Intel Posts 10th Gen Core Power Limit and Tau Values

[btarunr]

Intel today updated the public data-sheet of its 10th Gen Core "Comet Lake-S" desktop processor to reveal precise power limit and tau values of each specific SKU. PL 1 or power level 1 is interchangeable with the processor's TDP as a power value. PL 1 is sufficient for a processor to sustain its base frequency (nominal clocks). For example, a processor with 65 W TDP has PL 1 at 65 W. PL 2 is what affords the processor the power to seek out boost frequencies. This value varies with between model to model, with the unlocked K/KF SKUs getting higher PL 2 values than the locked ones. The company also disclosed Tau. This is a timing variable that tells the processor how long (in seconds) can it stay within PL 2, before having to retreat to PL 1.



View at TechPowerUp Main Site

 

[cucker tarlson]

not great,not terrible.

 

[Deleted member 172152]

I can't buy an Intel processor anymore. The power bill alone puts me off since my pc is used by multiple people during most of the day.

Add to that a new motherboard, even if I had an Intel motherboard to start with, and the cost of the processor itself and for me it is an AMD clean sweep from budget to high end.

 

[Vayra86]

Intel technology 2020: add more useless stats to push cpus even further out of their comfort zone. This is their disclaimer, nothing else. And even then, those power figures are retarded.

65W > 224W. Lol, imagine if you bought that 70W TDP cooler for it. After all, that is not a K- CPU... 28 seconds...

not great,not terrible.

Pretty terrible when only your i3s have some semblance of logic to their turbo left, pushing 90W on it vs 65W base. Those are quads. Intel still rocking that 2016 mojo at heart. We gained a few hundred mhz since Sandy Bridge, yay. 8 gens

 

[cucker tarlson]

Intel technology 2020: add more useless stats to push cpus even further out of their comfort zone. This is their disclaimer, nothing else. And even then, those power figures are retarded.

65W > 224W. Lol, imagine if you bought that 70W TDP cooler for it. After all, that is not a K- CPU... 28 seconds...



Pretty terrible when only your i3s have some semblance of logic to their turbo left, pushing 90W on it vs 65W base. Those are quads. Intel still rocking that 2016 mojo at heart. We gained a few hundred mhz since Sandy Bridge, yay. 8 gens

10 core on 14nm at +5GHz ? seems standard to me.

 

[Vayra86]

10 core on 14nm at +5GHz ? seems standard to me.

PL2 is effectively saying 'you can't keep me in turbo like you used to'. Its a boundary that is just there because Intel needs to push that turbo into the red - because it NEEDS to top that bench chart. Any sensibility is out the window.

 

[cucker tarlson]

doesn't manual oc override those anyway ?

 

[Vayra86]

doesn't manual oc override those anyway ?

If you buy a K CPU. If you don't, joke's on you. Now, take a long look at that 10900 again. 65W. 224W.

 

[cucker tarlson]

I think xtu can still do that on any cpu.and bios limits too.it is very confusing tho.the hell is going on with that 10700

 

[Vya Domus]

250W for 10 cores is staggering, AMD can fit 64 cores in that power budget. With lower clocks yes, but still, that's 6.4x more cores.

 

[cucker tarlson]

250W for 10 cores is staggering, AMD can fit 64 cores in that power budget. With lower clocks yes, but still, that's 6.4x more cores.

apples to oranges.
yes,zen is mighty more efficient,but still stupid comparison.
compare intel's 56 vs amd's 64,that's more consistent as far as comparisons go.
still beats the crap out of intel,but not 6.5x,nowhere near.225 at 3.4 vs 400 at 3.8. 2x maybe ?

epyc 32c 7542 225w at 3.4G,xeon plat 9222 32c 250w at 3.7G

no one is saying intel is as efficient,but be objective in comparisons.

 

[W1zzard]

doesn't manual oc override those anyway ?

