Question Good or bad to update mobo bios for stable 13700?

[StupidComputers]

13700k on msi 690, Nh-D15, 32gb 6400/C32, 3080 ftw3, 850w ti

I forgot all about the intel 13/14 gen issues, as mine doesn't crash. I just came across it again (too busy for too long, but that's done tg).

I see everyone screaming to update your bios right away! Few go on to mention that you need to select the profile for Intel default voltages etc after that.

So here's my take: Why do I want these idiots dictating my voltages again - especially as I've read that after this "fix" the chips run much hotter and throttle.

When I decided/just gave up and bought this, I knew it was halfway to an igniter. I had no intention of letting do what it wanted. When I went to poke around whilst learning to undervolt it, I saw stupidly high voltages - just like I've seen with other Intel builds for over a decade. But I also found MSI's Lite Load tool, aaaand no need to learn too much about undervolting. Picked almost or the lowest stable mode, got within 500pts of r23 cpu default score at 10C cooler, no throttling, called it job done.

So I have never crashed how they say, (or at all since initial setup), zero instability, I'm permanently and ideally undervolted. I can't see any reason to change the bios or my voltage profile. Maybe the microcode without their new voltages prevents problems without causing new ones? I have been absent in focus for a while, so I thought I'd check with those in the know.

To install new bios and/or profile or just stay with what seems to be pretty damn solid? Thanks.

 

[ex_bubblehead]

Unless the notes accompanying the BIOS update specifically state that it fixes a problem that you have been encountering there is never a reason to update BIOS "just because".

 

[thestryker]

Since you note that you're undervolting it shouldn't matter at all. The microcode changes are to address stock VID requests and cap voltage so as long as you've manually tuned your system it's fine. Should you stop doing this for any reason then you'd want to look into updating the BIOS.

 

[StupidComputers]

Since you note that you're undervolting it shouldn't matter at all. The microcode changes are to address stock VID requests and cap voltage so as long as you've manually tuned your system it's fine. Should you stop doing this for any reason then you'd want to look into updating the BIOS.

Thanks, that was my guess. I should point out that I didn't manually undervolt it. MSI has "Lite Load" profiles which have their set of researched value combinations for undervolting, so there may be some voltages left unaddressed, I'm not super sure - I really should check.

My assumption is no instability no problem. And for whatever reason, their profile seems to be solid. Sigh, I'll need to take a look at what's being pulled.

 

[StupidComputers]

Unless the notes accompanying the BIOS update specifically state that it fixes a problem that you have been encountering there is never a reason to update BIOS "just because".

And I never have lol, well maybe before I learned that especially with BIOS, if it ain't broke don't fix it. The problem here is that it might be broke and getting more broke in secret.

Stupid intel and asus, they've been overvolting their defaults for over 10 years - that I'm aware of, maybe longer, probably longer.

 

[thestryker]

Thanks, that was my guess. I should point out that I didn't manually undervolt it. MSI has "Lite Load" profiles which have their set of researched value combinations for undervolting, so there may be some voltages left unaddressed, I'm not super sure - I really should check.

My assumption is no instability no problem. And for whatever reason, their profile seems to be solid. Sigh, I'll need to take a look at what's being pulled.

Yeah definitely investigate to make sure all is good. Given the nature of the problem it's not something you really want to take chances on as the effects could appear down the road.

 

[StupidComputers]

Yeah definitely investigate to make sure all is good. Given the nature of the problem it's not something you really want to take chances on as the effects could appear down the road.

It looks like besides some odd voltage readings (VIN7) which should be just nonsense, the CPU vid seems to be pulling around 1.28v under load, with spikes of 1.33v according to the max reading - but I'm not really seeing it in the graph.
This is a bit higher than I usually like, but I'm guessing shouldn't be too bad.

What kind of voltages are damaging these chips, do you know? I can't seem to find that info, all the search engines give me are articles which are written for people who don't know anything - which is odd because they know they're talking to system builders who will have to swap their own chips. Maybe they guess everyone is just playing Legos ?

*Nevermind, I'm dumb, clearly the writers are typical modern mainstream reporters; understand little, investigate nothing, regurgitate press releases.

 

[geofelt]

Intel has found and fixed the root cause of 13/14th gen issues.
Here is their official update as of 9/25/2024:
https://community.intel.com/t5/Blog...sktop-Instability-Root-Cause/post/1633446#M40

The warranty on these products has been extended by 2 years.

A user only needs to verify that their motherboard includes the fixed bios.

My advice is to update all 13/14th gen motherboard bios updates to currency.

