News New firmware addresses Intel CPU gaming instability — Asus Intel Baseline Profile fixes crashing on Raptor Lake and Raptor Lake Refresh chips

[Pierce2623]

Glad to see this was addressed quickly. Hopefully this will be the end of instablity issues for intel. I look forward to someone posting some test results on a previoisly unstable cpu (looking at you Jarred...hope this helps you).

It seems strange how far they’re having to pull back to find stability when my 13700k is perfectly happy at 5.6 all core. I guess maybe it’s because I’m using a sane voltage that is significantly less than stock settings would push at 5.6GHz.

 

[wingfinger]

i'll say it: every "default" bios should be the CPU makers default requirement settings.

They make the chip & know it best and if you want to mess with that make it the optional settings.

Stability > performance for 90% of PC users.

I have bought ASUS motherboards ever since those early-on intel ones, but one of the things that always irked me about ASUS was the idea that they were adding secret hardware related "optimizations" and teaks that were basically always there ... unless maybe you somehow knew how to counter them. They know better than intel?

I like running my hardware as stock as possible. I don't want things running stressed or hot. I don't want hitches due to time collisions, recoverable errors, retrain, throttle, or whatnot. I want my hardware running in-spec, cool and smooth. No issues.

Some people like overclocking and tweaking hardware. That is for people that want to take those risks. It is also for people who like running stability tests or don't care.

If I want to tweak something, I'll tweak software; I can always reinstall it fresh.

For that reason, I was wondering if the TUF line would be different.

 

[bit_user]

I like running my hardware as stock as possible. I don't want things running stressed or hot. I don't want hitches due to time collisions, recoverable errors, retrain, throttle, or whatnot. I want my hardware running in-spec, cool and smooth. No issues.
...
For that reason, I was wondering if the TUF line would be different.

I recently ordered this "workstation" board:

Pro WS W680-ACE｜Motherboards｜ASUS Global

Intel® W680 LGA 1700 ATX motherboard, DDR5, PCIe 5.0, 2.5 Gb, SlimSAS, enhanced power and cooling solutions

www.asus.com

Hasn't arrived, yet. According to the users manual, it seems to have many of the same optimizations available as gaming boards, though it seems like you can switch them all off.

 

[TerryLaze]

I guess the key question is whether the voltage/frequency curve is the same for all CPUs of a given model, or is there some calibration to exploit better quality chips? My guess is the former: that they determine a V/F curve that works for the lowest-quality CPUs in that "bin" and that's just what they all use.

My guess would be that the curve is there for software stability so chip quality doesn't matter, enough at least.
They are not going to go as far down as to be just on the cusp of crashing, absolutely everything has to be able to run stable and that with a safety margin.

That's also why intel states upper limits for power, which get ignored by everybody, up to that amount of power the voltage will regulate and be stable and safe for the CPU, always.

 

[thestryker]

I guess the key question is whether the voltage/frequency curve is the same for all CPUs of a given model, or is there some calibration to exploit better quality chips? My guess is the former: that they determine a V/F curve that works for the lowest-quality CPUs in that "bin" and that's just what they all use.

I too think the curve is likely SKU derived, but I don't remember seeing anyone say one way or the other.

Buildzoid mentioned in a comment on I think it was one of JayzTwoCent videos about how removing power limits causes Intel CPUs to go up the voltage curve so long as cooling is good whether it needs to or not. This likely partly explains the variance you see in unlimited power performance testing where the power consumption doesn't necessarily match performance.

 

[Sleepy_Hollowed]

This is pretty wild.

I've not used their stuff, but I remember some wild default settings back in the day, I'm amazed they've kept with this for Intel of all people. I've not had any issues with most Ryzen vendors, but one would think they'd stick to the design defaults as their defaults, what a wild thing for ASUS to do.

I think AMD vendors learned with how hot their stuff was in the Piledriver era.

 

[NinoPino]

So, yet again Asus thinks they know better than AMD and Intel; causing mayhem and destruction along they way.

