Superior Stability by GIGABYTE BETA BIOS with Intel Baseline on Z790/B760 Motherboards

[chrcoluk]

For the testing on the temperature page, yes, so you can get a feel for the actual difference instead of comparing "throttled" and "throttled"

Oh if its only for that page, I think I may have misunderstood your previous comments, thanks for answering.

nah, it may be just a hype, why would anyone first ignore even the intel "Perf" spec to damage consumer CPUs and than offer the "Base" spec to care for the involved CPUs aferwards?
totally insane, right?
btw the picture is from 2021 ... but Intel let the board makers just ignore all specs ... and "tada" surprise it did hurt some CPUs

Here is a quote I found from their statement, seems you on the ball with what you found.

 

[Upgrayedd]

Just don't pick a CPU that needs 100 more watts to stay somewhat relevant.

Intel will come back, they have too much knowledge and money to not to, but right now for gaming, frankly a 7800X3D is the most-efficient and balanced choice

I think these talking points have become kind of pointless with how much the GPU pulls that typically gets paired with a CPU like the 7800X3D. Its very fast at gaming and efficient but the high end GPUs needed to balance that CPU really negate the whole point.

Its definitely not the most balanced. That'd be an Intel chip. Gaming, sure the 7800X3D is best. Gaming and anything else productivity oriented, Intel.

 

[GoldenX]

"Release it, just don't say we're sorry"

 

[Minus Infinity]

I think these talking points have become kind of pointless with how much the GPU pulls that typically gets paired with a CPU like the 7800X3D. Its very fast at gaming and efficient but the high end GPUs needed to balance that CPU really negate the whole point.

Its definitely not the most balanced. That'd be an Intel chip. Gaming, sure the 7800X3D is best. Gaming and anything else productivity oriented, Intel.

I'd probably just buy a 14700K if I were buying tomorrow and power limit it to say 150W max or undervolt it as I do a lot more than gaming. I know for some software I use like COMSOL Intel slaughters AMD. For overall productivity Intel wins IMO and I'm an AMD user and for gaming let's face it, 14700 is still excellent.

In reality I'm waiting for Zen 5 vs Arrow Lake before deciding on my next upgrade path. If leaks are correct Zen5 is already faster than Zen 4 X3D for gaming, so should be a killer all-round cpu.

 

[stimpy88]

I know I said that in the second paragraph.

I quoted the wrong post. Sorry.

 

[b1k3rdude]

Here is a quote I found from their statement, seems you on the ball with what you found.

So its exactly what most of us correctly indicated it was.

In reality I'm waiting for Zen 5 vs Arrow Lake before deciding on my next upgrade path. If leaks are correct Zen5 is already faster than Zen 4 X3D for gaming, so should be a killer all-round cpu.

Same, but Zen5/6 will need to be AM5 & WIn10 compataible. Its one of the main reasons I would like top continue with AMD over intel. Failing to do either will result in large drop in market share

 

[persizi]

It's a good thing that the CPUs doesn't popping like popcorn damaging the whole socket because of SOC overvoltage. Oh wait...

 

[john_]

reviewers need start putting faster ram in the Intel systems. Using the same 1:1 optimized for AMD kits in the Intel system and calling it good is rather unbalanced. I've built a i7 on 7600 timings at 8Ghz that was 100% stable. Seeing reviews use the same plain 6000MHz kit is kinda ignorant because they find the right kit for AMD but just ignore Intel's capabilities.

How much more the same capacity of memory at 8GHz costs compared to 6GHz DDR5?

Also how about testing AMD CPUs and Intel CPUs while using their stock coolers, or coolers that don't cost over $60, instead of pairing them with $200 cooling solutions? I mean, how many people throw a $200 cooling solution over an R9 7950X or an i9 14900K? Also it would have been nice to see tests where average motherboards where used, motherboards at the $200-$300 price range instead of motherboards that cost $500. Too see how these CPUs, both R9 7950X and i9 14900K behave when the VRMs are not of the best and most expensive kind.
Well? Wouldn't that be more accurate and fair?

 

[HD64G]

Here is the reason why Intel not only allowed but wanted reviews to be done without power limits:

An increase of more than 30% in wattage for a score increase of less than 10% just to surpass the already known and on sale 7950X. Intel very well knew what they needed to do in order to get some wins and sale those CPUs or else those would be DOA even for the die-hard Intel fans.