 

[Moonstick2]

Unless the notes accompanying the BIOS update specifically state that it fixes a problem that you have been encountering there is never a reason to update BIOS "just because".

While that's generally sound advice, the issue with the 13/14th gen is that the microcode errors mean that the Vmin Shift problems can start to appear in time, and if they do start appearing then the CPU is already damaged.

Intel recommend the update "as a preventative mitigation for processors not experiencing instability symptoms."

 

[Moonstick2]

To install new bios and/or profile or just stay with what seems to be pretty damn solid? Thanks.

How long have you had the 13700K? The earliest ones will be coming to the end of the extended warranty period anyway.

If you've had it and run it as it is now for a good long while, you're probably alright to leave it well enough alone. But nobody can guarantee it.

 

[StupidComputers]

How long have you had the 13700K? The earliest ones will be coming to the end of the extended warranty period anyway.

If you've had it and run it as it is now for a good long while, you're probably alright to leave it well enough alone. But nobody can guarantee it.

Is it bad that I can't remember if it's been one year or two? But they haven't been out long enough to reach the end of normal warranty period, much less the extended..

 

[StupidComputers]

While that's generally sound advice, the issue with the 13/14th gen is that the microcode errors mean that the Vmin Shift problems can start to appear in time, and if they do start appearing then the CPU is already damaged.

Intel recommend the update "as a preventative mitigation for processors not experiencing instability symptoms."

That's the debate. I have literally zero instability. The chip is undervolted since day 1, no more throttling, r23 barely hits 90C. Not to mention that their "fix" causes temperature, throttling and performance issues.

I'm thinking I'll keep trusting myself over sketchy Intel, and if the processor shows any damage in the next 3 years, I'll just go through their pita rma process which I'm sadly already familiar with. Hell, I don't want just another of the same chip anyway - especially a refurb!

 

[StupidComputers]

Intel has found and fixed the root cause of 13/14th gen issues.
Here is their official update as of 9/25/2024:
https://community.intel.com/t5/Blog...sktop-Instability-Root-Cause/post/1633446#M40

The warranty on these products has been extended by 2 years.

A user only needs to verify that their motherboard includes the fixed bios.

My advice is to update all 13/14th gen motherboard bios updates to currency.

Did you read through the comments in your link? Those relating their experience who are actually knowledgeable say it's not really a fix.

And that the Intel diagnostic tool is total garbage, which I found out myself in 2014. We're talking Samsung style results. I had a high voltage chip then that reported as fail, during the rma Intel updated their diagnostic tool and then the faulty chip came back as pass on the updated version. I've actually NEVER seen a chip fail since that update. Luckily they had already accepted the RMA from the original diagnostic's fail.

So I ask you, on what basis are you recommending the bios update?

 

[geofelt]

Who do you trust?
The engineers who designed and built the processors, or laymen(including me) with opinions?

 

[thestryker]

It looks like besides some odd voltage readings (VIN7) which should be just nonsense, the CPU vid seems to be pulling around 1.28v under load, with spikes of 1.33v according to the max reading - but I'm not really seeing it in the graph.
This is a bit higher than I usually like, but I'm guessing shouldn't be too bad.

What kind of voltages are damaging these chips, do you know? I can't seem to find that info, all the search engines give me are articles which are written for people who don't know anything - which is odd because they know they're talking to system builders who will have to swap their own chips. Maybe they guess everyone is just playing Legos ?

*Nevermind, I'm dumb, clearly the writers are typical modern mainstream reporters; understand little, investigate nothing, regurgitate press releases.

I believe Intel capped the voltage at 1.55v.

You're not going to see the high voltages on full load. It typically appears on lightly threaded loads which push the cores it does use to the max. This is one of the reasons minecraft servers were fairly reliably. killing these parts.

 

[StupidComputers]

Who do you trust?
The engineers who designed and built the processors, or laymen(including me) with opinions?

Honestly, neither.

The engineers missed the huge problem, can't fix it properly, and most importantly they don't contradict the company. So you're not even trusting the engineers, you're trusting what the corporation is telling/selling you, which is always a terrible idea. They should have done a lot more than just a software patch and extended warranty, but profit margins are their eternal tunnel vision focus.

Frankly, from their perspective the extended warranty is genius. I definitely do not intend to buy Intel next time, for more reasons than just this, but if they replace the chip down the line with a newer socket model, then we'll have to buy new mobos, and if they make that socket viable for a while, then many people will buy another Intel chip to save on mobo, maybe RAM too.

 

[Lafong]

Who do you trust?
The engineers who designed and built the processors, or laymen(including me) with opinions?