Nice...

For now it is limited to Intel's motherboards.
And it is not only ASUS, all motherboards makers.

 

[TerryLaze]

I think AMD vendors learned with how hot their stuff was in the Piledriver era.

No they didn't, because default mobo settings was what caused the ryzen 7000 series to blow up until AMD forced agesa updates.

 

[SyCoREAPER]

I knew /posted this was a BIOS issue, sure enough here we are. Also surprise, ASUS is the main culprit once again. So glad I didn't go ASUS, they've been screwing motherboards up for the past 2 or 3 years.

 

[NinoPino]

I knew /posted this was a BIOS issue, sure enough here we are. Also surprise, ASUS is the main culprit once again. So glad I didn't go ASUS, they've been screwing motherboards up for the past 2 or 3 years.

This is a problem also with other brands, ASUS seems to be the first brand to officially fix it.
Also, this is not a real "BIOS issue" in the sense that it was intended to work this way.

 

[SyCoREAPER]

This is a problem also with other brands, ASUS seems to be the first brand to officially fix it.
Also, this is not a real "BIOS issue" in the sense that it was intended to work this way.

It probably is an issue with other boards rkk, I'm just saying to add this to the existing list of ASUS issues.

As for whether it's a BIOS fault, that's subjective. OEMs like pushing the limits even if it's just a little outside of spec. One does it, they all do it to keep up.

The 13 and 14 i9s are such furnaces that even a small miscalculation in PD, timing, etc could send the CPU into panic. And the cooling I see on a lot of PCs is inadequate IMO for a i9 so it's not that surprising that many are having issues but many also aren't.

And this is 100% my opinion and bias, a i9 is overkill for gaming. The small performance uplift isn't worth the extra heat. That doesn't excuse the situation but I figured I'd throw that out there while i9 is a topic of discussion.

 

[TerryLaze]

The 13 and 14 i9s are such furnaces that even a small miscalculation in PD, timing, etc could send the CPU into panic. And the cooling I see on a lot of PCs is inadequate IMO for a i9 so it's not that surprising that many are having issues but many also aren't.

The problem is the exact opposite, when the CPU gets too hot it just throttles down and you wouldn't get any crashes.

The problem is with people that do have not just enough cooling but such an overkill of cooling that the CPU can boost higher than the CPU can deal with, instead of throttling under the pressure of shader compiling and being ok, it keeps boosting because the cooling is there.

 

[SyCoREAPER]

The problem is the exact opposite, when the CPU gets too hot it just throttles down and you wouldn't get any crashes.

The problem is with people that do have not just enough cooling but such an overkill of cooling that the CPU can boost higher than the CPU can deal with, instead of throttling under the pressure of shader compiling and being ok, it keeps boosting because the cooling is there.

My observation still stands but just not correctly. That said, thats where the BIOS should be stopping it. IIRC (and I very well might not be) many BIOS are created by one company and customized for the components and interface by the OEMs. That may have been a dream though.

Regardless, should the CPU have hardcoded logic to stop boosting? Absolutely.
Should board makers have had BIOS safeguard in the default profiles?
Absolutely.

 

[TerryLaze]

Regardless, should the CPU have hardcoded logic to stop boosting?

They have, that's what makes them locked...
An unlocked CPU should be unlocked.

 

[Amdlova]

What CPU makers need It's say on the box what voltage cpu use...
I don't know what asrock put on the cpu... I see a perfect 35w cpu going UP 1.235mv and its fine
Only option I have on bios is adjust the power level from 1 - 5 put it on nivel 5 less voltage possible :S

 

[TheHerald]

I don't understand this. If undervolting is so great and so safe, then why doesn't it do that out of the box?

Personally, I just want something stable. However, I also don't like wasting power. So, I wish the CPU would just undervolt itself if/when it's safe to do that and then the advice would always be: "don't touch it, unless you like playing with fire."