UPDATE: A few minutes after my initial post the following video dropped and Intel's mic also ; )

 

[john_]

It's a good thing that the CPUs doesn't popping like popcorn damaging the whole socket because of SOC overvoltage. Oh wait...

It's probably not such a good thing. If they where popping, this problem would have been seen press coverage 2 years ago and been fixed 2 years ago. People would have gotten new stuff, working properly stuff through RMA. Staying for so long under the carpet was probably the reason many Intel customers would have to face constant instability not knowing what is going on with their systems.

 

[persizi]

It's probably not such a good thing. If they where popping, this problem would have been seen press coverage 2 years ago and been fixed 2 years ago. People would have gotten new stuff, working properly stuff through RMA. Staying for so long under the carpet was probably the reason many Intel customers would have to face constant instability not knowing what is going on with their systems.

Of course. Dealing with RMA on both CPU and MB is way easier that change a couple of settings in the BIOS.

 

[john_]

Of course. Dealing with RMA on both CPU and MB is way easier that change a couple of settings in the BIOS.

You mean, having to deal with instability for weeks or even months, installing re installing windows, games, drivers, going mad not knowing what is going wrong with your system and then finding out after all these trouble that it was motherboard manufacturers pushing your CPU outside of it's limits with Intel looking the other way, it's easier than having to do an RMA, because now it's just a couple of settings in the BIOS to save your already degraded CPU?
OK. If you say so, probably you had never the fun of troubleshooting an unstable system that was crushing when running certain applications while looking stable under other stress testing applications.

 

[persizi]

You mean, having to deal with instability for weeks or even months, installing re installing windows, games, drivers, going mad not knowing what is going wrong with your system and then finding out after all these trouble that it was motherboard manufacturers pushing your CPU outside of it's limits with Intel looking the other way, it's easier than having to do an RMA, because now it's just a couple of settings in the BIOS to save your already degraded CPU?
OK. If you say so, probably you had never the fun of troubleshooting an unstable system that was crushing when running certain applications while looking stable under other stress testing applications.

Sorry that you had to to deal with instability for months, installing re installing windows, games, drivers etc.

 

[Crackong]

Hey we sucessfully implement the new 'One Click DC' feature.
Sorry for the stability issue , we pushed too far to beat AMD in the hall of fame .
We wanted good launch reviews so we pushed to far, and we won't change the performance claim
Now your CPU runs stable.

UPDATE: A few minutes after my initial post the following video dropped and Intel's mic also ; )

There is a fasinating point in Steve's video, that the "intel baseline" profile isn't the same across motherboard vendors.
It looks like the "intel baseline" profile isn't provided by Intel and the motherboard vendors have to bake their own "intel baseline" .

 

[nguyen]

I have been saying the same, running unlimited power has been the norm for Intel for the past decade.
The problem is that this practise has been relatively safe, until 13th/14th gen K SKU that is. Now Intel is putting the blame on mobo makers LOL

 

[john_]

Sorry that you had to to deal with instability for months, installing re installing windows, games, drivers etc.

Well, if you are trying to be sarcastic or ironic, then maybe you have no idea how it is of having stability issues that you can't explain.

 

[Upgrayedd]

How much more the same capacity of memory at 8GHz costs compared to 6GHz DDR5?

Also how about testing AMD CPUs and Intel CPUs while using their stock coolers, or coolers that don't cost over $60, instead of pairing them with $200 cooling solutions? I mean, how many people throw a $200 cooling solution over an R9 7950X or an i9 14900K? Also it would have been nice to see tests where average motherboards where used, motherboards at the $200-$300 price range instead of motherboards that cost $500. Too see how these CPUs, both R9 7950X and i9 14900K behave when the VRMs are not of the best and most expensive kind.
Well? Wouldn't that be more accurate and fair?

Im talking about reviews, not a direct comparison. In a comparison you could keep the ram the same and compare how each CPU fairs with that same ram. In a CPU review you want to see the full capabilities and potentials of the product. You don't want anything limiting its potential like a weak cooler or slower ram than it can handle. Swapping ram used to be okay in reviews before gears and infinity cache, but now it's not okay for reviews. Especially when the ram is only optimized for one brand. If they wanted to swap a basic 4800MHz, okay sure. Hand selecting an AMD tuned kit then just totally ignoring any Intel optimizations on the ram side is bad practice.

