It's not for stability reasons.Just saw this:
Intel Core Ultra CPUs May Feature Lower Clock Speeds to Avoid Stability Issues
Arrow Lake-S, Intel's upcoming generation of "Core Ultra" desktop CPUs, may serve as a new challenge for Intel's marketing department, according to new rumors shared on China's Weibo forums recently that suggest the new flagship, the Core Ultra 9 285K, will launch with lower clock speeds to...www.thefpsreview.com
well, stability is in the title of the article. And lowing clock speeds.It's not for stability reasons.
Delidded ? Jow is that any kind of 'stock' situation ?Yeah, looks like properly cooled delidded 14900KS CPUs may remain the fastest Intel CPUs for a while, except most reviewers may not choose to use them that way or even choose them at all. Gonna be tough for them if the performance gap between ARL and Zen 5 exceeds 10%.
Believe it or not, that IS stock for the 14900KS. It is simply impossible to air cool or watercool it properly without delidding. It's absolute trash without delidding.Delidded ? Jow is that any kind of 'stock' situation ?
Perhaps from their perspective it makes sense? Most of your sales are prebuilts where this isn't an issue for the vast majority of units. The platform is dead and getting massively outsold by the competition in retail already. Why waste time and energy on it when the new hotness will be available soon.Still a mess and apparently Intel is unperturbed by the situation.
Good question 3Dcenter, that's what I want to know.- Auto profiles still covered by Intel's warranty?
No, those useless silicon junk pieces are sold with a lid. Stock is with a lid. Whether they're stable or not is different problem.Believe it or not, that IS stock for the 14900KS. It is simply impossible to air cool or watercool it properly without delidding. It's absolute trash without delidding.
Only someone ignorant would use that with the lid on. At stock, it will keep throttling and never realize its full potential, unless you are fine with running only games and no serious multicore workloads but then it will perform almost similar to 14900K so what's the point of spending extra money?Stock is with a lid.
That's an AiO. Pretty sure it isn't just the lid that's the problem there. Full custom water could probably handle one of those things without a delidding.Start reading that article by searching the string "It was in our first real performance test".
And a chiller.Full custom water could probably handle one of those things without a delidding.
Given that the delid only got them ~18C off the recorded max temps, I'd say using an aggressive pump config with a large rad (560mm or larger) would do the trick, even without a chiller. AiOs don't have the best blocks and definitely don't have the best pumps. Granted a 360mm rad should be able to handle that kind of a heat load, so maybe even a 560mm rad would be overkill if the only thing being cooled is the CPU, but that's not gonna be the only thing.And a chiller.
What about the Subzero EVO here?AiOs don't have the best blocks and definitely don't have the best pumps.
Intel's proposed cooling solution:And a chiller.
You really think so?Going to be hilarious to watch all of this become a non factor in the next 6 months when suddenly Intel CPUs draws less power than their AMD counterparts.
Turin power consumption increases by +25% and the top SKU eclipses 500W. It’s certainly going to be more costly and harder to cool but with the performance increases you can sort of excuse the power consumptionYou really think so?
I still think the Ampere to Ada was the funniest efficiency doesn't matter switch.
And I'm sure Markfw will be on the same train because I don't see that changing with Turin. He'll be free to complain about Intel server parts being inefficient for years.
Going to be hilarious to watch all of this become a non factor in the next 6 months when suddenly Intel CPUs draws less power than their AMD counterparts.
You think they're going over 230W? For the X3D parts too?The desktop SKUs are what interests me though since those have even larger increases in power consumption proportionally. The people that X670E motherboards with overbuilt VRM are going to be glad that they did.
No, 232W is a hard limit.You think they're going over 230W? For the X3D parts too?
It’s in shipping manifests and in some of the open source code that’s received Zen 5 patches iirc. You don’t have to believe me, the 8C SKU appears have a 170W TDP / 232W PPT.And in my pulled from my bum opinion this post will not age well.