Unfortunately, that question immediately leads to a can of worms.

Was it non-engineers who ultimately approved the release of the original design? Or the fix for that matter? Being on the outside, we can only speculate who has the final say. One guy? A vote among 23? 23, of which 13 are in marketing and PR?

Any decision takes a leap of faith, but preconceived notions will likely prevail in the process.

Considering how annoying the RMA process is, if I had a 13700 running OK as of today, I'd probably leave it alone with crossed fingers.

 

[StupidComputers]

I believe Intel capped the voltage at 1.55v.

You're not going to see the high voltages on full load. It typically appears on lightly threaded loads which push the cores it does use to the max. This is one of the reasons minecraft servers were fairly reliably. killing these parts.

Zoiks, I was just reading about the idle/low use voltages being the issue. I don't play Minecraft, any other ways to test? I play a crysis wars mod from many years ago, I'll check that next time.

I also read turning off the E cores might be wise, and I feel like I always meant to do that, but don't remember why.

When you say Intel capped voltage to 1.55v, is that before or after the fix? Cuz that's pretty high.

 

[StupidComputers]

Unfortunately, that question immediately leads to a can of worms.

Was it non-engineers who ultimately approved the release of the original design? Or the fix for that matter? Being on the outside, we can only speculate who has the final say. One guy? A vote among 23? 23, of which 13 are in marketing and PR?

Any decision takes a leap of faith, but preconceived notions will likely prevail in the process.

Considering how annoying the RMA process is, if I had a 13700 running OK as of today, I'd probably leave it alone with crossed fingers.

Exactly. The engineers are certainly not in charge.

 

[thestryker]

When you say Intel capped voltage to 1.55v, is that before or after the fix? Cuz that's pretty high.

After, and you can just run Cinebench single threaded.

I think this was a pretty good piece of analysis on it:

View: https://youtu.be/1bEv74JrHQo?si=JaWse1jmvfZjIMOW

 

[Moonstick2]

Is it bad that I can't remember if it's been one year or two? But they haven't been out long enough to reach the end of normal warranty period, much less the extended..

Ah, my mistake, I misread the two-year extension as a two-year warranty.

...if the processor shows any damage in the next 3 years, I'll just go through their pita rma process which I'm sadly already familiar with.

The only possible caveat there is whether in a year or two they might claim that the warranty doesn't cover chips that have been left to run under the old BIOS instead of the Intel recommended update that was designed to prevent the damage that's since occurred. Probably unlikely, but the number of times a large rich corporation treats a single customer unfairly and obnoxiously in an attempt to save a couple of hundred dollars never ceases to amaze.

 

[StupidComputers]

After, and you can just run Cinebench single threaded.

I think this was a pretty good piece of analysis on it:

View: https://youtu.be/1bEv74JrHQo?si=JaWse1jmvfZjIMOW

Hmm. Idk about Jay, but that's a positive for him. Most of these guys are bought *cough*Linus.. i don't trust anything anymore, one bitten twice shy, 10 times bitten..

It's odd that msi would pump this out so fast when they are the only ones who made undervolting easy, ergo the fewest affected customers. Curious..

Anyway, this guy is still seeing over 1.5v after update. I've done heavy crisis singlecore now issue, and even maxing out the new Indiana Jones game doesn't hit 1.35v, and I'm still getting more than 10% better than the frames others with similar or slightly better rigs are getting.
Seems like a foolish idea to fix what clearly ain't broke.

 

[StupidComputers]

Unless the notes accompanying the BIOS update specifically state that it fixes a problem that you have been encountering there is never a reason to update BIOS "just because".

So this is the really the shortest version of the correct answer, but on mobile or desktop through different browsers, there's no way to select the answer or even edit the original post, I can only edit comments after.. am I missing something? It never used to be this way.

 

[USAFRet]

So this is the really the shortest version of the correct answer, but on mobile or desktop through different browsers, there's no way to select the answer or even edit the original post, I can only edit comments after.. am I missing something? It never used to be this way.

To the right of the relevant comment, there is a checkmark.
Click that.

 

[ex_bubblehead]

So this is the really the shortest version of the correct answer, but on mobile or desktop through different browsers, there's no way to select the answer or even edit the original post, I can only edit comments after.. am I missing something? It never used to be this way.

The default thread type is "Discussion" which does not offer a "Best Answer". When starting a new thread it is on you to pay attention and choose the correct thread type. The moderators and administrators are able to edit the thread type after the fact, but you have to make that request to us. The change of thread type in here has been applied retroactively.