No, cheaping out on testing equipment would not be more accurate.
24GB sticks cost roughly the same at 6000MHz as they do at 8000MHz.

 

[chrcoluk]

Here is the reason why Intel not only allowed but wanted reviews to be done without power limits:
View attachment 345421View attachment 345422
An increase of more than 30% in wattage for a score increase of less than 10% just to surpass the already known and on sale 7950X. Intel very well knew what they needed to do in order to get some wins and sale those CPUs or else those would be DOA even for the die-hard Intel fans.

UPDATE: A few minutes after my initial post the following video dropped and Intel's mic also ; )

HUB jumping on the blame Intel for all of it bandwagon based on what media contacts have said, rather than what engineers have said, and then ignored that ASUS and gigabyte who now both have baseline settings dont actually match up when its enabled on both boards one has 253w at baseline and the other much lower, so one of them is doing it wrong, and on both boards the baseline is still not the default he had to manually enable it (intel asked them to put it as default). So I dont agree with his interpretation.

I have been saying the same, running unlimited power has been the norm for Intel for the past decade.
The problem is that this practise has been relatively safe, until 13th/14th gen K SKU that is. Now Intel is putting the blame on mobo makers LOL

Just because they got away with it for years it doesnt mean they innocent.

The blame is on both sides really, Intel should have had a firmer grip for sure. But we cant say Asus and co are innocent when they using blatant silly defaults.

We can even see on these new bios, gigabyte's baseline settings are different to Asus, So another vendor mistake.

Looking at what @_Flare posted it looks like Gigabyte got it right and Asus got it wrong.

 

[Crackong]

We can even see on these new bios, gigabyte's baseline settings are different to Asus, So another vendor mistake.

Looking at what @_Flare posted it looks like Gigabyte got it right and Asus got it wrong.

You mean Power Limit = 188W is the 'right' baseline setting?
Then why Intel themselves still uses 253W as performance index ? oh and also in a Asus Motherboard.

 

[Broken Processor]

Looking at what @_Flare posted it looks like Gigabyte got it right and Asus got it wrong.

And I suspect this is why Intel hasn't been forcing default power profiles. Who would pay so much for that performance compared to the competition even the toughest fanboy would have to take a large gulp of something strong before entering their card details. I suspect if indeed Gigabyte got it right a lot of people will be very unhappy. In the grand scheme of things the performance difference is nothing but perception is everything.

 

[john_]

Im talking about reviews, not a direct comparison. In a comparison you could keep the ram the same and compare how each CPU fairs with that same ram. In a CPU review you want to see the full capabilities and potentials of the product. You don't want anything limiting its potential like a weak cooler or slower ram than it can handle. Swapping ram used to be okay in reviews before gears and infinity cache, but now it's not okay for reviews. Especially when the ram is only optimized for one brand. If they wanted to swap a basic 4800MHz, okay sure. Hand selecting an AMD tuned kit then just totally ignoring any Intel optimizations on the ram side is bad practice.

No, cheaping out on testing equipment would not be more accurate.
24GB sticks cost roughly the same at 6000MHz as they do at 8000MHz.

I am also talking about reviews. With the latest fiasco with Intel CPUs becoming unstable and degrading, it's obvious that testing a CPU under perfect conditions is misleading. Using the best mobo available, the best cooling system available, the best and fastest RAM on the platform and probably the fastest SSD available today, combined with the fastest Video card, while showing the performance a CPU can reach without limitations, is absolutely misleading. Reviews should be done with a good midrange, let's say $300 motherboard, a very good air cooler at $60-$80, a typical NVMe SSD at 3-7GB/sec and the best VFM RAM. If that's 6000MHz, then 6000MHz it is. Then reviews can add a page or two, about the "Perfect system" and show to their readers what the CPU can do under perfect conditions.
Reviews today are totally misleading.

And why use 8000MHz RAM when there are even faster RAM out there?

