News Intel 35W Comet Lake CPU Spotted With Up To 123W Peak Power Draw

[Admin]

According to a SiSoft test entry, the Core i9-10900T pulls up to 123W at peak operating clocks.

Intel 35W Comet Lake CPU Spotted With Up To 123W Peak Power Draw : Read more

 

[TJ Hooker]

Given how many motherboards seem to more or less remove the power (PL2) and time limits for turbo boost, it's probably not unusual for a -T CPU to perform similarly to a non-T version unless you make sure the BIOS is configured to use Intel's recommended values.

 

[2Be_or_Not2Be]

Intel should only be able to list the TDP of their processors at the highest power draw during actual operation, whether that's single-core/all-core turbo. People are really misled when they see a CPU listed at 35W when, like this situation, they hit close to 125W in actual operation.

AMD is much better at this (CPU power draw stays pretty close to rated TDP, but can get 10-15% higher), but all CPU mfgs should be required to list their CPUs at max power draw. (/rant)

 

[travsb1984]

So it only draws 35 watts when it's basically at idle... Go to know. It's a essentially a 110 TDP chip that can throttle down if cooling is insufficient.

 

[joeblowsmynose]

Intel should only be able to list the TDP of their processors at the highest power draw during actual operation, whether that's single-core/all-core turbo. People are really misled when they see a CPU listed at 35W when, like this situation, they hit close to 125W in actual operation.

AMD is much better at this (CPU power draw stays pretty close to rated TDP, but can get 10-15% higher), but all CPU mfgs should be required to list their CPUs at max power draw. (/rant)

AMD tdp numbers since Zen2 also are pretty sketchy, just from the formula they use (expected cooling power somehow gets factored into the formula, which it shouldn't be)

Intel actually has two TDP numbers for each of its chips and both are accurate in their own way.

PL1 TDP 35w (no boost - base clock)
PL2 TDP 105w (full max load - boosted) (I didn't see the document - this is just an guessed at example for reference)

And this is quite accurate, and aligns quite well with power consumption even.
The public never sees the second number though ... herein lies the problems with Intel's TDP numbers ...

The PL2TDP for 9900k is 250w - almost the exact power consumption of the CPU under heavy AVX load with all cores boosting.

I used to heavily mock Intel for their TDPs but now I think if we should move to a standard for both mfgrs, to use Intel's formula, except a requirement to list TDPs for both average working state and full load state (for both mfgrs), as opposed to just the former. Anyway a standard would a welcome here.

 

[truerock]

Intel should only be able to list the TDP of their processors at the highest power draw during actual operation, whether that's single-core/all-core turbo. People are really misled when they see a CPU listed at 35W when, like this situation, they hit close to 125W in actual operation.

AMD is much better at this (CPU power draw stays pretty close to rated TDP, but can get 10-15% higher), but all CPU mfgs should be required to list their CPUs at max power draw. (/rant)

I disagree. The problem isn't what a CPU manufacturer reports as TDP. The problem is that CPU technology currently supports variable power-draw for each core and the number of cores in operation can vary. The problem is that any one TDP number is not going to give a complete picture.

So I guess it would be great if there was an industry standard way to provide CPU power specs in a way that provided a reasonable idea of what might be expected in various scenarios - but, there isn't. So, yes... I guess I would want to know what the power draw is if I'm staring at a Wikipedia web page and what the absolute maximum is.

We are starting to see more often CPU specs with several to many TDP numbers.