Because intel (and amd of course) want to make sure that your cpu will be stable under all conditions up to 100c for at least as long as your warranty. So the extra voltage it ships with helps with stability at very high temperatures and after years of usage where electromigration will have caused some degration. On a high end build with a good cooler and temperatures in check voltage doesn't need to be that high

 

[vMax]

If you leave the power management settings at their defaults in BIOS and Windows and don't overclock, then all CPUs already undervolt themselves to the lowest programmed voltage and frequency they can run at (voltage and frequency stepping is a thing). Just open CPU-Z and you will see them undervolt to 0.8V when idle.

Yep, the CPU under balanced power settings in windows will automatically downclock and drop vcore but if you leave the BIOS settings at Auto or even worse 'multi-core enhancement' especially ASUS! as many of the Z series motherboards like to do, under load it will still pump way too much voltage and even when just running low level tasks I have seen vcore jump past 1.4v in HarwareInfo.

Now I understand that those that buy the K series CPU's are specifically marketed for enthusiasts and it is great when coupled to a Z series motherboard that you have access to really push the CPU and fine tune so many settings to get the very best out of the CPU and as importantly out of the RAM, but I think the baseline motherboard settings should always be Intel stock settings. I see too many people online who buy these CPU's and see very high temps and cannot understand why, when in reality, the temps are actually no problems at all, it is just the stock or enhanced settings are pushing too much vcore. Enthusiast and those that are into overclocking and undervolting know this but the average consumer most certainly does not.

AMD CPU's also run hot and are suppossed to as they have been 'engineered to run at 90 to 95 Degrees C'! ( AMD 50th level Jedi mind trick) boosting based on thermal headroom but even then I would find that to high, though I have also seen vcore at 1.4v+ which I personally would find way too much. 1.3v for longevities sake is my limit.....but saying that, in the good old days with my Sandy Bridge 2600K pushed to 4.5GHz for many years, it never seemed to degrade!

 

[SyCoREAPER]

They have, that's what makes them locked...
An unlocked CPU should be unlocked.

Just because it's unlocked doesn't means it SHOULD go freeroam, just that it can.

This applies to the CPU itself and the BIOS. Both should be locked as if it were bonestock unless a specific instruction is sent to supply more power.

I know what you're saying but it's a sloppy implementation on both sides.

As I mentioned a few posts back that my MB also did some goofy settings for my 13700k. It was running stupid hot and unstable because the default BIOS settings were like a n00b intern set them up.

This is/was all easily preventable but it's another case of basically a pissing contest to show on paper that one processor is better than the other rather than focus on a positive experience for the end user.

 

[TerryLaze]

Just because it's unlocked doesn't means it SHOULD go freeroam, just that it can.

This applies to the CPU itself and the BIOS. Both should be locked as if it were bonestock unless a specific instruction is sent to supply more power.

I know what you're saying but it's a sloppy implementation on both sides.

As I mentioned a few posts back that my MB also did some goofy settings for my 13700k. It was running stupid hot and unstable because the default BIOS settings were like a n00b intern set them up.

This is/was all easily preventable but it's another case of basically a pissing contest to show on paper that one processor is better than the other rather than focus on a positive experience for the end user.

Oh I completely agree, anytime that warranty voiding settings get applied there should be a warning were you have to type in a whole word to show that you understood and agree.

 

[CmdrShepard]

Oh I completely agree, anytime that warranty voiding settings get applied there should be a warning were you have to type in a whole word to show that you understood and agree.

If you want a babysitter in your BIOS I don't and I sure as hell don't want to have click and type all those Accept and Agree crap. That's just evil anti-user behavior because it puts the manufacturers' bottom line ahead of user experience. Say no to corporate bootlicking.

Instead, how about people who are clueless about low-level hardware settings don't buy top-end unlocked parts and don't pretend they know how to build a PC and set it up properly to manufacturers' specs?

If you can't handle BIOS settings then buy a brand name prebuilt PC or ask a techie friend to help you build one, and you won't have these problems. That way we won't have to lock every setting behind three layers of "Are you reaaally, reaaaally sure?!?" prompts for the rest of us who know what they are doing.

