News Intel 12th Gen Alder Lake T-Series 35W CPUs Reportedly Hit 4.9 GHz

[Admin]

FanlessTech leaks the specifications for Intel's 12th Generation Alder Lake 35W processors.

Intel 12th Gen Alder Lake T-Series 35W CPUs Reportedly Hit 4.9 GHz : Read more

 

[BillyBuerger]

The fact that they have no Gracemount cores on the more normal mobile CPUs seems an odd choice. The point of these cores is to be more power efficient for light tasks. That seems like the kind of thing that would be great for mobile devices where only these cores need to be used for normal things which could save on battery life. This feels to me a little like back in the Pentium/Celeron M days where only the more expensive Pentium would clock down to save power. I get they want to have their features that their most expensive parts have. But power saving features should be the primary thing that every mobile CPU should have. Save the performance things for the high end chips. But then people would be able to get by with the cheaper parts as most people don't actually need the highest performance and that doesn't help their margins.

 

[abufrejoval]

PCIe 5.0 doesn't come free in terms of power budget. So I wonder just how much juice might be left for the CPU parts, once the iGPU takes its 15 Watts and PCIe takes a full bite, too.

And I guess we're not talking 35 Watts at 4.9GHz all cores sustained.

Given that my Tiger Lake i7-1165G7 also hits that clock on a single thread for a second or so on 64 Watts PL2, Alder Lake's maximum clock for 35 Watts PL1 setting isn't much of a surprise, wonder where their PL2 is for that chip.

I also wonder if the top clocks for the efficiency cores are actually lowered, too. First hunch would be no, or very little.

 

[abufrejoval]

The fact that they have no Gracemount cores on the more normal mobile CPUs seems an odd choice. The point of these cores is to be more power efficient for light tasks. That seems like the kind of thing that would be great for mobile devices where only these cores need to be used for normal things which could save on battery life. This feels to me a little like back in the Pentium/Celeron M days where only the more expensive Pentium would clock down to save power. I get they want to have their features that their most expensive parts have. But power saving features should be the primary thing that every mobile CPU should have. Save the performance things for the high end chips. But then people would be able to get by with the cheaper parts as most people don't actually need the highest performance and that doesn't help their margins.

Normally product differentiation for a single die design would be based on binning. And I agree, it's hard to imagine that there should be significant numbers of chips with defects or voltage issues in that Gracemont area unless yields are really bad overall.

I am more inclined to believe that Intel's product segmentation is more based on psychology than technology: they work very hard to make Gracemont cores a thing to lust for, even if it costs them very little transistor budget. And then they take exactly that hot new gadget away from you, the moment you try to skimp on price. To me there is really no two ways about it: They really want you to buy the most expensive product, with "all the magic".

But the reality is still the same as with all those old Pentiums that wouldn't ever Turbo up or down: much more energy is saved when you simply stop the clock and CPUs have done that as far back as the 80486SL (probably earlier).

I remain doubtful that the Gracemont cores will ever pay their markup in energy savings or productivity gains, but who in this readership can resist wanting to try for himself? Pencil pushers will show no mercy to Gracemont and they may just be right.

 

[TerryLaze]

The point of these cores is to be more power efficient for light tasks.

I doubt that because the gracemont cores would have to run these lighter tasks at much higher clocks, so they wouldn't be that much more efficient, or would be much slower at finishing them and I don't see either of the two scenarios being something intel is hot for.

I believe the gracemont cores are there to give them a pretty big boost in multithreading tasks without decreasing the single thread speed at all and without using too much more power.

 

[mdd1963]

So within the official '35 Watt TDP' envelope, 1 core will allegedly hit 4.9 GHz for brief bursts of up to 30 seconds or so before all clocks ultimately stabilize at something far less, likely 3.0 GHz or so... (If users and/or manufacturers allow within BIOS settings, undoubtedly it can clock at clock speeds closer to 4.5-4.7 GHz for longer at something much closer to 65-80 Watts; but as 35W grabs press attention, we'll ignore the fact that for that to happen at Intel specs, it's likely for a burst up to 10 seconds or so..)

 

[Roland Of Gilead]

PCIe 5.0 doesn't come free in terms of power budget. So I wonder just how much juice might be left for the CPU parts, once the iGPU takes its 15 Watts and PCIe takes a full bite, too.

And I guess we're not talking 35 Watts at 4.9GHz all cores sustained.

Given that my Tiger Lake i7-1165G7 also hits that clock on a single thread for a second or so on 64 Watts PL2, Alder Lake's maximum clock for 35 Watts PL1 setting isn't much of a surprise, wonder where their PL2 is for that chip.

I also wonder if the top clocks for the efficiency cores are actually lowered, too. First hunch would be no, or very little.

Agree with you there! TDP should be changed to PL2 instead of this misnomer on 35w CPU's. It's misleading for those not in the know.