You can even override those on non-K CPUs, on any motherboard/chipset

 

[Caring1]

PL2 is effectively saying 'you can't keep me in turbo like you used to'. Its a boundary that is just there because Intel needs to push that turbo into the red - because it NEEDS to top that bench chart. Any sensibility is out the window.

Isn't that 225W TVB not Turbo which can be sustained?

 

[Vayra86]

You can even override those on non-K CPUs, on any motherboard/chipset

That is good, then.

 

[ratirt]

250W?
Holly crap :/

 

[Darmok N Jalad]

I can't buy an Intel processor anymore. The power bill alone puts me off since my pc is used by multiple people during most of the day.

Add to that a new motherboard, even if I had an Intel motherboard to start with, and the cost of the processor itself and for me it is an AMD clean sweep from budget to high end.

The only way you’d know for sure is to measure your consumption while in use. Desktop tasks are not going to be crazy like this, so you’re only paying the premium under high-demand tasks. To me, the bigger issue is the higher max-TDP just makes everything more expensive—motherboards and coolers have to be beefier to get the advertised performance.

 

[1d10t]

Now that something to brag about, 10s - 15s longer than AMD

The only way you’d know for sure is to measure your consumption while in use. Desktop tasks are not going to be crazy like this, so you’re only paying the premium under high-demand tasks. To me, the bigger issue is the higher max-TDP just makes everything more expensive—motherboards and coolers have to be beefier to get the advertised performance.

In previous chipset, MCE is enable by default on Z motherboard. Now even with B motherboard, "MCE" thingy like assorted bclk boost also enable by default, kinda bummed unsuspecting user isn't ?

 

[thesmokingman]

apples to oranges.
yes,zen is mighty more efficient,but still stupid comparison.
compare intel's 56 vs amd's 64,that's more consistent as far as comparisons go.
still beats the crap out of intel,but not 6.5x,nowhere near.225 at 3.4 vs 400 at 3.8. 2x maybe ?

epyc 32c 7542 225w at 3.4G,xeon plat 9222 32c 250w at 3.7G

no one is saying intel is as efficient,but be objective in comparisons.

Yea, stop looking at benchmarks and metrics, dammit!

 

[cucker tarlson]

Yea, stop looking at benchmarks and metrics, dammit!

reported for lq as it's clearly not what I meant.
you can't take a 5.3GHz desktop part and compare against 3.4GHz epyc to come up with a crazy 6.4x number.

 

[thesmokingman]

reported for lq as it's clearly not what I meant.
you can't take a 5.3GHz desktop part and compare against 3.4GHz epyc to come up with a crazy 6.4x number.

3990x is 280w too.

 

[cucker tarlson]

then do take time to compare it against an intel high core count one instead of spamming this thread.

 

[thesmokingman]

then do take time to compare it against an intel high core count one instead of spamming this thread.

What is your problem. The post that made you go into fandom simply showed the stark difference in what each company can do with the wattage used. Clearly one is stretching that power a helluva lot further than the other. It's your own bias that you cannot admit the obvious.

 

[cucker tarlson]

What is your problem. The post that made you go into fandom simply showed the stark difference in what each company can do with the wattage used. Clearly one is stretching that power a helluva lot further than the other. It's your own bias that you cannot admit the obvious.

what is yours ?
that other guy took the highest clocked desktop intel vs lowest clocked epyc,and you're sweating that this is somewhat fair.

 

[thesmokingman]

what is yours ?
that other guy took the highest clocked desktop intel vs lowest clocked epyc,and you're sweating that this is somewhat fair.

You didn't like the epyc example so why not the worlds fastest cpu? The only one sweating is you since you hit the report button. I suggest you calm down.

 

[cucker tarlson]

You didn't like the epyc example so why not the worlds fastest cpu? The only one sweating is you since you hit the report button. I suggest you calm down.

no,I liked the epyc example.just against epyc's counterpart.
not a +5GHz gaming cpu.
keep trolling,both of you,that'll be constructive for the thread.