 

[TerryLaze]

If you want a babysitter in your BIOS I don't and I sure as hell don't want to have click and type all those Accept and Agree crap. That's just evil anti-user behavior because it puts the manufacturers' bottom line ahead of user experience. Say no to corporate bootlicking.

So....telling the user what's going on and asking them for permission is anti-user...and evil on top of that...

If you can't handle BIOS settings then buy a brand name prebuilt PC or ask a techie friend to help you build one, and you won't have these problems. That way we won't have to lock every setting behind three layers of "Are you reaaally, reaaaally sure?!?" prompts for the rest of us who know what they are doing.

Did you even read any part of this thread?!
The whole point is that pre-builds use unstable settings, since the mobos they use use unstable settings, and a user that is being protected by your idea of pro-user would have no idea what's going on.

 

[CmdrShepard]

So....telling the user what's going on and asking them for permission is anti-user...and evil on top of that...

The kind of user that needs to be told "what's going on" and "asked for permission" shouldn't even be there in the first place. Yes, it's anti-user and evil to those users who know that stuff.

The whole point is that pre-builds use unstable settings, since the mobos they use use unstable settings, and a user that is being protected by your idea of pro-user would have no idea what's going on.

What I am saying is that users who don't know those things shouldn't be buying enthusiast grade parts to begin with instead of all of us who do know being saddled with more obstacles just to make it safe for ignorant people.

There are plenty of locked CPUs and non-O/C oriented mainboards with simpler (non Z-series) chipsets to choose from which don't behave like that. The problem is as usual with all those new generations of people who want instant gratification in whatever they do.

They want to jump into a multiplayer game which takes 150 hours to master and start at the highest possible difficulty level others on the team be damned.

They want to build a high-end PC without having any background in electrical engineering, hobby electronics, or even basic high school physics. CPU specifications and BIOS manuals? Who reads that? Reading is for nerds.

And because it is not politically correct to put them in their place by telling them how much they suck and send them off until they learn everyone else is now paying the price of their entitlement and "freedoms".

I agree that no values should be set out of the manufacturer's specification, but what some of you seem to be missing is that both RAPL and current limit values aren't out of spec -- they simply mean "unlocked" or "no limit" on K CPU parts. Could the mainboard manufacturers do a better job and fine-tune for a perfect balance of power consumption and performance at default settings? Certainly, but both the boards and the CPUs are marketed towards people who want to tweak stuff -- if you tweak it for them what's the point?

TL;DR -- if you want safe defaults either don't buy unlocked parts or learn how to configure them, but don't put more warnings and Agree buttons or worse yet verification boxes in my face.

 

[TerryLaze]

The kind of user that needs to be told "what's going on" and "asked for permission" shouldn't even be there in the first place. Yes, it's anti-user and evil to those users who know that stuff.

What I am saying is that users who don't know those things shouldn't be buying enthusiast grade parts to begin with instead of all of us who do know being saddled with more obstacles just to make it safe for ignorant people.

There are plenty of locked CPUs and non-O/C oriented mainboards with simpler (non Z-series) chipsets to choose from which don't behave like that. The problem is as usual with all those new generations of people who want instant gratification in whatever they do.

They want to jump into a multiplayer game which takes 150 hours to master and start at the highest possible difficulty level others on the team be damned.

They want to build a high-end PC without having any background in electrical engineering, hobby electronics, or even basic high school physics. CPU specifications and BIOS manuals? Who reads that? Reading is for nerds.

And because it is not politically correct to put them in their place by telling them how much they suck and send them off until they learn everyone else is now paying the price of their entitlement and "freedoms".

I agree that no values should be set out of the manufacturer's specification, but what some of you seem to be missing is that both RAPL and current limit values aren't out of spec -- they simply mean "unlocked" or "no limit" on K CPU parts. Could the mainboard manufacturers do a better job and fine-tune for a perfect balance of power consumption and performance at default settings? Certainly, but both the boards and the CPUs are marketed towards people who want to tweak stuff -- if you tweak it for them what's the point?

