to benchmarks, because removing the power limit makes the limit thermal, which makes performance proportional to the efficacy of the cooling solution and other variables like the ambient temperature in the room.
There surely is an upper limit? Would cooling the CPU with liquid nitrogen, for example, make it perform appreciably faster even at the same stock clock?
There would be an upper limit, but where is it? Maybe something less exotic than liquid nitrogen (like water cooling), though who knows without trying it. Maybe even a solid air cooler could do it, or at least get within a few percent of it.
The concern is that a lot of the cooling solutions that are common in the market could be substantially worse than that, or even motherboards that can't supply that much power and are correspondingly configured not to.
It implies that a small form factor machine may be a lot slower even if it has the same processor in it.
The desktop user may be less happy if the lifetime of their processor is 1 year instead of >10. Thats the other downside of running hot: shorter time to failure.
In all these cases the CPU isn't going over its designed temperature limit, so unless Intel are messing with those numbers too now (AFAIK the expected lifetime is based on continuous running at the maximum temperature, like many other electronic components are rated at) the lifetime won't be shorter than the spec.
(I'm curious if these new CPUs based on smaller processes do have a measurably shorter lifetime in practice; a lot of people can probably tell stories of decade-old Prescotts and such which are still in active use.)
Longevity is, supposedly, also dependant on current, and hence voltage. Unfortunately processor reviews don't consider the lifetime of the product, as that is probably only known to the manufacturer
TDP="Thermal design power"... (the amount of designed cooling power to get max cpu performance). ̶W̶h̶e̶r̶e̶ ̶i̶s̶ ̶t̶h̶e̶ ̶m̶y̶s̶t̶e̶r̶y̶?̶
Mystery might be also in the nuance of turbo vs base frequency draw. So even if you design for TDP as the "max thermal dissipation" you might get only base frequency performance.
On a side note people often confuse TDP with the max power draw of the processor. Intel does not specify total power draw probably because processor scaling is such an important consideration. If you have a system, the easiest way to figure out what max draw from the power source is to consult the power brick.
> Mystery might be also in the nuance of turbo vs base frequency draw. So even if you design for TDP as the "max thermal dissipation" you might get only base frequency performance.
That is really what is happening. If you have '4 core 3Ghz Base 4.5Ghz turbo 95W TDP' what that means is that the CPU will pull, by spec, a maximum of 95W with all cores running at the base frequency under full load. Turbo can exceed this power spec and maximum power draw under turbo is ill-defined.
> On a side note people often confuse TDP with the max power draw of the processor.
It's the same thing, any 'power' a CPU 'draws' is converted to heat, by definition.
I'm having a hard time reconciling the Anandtech definition of Tau:
> Tau is a timing variable. It dictates how long a processor should stay in PL2 mode before hitting a PL1 mode.
with the Intel one:
> Turbo Time Parameter (Tau): An averaging constant used for PL1 exponentional weighted moving average (EWMA) power calculation.
I'm assuming that this power calculation is calculated this way, in a similar way to how Unix load averages are. But the thing is that they are an exponentially weighted moving average. I understand why: moving averages need the storage of all the values over that time period and the EWMA only needs the last value of the moving average.
It's not exactly clear how tau relates to the constant used for the exponentially weighted moving average. Assuming I want to minimize the difference between the EWMA and the moving average over a time interval, there appear to be many valid ways to do so -- and this appears to depend on the statistical properties of the variable being averaged. But the Anandtech article shows it as if it's a constant time, which is subtly different.
IIRC until the recent fix they pushed in a macOS update for the new i9 models, MacBook Pros don't have any of their power limits set (or have them on the defaults at 100W). The processor runs at full turbo, until the CPU gets to 102 C and throttles itself to keep it at about 99C.
TDP means what it always has. It's the maximal thermal power output of the processor. Input electrical power may fluctuate, but TDP was never an electrical parameter.
On modern Intel processors, TDP isn't the maximum thermal output power of the processor. It's the thermal output power when it's throttling back its speed by a particular, arbitrarily-defined amount for thermal reasons. It's basically a marketing number. The electrical input power on the other hand is pretty close to the actual thermal output power and is roughly the number you should actually use when speccing out a cooling system for an Intel processor.
The difference between input power and thermal output of a CPU is negligible. A tiny amount of electricity is used for external signalling, the rest gets converted to heat. So in practice TDP can be treated as a limit on input power. In fact that's exactly what Intel CPUs do by setting PL1 equal to TDP by default.
Negligible over time. What has everyone inexplicably confused is that electrical draw is allowed to be higher instantaneously than the output cooling will allow to accommodate bursts of computation. And some how everyone thinks that's a bad thing.
Again, TDP means the same thing now that it did 15 years ago.
What you've described is how it used to be, maybe you gets few seconds of usage over the TDP, but then the processor throttles back to hit the spec. That can still happen (and often does in laptops etc), but as described in the article, and many others, many to most motherboards sold as individual components are configuring the chips so that they will run at a higher power state for an indefinite time, so long as the temperatures stay within operating limits.
