News U.S. gov't mulls taking a stake in Intel to speed up Ohio campus fab build out

[spongiemaster]

No it really wasn't. It was all the reviewers forcing mobos to use unlimited power while also not using this additional power to make the cpu clock higher.
And they still do it today, even after all of this time of intel telling us to use stock/default settings reviews show intel CPUs with 300W+ calling it out of the box.

The inferior node absolutely put Intel at a disadvantage. Let's not deny reality. Unlimited power settings didn't help, but Alder Lake/Raptor Lake needed more than the stock settings to compete in the benchmark wars. Intel needed significantly higher clock speeds on an inferior node to be performance competitive and that meant higher power consumption.

 

[TerryLaze]

The inferior node absolutely put Intel at a disadvantage. Let's not deny reality. Unlimited power settings didn't help, but Alder Lake/Raptor Lake needed more than the stock settings to compete in the benchmark wars. Intel needed significantly higher clock speeds on an inferior node to be performance competitive and that meant higher power consumption.

But they didn't get significantly higher clocks in the reviews, that's the whole point. Lifting the power limits doesn't increase the clocks at all it only forces more power usage.

 

[thestryker]

No it really wasn't. It was all the reviewers forcing mobos to use unlimited power while also not using this additional power to make the cpu clock higher.

This might be true of some reviewers, but it's still true Intel is using a lot more power for the same performance:

https://www.techpowerup.com/review/intel-core-i9-14900k/22.html

Peak power consumption isn't significantly worse (RPL v Zen 4), but the in between definitely is. It's one of the big issues with the underlying ADL/RPL architectures. This was fixed with the release of ARL and while it has higher peak power consumption than Zen 5 it is similar or better everywhere else.

The clock speeds on RPL are much higher than AMD or even ARL run. Yes it was part of the design of both the node and CPU, but it harms efficiency potential just the same.

 

[TerryLaze]

https://www.techpowerup.com/review/intel-core-i9-14900k/22.html

Peak power consumption isn't significantly worse (RPL v Zen 4), but the in between definitely is. It's one of the big issues with the underlying ADL/RPL architectures.

If you compare this to the 7950x the difference in average, not max, power draw is ~42W , not the 100-150+ watts difference most reviews are trying to tell you, the difference with the 9950x is ~45W.
It's in the link you posted just a little bit lower, power consumption (47 test average)

Point being, that little of a difference in power draw will not cause anybody to think that intels node is that much worse, especially since most of the CPU is made from e-cores which are not made for max power efficiency while running server loads, still the difference is small.

 

[thestryker]

If you compare this to the 7950x the difference in average, not max, power draw is ~42W , not the 100-150+ watts difference most reviews are trying to tell you, the difference with the 9950x is ~45W.

...

Point being, that little of a difference in power draw will not cause anybody to think that intels node is that much worse,

It's a third more power consumption so even if the number doesn't seem large that isn't a "little" difference. Intel 7 is certainly worse than N5 by a decent margin in terms of efficiency and nobody should really think otherwise.

 

[TerryLaze]

It's a third more power consumption so even if the number doesn't seem large that isn't a "little" difference. Intel 7 is certainly worse than N5 by a decent margin in terms of efficiency and nobody should really think otherwise.

And that's fine, 170 compared to 130 is fine, this does not warrant people running through forums shouting "oven" "nuclear reactor" and stuff like that.
350W compared to 130W on the other side does.