TL;DR -- if you want safe defaults either don't buy unlocked parts or learn how to configure them, but don't put more warnings and Agree buttons or worse yet verification boxes in my face.

You are talking about something completely different from everybody else.

You are talking about somebody who knows manually changing settings versus everybody else talking about mobo makers using whatever settings they want without informing you.

 

[SyCoREAPER]

The kind of user that needs to be told "what's going on" and "asked for permission" shouldn't even be there in the first place. Yes, it's anti-user and evil to those users who know that stuff.

What I am saying is that users who don't know those things shouldn't be buying enthusiast grade parts to begin with instead of all of us who do know being saddled with more obstacles just to make it safe for ignorant people.

There are plenty of locked CPUs and non-O/C oriented mainboards with simpler (non Z-series) chipsets to choose from which don't behave like that. The problem is as usual with all those new generations of people who want instant gratification in whatever they do.

They want to jump into a multiplayer game which takes 150 hours to master and start at the highest possible difficulty level others on the team be damned.

They want to build a high-end PC without having any background in electrical engineering, hobby electronics, or even basic high school physics. CPU specifications and BIOS manuals? Who reads that? Reading is for nerds.

And because it is not politically correct to put them in their place by telling them how much they suck and send them off until they learn everyone else is now paying the price of their entitlement and "freedoms".

I agree that no values should be set out of the manufacturer's specification, but what some of you seem to be missing is that both RAPL and current limit values aren't out of spec -- they simply mean "unlocked" or "no limit" on K CPU parts. Could the mainboard manufacturers do a better job and fine-tune for a perfect balance of power consumption and performance at default settings? Certainly, but both the boards and the CPUs are marketed towards people who want to tweak stuff -- if you tweak it for them what's the point?

TL;DR -- if you want safe defaults either don't buy unlocked parts or learn how to configure them, but don't put more warnings and Agree buttons or worse yet verification boxes in my face.

Glad you've always been so perfect and never had to learn anything. That horse you're on will need a step latter.

Look everyone! The second coming of Christ is upon us!

Stop being so critical and derogatory of less experienced users. A one time accepting the risks going to give you schlong cancer.

Some people have faith that the tech companies, you know, the ones that get paid millions of dollars? Will actually configure and design things to work safely out of the box and let us configure if from there. You have some skewed view of what unlocked means. Unlocked means safe factory settings Aka a baseline, to tweak off of. Not figuring out where stable is (going backwards and then tweaking. Those kind of views are also why we have stupid motherboards that don't let you do anything unless you shell out $400+ dollars to get what used to be standard features.

This is why there is a constant Intel and AMD fangirl pissing match. One screws up and the other team loses their farmers market worth of bananas and pounces.

TLDR: Take a breath. Hitting accept to modifying non-typical BIOS settings once isn't going to hurt you. If you're good you might even get a lollipop.

 

[CmdrShepard]

You are talking about something completely different from everybody else.

Yes, because everyone else is ignoring it and it should be addressed.

You are talking about somebody who knows manually changing settings versus everybody else talking about mobo makers using whatever settings they want without informing you.

Yes, those mainboards and K CPUs are marketed towards people who know how to manually change settings.

But please, do enlighten us, what would the mobo makers have to do for you to consider yourself informed?

Because, as far as I know (and I know a lot on the subject), the first time you insert a new CPU in the socket and boot the system the first message you see is along the lines of "It looks like you have just installed a new CPU, press F1 to enter Setup to configure it".

And when you do it's all there black on white -- it just needs someone who knows how to read.

Also, for at least a decade or so, the BIOS-es have safe defaults they use to boot your CPU and run BIOS itself, and your O/C settings kick in only when the OS starts booting.

So, until you enter setup and either confirm or change default settings mobo makers don't use "whatever settings they want without informing you" so please stop spreading FUD.