PS AMD recommends 6000MHz RAM, Intel CPUs support up to 5600MHz. Is there an Intel recommendation about needing 8000MHz, 9000MHz, 10000000000MHz ram to achieve scores not limited from RAM speeds?

I would like the link of those 24GB 8000MHz sticks that cost the same as 6000MHz.
Then I want a link with the cheapest 8000MHz ram compared to the cheapest 6000MHz ram. And this time not limit ourselves to 24GBs sticks that is a very specific product case and could not be saying the absolute truth about prices.

 

[OkieDan]

All my CPU reviews have been done at Intel stock settings, for the last 15 years or so.

Each Intel review has a second full run "Power Limits Removed", which is probably what you expected the default to be

What about motherboard reviews, are Intel / AMD specs used for those reviews?

 

[john_]

HUB jumping on the blame Intel for all of it bandwagon based on what media contacts have said,

HUB while rushing to point at the problem, in my opinion they are rushing to save face. They had reviews in the past of Intel motherboards not utilizing Intel CPUs at their maximum throwing the blame on motherboards and their componets. They even mentioning it in their video, at the beggining. But had they ever done a video blaming Intel and accusing them of scamming consumers? I doubt. Instead they are doing a number of videos where they directly blame AMD of scumming consumers.

IN MY OPINION, HUB are hypocrites who rushed to do this video for their own benefit.

Here is the title of that video that shows motherboards not utilising Intel CPUs at their full potencial when using stock power limits.

And this is their take on that. It's a mess, but "let's unlimit them and see that it's fine!"

Now, what title should we put in a motherboard review about AMD motherboards that we don't like? Well, it's AMD platform, so let's be more aggressive, let's make AMD platform look like "crap" when going for a cheaper board

So, Intel messed up with power limits? Let's be ungry about it. What our viewers are probably feeling. And you know, Intel "screwing up" can be just a mistake by them. They should just be more carefull in the future, right?

How about AMD? Well, AMD..... Let's just call them scammers. But, let's also put a questionmark in the end just to be safe. Right?

HUB, IN MY OPINION always, is an indication of the problems of tech press.
Waiting for GN video to throw more laught with tech press being "ungry" at Intel because they "didn't knew" either.

 

[chrcoluk]

You mean Power Limit = 188W is the 'right' baseline setting?
Then why Intel themselves still uses 253W as performance index ? oh and also in a Asus Motherboard.


View attachment 345526

Yes 188w appears to be the right setting. PL4 is 238w close to the TDP.

Superior Stability by GIGABYTE BETA BIOS with Intel Baseline on Z790/B760 Motherboards

GIGABYTE TECHNOLOGY Co. Ltd, a leading manufacturer of motherboards, graphics cards, and hardware solutions, released the latest beta BIOS with Intel Baseline feature on Z790, B760 series motherboards for enhanced stability, regarding the feedback from Intel that high power consumption settings...

www.techpowerup.com

btw the picture is from 2021

If they were not published anywhere, where did flare get his info from? note also his info is years old so its been published for a while. Also how did Gigabyte end up with the correct settings if they not available?

A TDP doesnt mean thats what a bios should be configured to.

Will be keeping my PL2 as 175w for sure now. PL1 is 175w also at the moment, will think on that one dropping to 125w, probably not a big deal as now I no longer software encode I wont have any workload that uses anything like that power draw. I havent even got a game to use over 50w yet.

I do agree with flare Intel are guilty of letting it all get to this point, its just that I think the board vendors are guilty as well, thats where I disagree with HUB.

 

[Crackong]

Yes 188w appears to be the right setting. PL4 is 238w close to the TDP.

If they were not published anywhere, where did flare get his info from? note also his info is years old so its been published for a while. Also how did Gigabyte end up with the correct settings if they not available?

A TDP doesnt mean thats what a bios should be configured to.

Will be keeping my PL2 as 175w for sure now. PL1 is 175w also at the moment, will think on that one dropping to 125w, probably not a big deal as now I no longer software encode I wont have any workload that uses anything like that power draw. I havent even got a game to use over 50w yet.

I just find all these very confusing.

Buildzoid @Actually Hardcore Overclocking just posted a video discussing Gigabyte baseline profile and it seems the voltage is skyrocketed
It doesn't seem 'Normal'..