It would be nice to know what the expected power input / heat output is in that state, so you can get cooling to match, and have it in turbo mode at all times when the CPU is being stressed.
Cooling to the TDP only will (or should) get you the base clocks and a brief amount of turbo, but is likely to thermal throttle quickly.
This also brings more into question the adequeacy or not of the thermal interface material used between the CPU die and the heatspreader -- if the CPU will run turbo for longer if you keep it cooler, there's more frustration about less heat efficient solutions.
Shouldn't chip makers provide you a Max power usage number so you would know what power supply you need, if the CPU can draw double for an interval, then the GPU would do the same and your power supply does not enough extra power you maybe in trouble.
The "scandal" was caused by the fact reviewers were getting different results and people would accuse people on not doing the benchmark right.
It's how much it's designed to dissipate right? Which depends on the cooling so by your meaning it's a vague number isn't it?
I think from power/watt number on a cpu people expect two thing. How efficient it is for the performance (electric bill) where you need input power. How better a cooling I need will need the thermal power.
But even with that the issue of Turbo makes it harder to quantify those two because it all depends on workload. If I recall correctly though even with all that if I understand correctly the peak power draw in intel gets throttled much easier than AMD whereas their TDP number (compared to AMD) artificially inflates their efficiency reputation. I do think they had been much more efficient than AMD specially pre Zen.
In that case, reviewers and tech sites should either prioritize peak power or at least mention peak power numbers every single time TDP is mentioned, too, because peak power is at least as important as TDP. I know from past experience with laptops that tend to get hot, that if the laptop uses significantly more power than what you'd expect from the TDP, you will have a bad experience as a customer with that laptop. So at this point, TDP is definitely quite misleading to customers. Either scrap it completely or at least mention a more relevant metric to consumers.
> that if the laptop uses significantly more power than what you'd expect from the TDP
You're falling into the "TDP is power input" trap.. TDP and power input are not the same. You can have a chip that takes less power to operate at a TDP equivalent to another chip which takes more power to operate.
Edit: There is no way that 100% of output power in a CPU is thermal, else it would be nothing but a very expensive heater. Also, CPUs do not operate in a closed environment, like those replying to me here seem to think.
Input power equals output power is a fundamental physical rule, not a "trap". Yes, average (thermal) output power does not equal instantaneous (electrical) input power (ie thermal mass), but it averages out.
What the article is pointing out that TDP is thermal design power. Meaning a nameplate value that was targeted for "standard" operation, even though the processor can do higher. So if something has a TDP of 100W, yet the heatsink can maintain proper operating temperature while dissipating 150W, then the processor might sit there drawing (and dissipating) 150W the whole time.
As an aside, I really wish CPU cooler manufacturer/reviews would start simply stating the thermal resistance - basically the single number that characterizes its effectiveness. I know it would undermine the dog and pony show of building a test system with so-and-so's motherboard etc running benchmarks, but the entire goal of science is to weed out the extraneous narrative.
CPUs do no thermodynamic work. They do computational work, which does not count as work in work-heat-energy balance. All electrical energy consumed by CPU turns into increased internal energy of the surroundings, which is, in short, called heat.
Not all of the electrical energy consumed by the CPU is used by the CPU for computational work. Some of the electrical energy transfers out of the CPU to other components. The TDP of the CPU is not concerned with those other components. TDP is not a useful measurement of the amount of electrical energy a CPU can consume.
All consumed energy is used to produce heat, by definition of "consumption". If some electrical energy transfers out of CPU to other components (this is indeed possible), then by definition this energy is not consumed by the CPU. The meaning of TDP is explained in the article - Intel's value of TDP was such a measure, but is no longer so.
All output power is thermal though. This is basic thermodynamics. A negligible amount of input power will be distributed outside of the CPU die which will be dissipated as heat a short distance away. Pray tell, what other energy form do you think this input is becoming anyway?
Indeed there will be some piezoelectric effect and some EM emissions on the order of micro watts compared to the 10s of watts burned by the CPU. But the poster I was replying to seemed quite confused about the fate of power in a CPU used for computation: It’s dissipated as heat.
Well some power is used to run the fans which ends up partly turning into kinetic energy of air leaving the case. Sure, that may also manifest itself as heat somewhere down the line, but not in a way that is relevant for this discussion (and indeed you have to draw the line somewhere).
It's probably fairly neglible though, probables less than a watt in most systems.
I think the basic problem is that power management on modern CPUs has gotten complex enough that you can't really summarize a CPU's power consumption in one or two numbers anymore. To really get a clear picture takes a full blown datasheet with tables showing power consumption under a range of conditions. Frankly, most enthusiasts aren't equipped evaluate such detailed information. Instead people rely on a few sparse data points provided by tech review sites and generally just throw a really big cooler on their CPU and hope for the best.