Thursday, September 12th 2024

Intel Z890 Chipset Motherboards to Launch with Default Power Profile Out of the Box?

Intel is expected to launch its next-generation Core Ultra 200 "Arrow Lake-S" desktop processors on October 24, 2024, and since these chips introduce the new Socket LGA1851, they launch alongside new motherboards. The first wave of Core Ultra 200 series processors will be "K" or "KF" SKUs targeting gamers and enthusiasts, and the first compatible motherboards will be based on the premium Intel Z890 chipset. Intel is reportedly being extra careful not to repeat the "Raptor Lake" fiasco that saw motherboards power its 13th- and 14th Gen Core "Raptor Lake" processors with elevated voltages, causing their irreversible physical degradation over time. To this effect, Intel is reportedly getting its motherboard partners to ship their Intel Z890 chipset motherboards with Intel Default Power Profile out of the box.

A motherboard-level Power Profile dictates the processor base power (PL1), maximum turbo power (PL2), and IccMax values. We don't know these values for "Arrow Lake-S," particularly its top Core Ultra 9 285K part; but to illustrate what Default Power Profile out of the box means, we have to look at "Raptor Lake-S." For a Core i9-14900K, the Performance Power Profile—which is what Intel Z690 and Z790 motherboards enable out of the box—provide a PL1 of 125 W, PL2 of 253 W, and IccMax of 307 W, which is what Intel considers stock for this processor model.
The Extreme Power Profile, which can be enabled in the UEFI setup program, holds PL1 and PL2 at 250 W, and unlocks 400 W IccMax (something that requires two 8-pin EPS power connectors to be plugged in, each capable of 225 W). The Insane Power Profile, available on select motherboard models, completely unlocks PL1 and PL2 to "4096 W" (i.e. anything that the motherboard VRM is capable of providing), and IccMax to 511 W (the socket's design limit). The Intel Baseline Power Profile (aka Default Power Profile) runs contrary to these three.

On the Core i9-14900K, the Default Power Profile gives the processor a PL1 of 125 W (which is also the chip's base power value), but the PL2 maximum turbo power is limited to 188 W, and the IccMax is capped at 249 W. These severely reduce the processor's boost frequency residency even at stock settings.

If the rumors of Intel getting motherboard vendors to ship Z890 chipset motherboards with Default Power Profile out of the box are true, then out of the box, these processors will likely not perform as Intel intended, and the company's guidelines to reviewers and end-users would be to manually enable at least Performance Power Profile in the UEFI setup, to at least get stock performance out of these processors.
Sources: Benchlife.info, VideoCardz, Hardware Busters
Add your own comment

16 Comments on Intel Z890 Chipset Motherboards to Launch with Default Power Profile Out of the Box?

#1
nguyen
Yeah sure but I'm just gonna put in even lower TDP and current limits, if I happen to buy Arrow Lake CPU :D
Posted on Reply
#2
dj-electric
Lord have mercy the next 6 weeks are going to be motherboard fiesta.
Posted on Reply
#3
Outback Bronze
What's wrong with the Insane Profile out the box? Did something happen?? :)
Posted on Reply
#4
nguyen
Outback BronzeWhat's wrong with the Insane Profile out the box? Did something happen?? :)
Probably something related to Intel not making enough CPU for RMA
Posted on Reply
#5
Waweq
Intel and their voltages :kookoo::kookoo::kookoo::kookoo::kookoo::kookoo::kookoo:
Posted on Reply
#6
las
Sounds like OC headroom to me.

Can't wait to see what Intel can do with TSMC 3N - AMD will probably be rushing out 9000X3D very fast afterwards... AMD going with 4N might bite them in the A
Posted on Reply
#7
Kirederf
If the rumors of Intel getting motherboard vendors to ship Z890 chipset motherboards with Default Power Profile out of the box are true, then out of the box, these processors will likely not perform as Intel intended, and the company's guidelines to reviewers and end-users would be to manually enable at least Performance Power Profile in the UEFI setup, to at least get stock performance out of these processors..
I don't agree with this. Stock performance would be with the 'Default Power Profile' since those are the 'stock' settings. You get extra performance by enabling the 'Performance Power Profile' but that's not 'stock'. Soon this will be enabled again by default, to get the extra performance. And we all know how this went.
Posted on Reply
#8
Event Horizon
Ship it in default mode but review it in performance mode. I guess that's kind of like the XMP/EXPO situation. Still not great.
Posted on Reply
#9
las
Event HorizonShip it in default mode but review it in performance mode. I guess that's kind of like the XMP/EXPO situation. Still not great.
Any decent review will state this clearly
Posted on Reply
#10
RogueSix
Reviews should definitely be conducted with the default profile only. Every decent CPU review usually has an OC section and that is where the testing of the 'Performance' and 'Extreme' profiles belongs.
Posted on Reply
#11
KaitouX
For reference on profiles, from Intel themselves:


According to this image, which is still referenced in the latest news about the Raptor Lake issue in the Intel forums, and as far as I know is exactly what still is in the datasheet, the baseline profile is not the default, and shouldn't be unless required. At least when talking about current CPUs Intel was also somewhat clear in saying "Intel is not recommending motherboard manufacturers to use ‘baseline’ power delivery settings on boards capable of higher values. Intel’s recommended ‘Intel Default Settings’ are a combination of thermal and power delivery features along with a selection of possible power delivery profiles based on motherboard capabilities."
My guess is that the default profile will simply be the Performance one, with the baseline and extreme profiles being there as alternatives, together with possible custom profiles by motherboard manufacturers. Which would be avoiding the situation of some motherboards defaulting to PL2=4096W and ICCMAX=500A like it was happening.
Posted on Reply
#12
Daven
What a novel idea, run default settings as the default.
Posted on Reply
#13
Solid State Brain
KaitouXFor reference on profiles, from Intel themselves:
PL1=PL2=253W cannot be sustained continuously below TjMax with air cooling unless you have some pretty serious air flow and a high-end cooler, so I can't see how it could be the default unless thermal throttling is the intended default operating state for these processors.
Posted on Reply
#14
KaitouX
Solid State BrainPL1=PL2=253W cannot be sustained continuously below TjMax with air cooling unless you have some pretty serious air flow and a high-end cooler, so I can't see how it could be the default unless thermal throttling is the intended default operating state for these processors.
I'm guessing the assumption is that if you're getting a 1X900KS you're going to be cooling it properly. For the other models 125W PL1 is considered the standard, but with the option for the manufacturers to set it to 253W. I mean depending on the motherboard, making the assumption the person will have enough cooling isn't that weird.
Even if in my opinion it should be lower, with the PL2 of the i7 range around 180W and the i9 around 200W, it doesn't really make sense in most cases to have the default be higher than that.
Posted on Reply
#15
Minus Infinity
WaweqIntel and their voltages :kookoo::kookoo::kookoo::kookoo::kookoo::kookoo::kookoo:
cRaptor Lake was a rush job. Insiders have admitted they did the refresh in record time to beat AMD in benchmark and it clearly shows. It's the ring bus that is experiencing the high voltage spikes. If we can believe Intel, Arrow Lake's new design means it won't happen again.

I'm prepared to give them a go. I want to move beyond 8 cores and don't want to deal with AMD's dual ccd design even more so that non-X3D models get core parking nonsense now. Intel 265KF is looking like a nice update from my 5800X.
Posted on Reply
#16
NoLoihi
Minus InfinityIt's the ring bus that is experiencing the high voltage spikes.
From what it seems like to me, it's a programming error. People have said these spikes happen right when coming out of C-states, so there's probably some control registers which aren't being set properly in time. Most importantly, I don't see how this would result in any appreciable performance gain.
I don't see why people constanly need to pull in all these allegations. Errors happen in products that haven't been rushed as well. Processor testing is (surely) highly automated, they might not have had a test case for that, maybe some engineer disabled it in a fat-fingering accident, who knows. They’ve had quite some time after release to find the issue, so even if you wish to imply the release was rushed beyond good conciousness, there were lengths of time after that to pick up any loose ends. Had they quietly released a microcode update with “fixes for voltage regulation in transitional operating states”, probably nobody whatsoever would have batted an eye.
It just doesn’t make sense to believe Intel sat on their knowledge for over a year, which is also why they themselves didn’t know what’s up for the first stretch of time.
Posted on Reply
Add your own comment
Dec 11th, 2024 20:31 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts