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

[chrcoluk]

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.
View attachment 345550
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
View attachment 345551


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?
View attachment 345552


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?
View attachment 345553



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.

Dont disagree, lots of clickbait on videos, playing to the audience.

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'..

That might be due to loadline, I think board vendors have also been undervolting chips based on something I read. (might have been a TPU post, might have been reddit). Of course there also could be a bug in the bios, sadly these days bios bugs are common. I agree its confusing as its not just about power limits. Waiting for a food delivery, when its here I will watch the video, always appreciate his content.

 

[pressing on]

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'.

I haven't seen all the video but so far Buildzoid has identified that although voltages are higher than before they are still within the Intel maximum allowed limits. On his machine CB15 and CB20 didn't run before without crashing/freezing but they run OK now with the Intel Baseline Profile (IBP).

The two BIOS settings responsible for the higher voltages are "SVID Behavior" and "IA CEP". "SVID Behaviour" is set by the IBP to 'Intel Fail-Safe'. From reports it seems that as @chrcoluk has suggested this setting is something to do with loadline, possibly the equivalent of setting loadline to the max. "IA CEP (Current Excursion Protection)" is set to "Enabled" by the IBP, this is associated with higher voltages. There is a third setting in the IBP which is "IA TDC Current Limit" set to "Intel's Default". There is some feedback on the web that this setting is limiting maximum current to less than the Intel default.

 

[chrcoluk]

Buildzoid got a really bad 14900k (very likely not media sample I guess), and his board on clear cmos defaults to 280/4095.

Also glad he hasnt jumped to conclusions and makes an effort to read the documentation.

--

 

[Crackong]

Buildzoid got a really bad 14900k (very likely not media sample I guess), and his board on clear cmos defaults to 280/4095.

Also glad he hasnt jumped to conclusions and makes an effort to read the documentation.

It is quite entertaining see what he found within Intel documentations.
He identifies Intel give advises in the spec document encouraging MB vendors to undervolt the CPUs and
It could be exactly the reason why the Gigabyte bumps loadline to 1.7v in the baseline profile.

 

[Upgrayedd]

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.

Well that's not really a CPU review then is it? That's a budget build review. Not everyone is on a budget but they'd all like to know what the capabilities are of the review subject. Why 8000? Cause they're capable. Btw the product page for the 7800X3D mentions 5200 being the supported max speed but they're capable of up 6000MHz for 3:2 or whatever it is now they're calling the sweetspot cause it's not 1:1 anymore.

3Gb chips are the newer chips that make 24GB and 48GB sticks. You wanted to limit everything else about a theoretical test system now you want to also limit ram choices? I chose them cause they're the newest kits.

 

[persizi]

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

Haven't watched the video but I'm pretty sure the voltage is because of the insane AC LL of 1.7mohm. Even in the Intel specs for i9 those values are 1.1mohm which is still insanely high. Some vendors instead fix the problems they are making them worse.

 

[john_]

Well that's not really a CPU review then is it? That's a budget build review. Not everyone is on a budget but they'd all like to know what the capabilities are of the review subject. Why 8000? Cause they're capable. Btw the product page for the 7800X3D mentions 5200 being the supported max speed but they're capable of up 6000MHz for 3:2 or whatever it is now they're calling the sweetspot cause it's not 1:1 anymore.

I like how you conveniently bypassed the part where I say that the reviews could include 1-2 pages with the CPU unlocked and top of the line, ultra expensive hardware.

I am just saying they should focus in what 90% of consumers more or less will use. And no, a $300 motherboard, a $60 cooler, a PCIe 3.0 or 4.0 NVMe with speeds at 3-7GBs/sec, is NOT budget build. It's mid range to hi end build. Just not enthusiast, "no limits", "money is not a concern" build. It's another thing to know where a hardware part can go under perfect conditions and a totally different thing what they will end up having in their own system. Insisting on reviews that lie to the average buyer, is not what we should be asking. And reviews today lie. Because people who can reproduce benchmarking configurations and get what those reviews promise them, are probably 3-5% of the consumers. The rest only read lies.

3Gb chips are the newer chips that make 24GB and 48GB sticks. You wanted to limit everything else about a theoretical test system now you want to also limit ram choices? I chose them cause they're the newest kits.

MY GOD!!!!!!!! You throw a claim about 8000MHz chips being the same price as 6000MHz chips. Then instead of giving links you throw a BS about limiting choices. YOU LIMIT THE CHOICE TO 24GB sticks, when the market is full of 16GB, 32GB and 64GB choices.
Damn you are a waste of time. Obviously someone who just wants to promote a certain narrative WITHOUT ARGUMENTS, because you have none. When someone asks you for a link, you throw that BS. Oh my God........

Don't waste my time kid.

 

[chrcoluk]

It is quite entertaining see what he found within Intel documentations.
He identifies Intel give advises in the spec document encouraging MB vendors to undervolt the CPUs and
It could be exactly the reason why the Gigabyte bumps loadline to 1.7v in the baseline profile.

That voltage bump is probably because a fixed LL is specified in the baseline, I suspect Asus have missed certain parts of it, hence Asus higher PL and lower voltage.

 

[Upgrayedd]

I like how you conveniently bypassed the part where I say that the reviews could include 1-2 pages with the CPU unlocked and top of the line, ultra expensive hardware.

I am just saying they should focus in what 90% of consumers more or less will use. And no, a $300 motherboard, a $60 cooler, a PCIe 3.0 or 4.0 NVMe with speeds at 3-7GBs/sec, is NOT budget build. It's mid range to hi end build. Just not enthusiast, "no limits", "money is not a concern" build. It's another thing to know where a hardware part can go under perfect conditions and a totally different thing what they will end up having in their own system. Insisting on reviews that lie to the average buyer, is not what we should be asking. And reviews today lie. Because people who can reproduce benchmarking configurations and get what those reviews promise them, are probably 3-5% of the consumers. The rest only read lies.

MY GOD!!!!!!!! You throw a claim about 8000MHz chips being the same price as 6000MHz chips. Then instead of giving links you throw a BS about limiting choices. YOU LIMIT THE CHOICE TO 24GB sticks, when the market is full of 16GB, 32GB and 64GB choices.
Damn you are a waste of time. Obviously someone who just wants to promote a certain narrative WITHOUT ARGUMENTS, because you have none. When someone asks you for a link, you throw that BS. Oh my God........

Don't waste my time kid.

Showing full potential is lying? Umm ..What??
I'm not your personal shopper. What you're asking for isn't a product review then and should probably be an entirely different article about average builds and any performance loss. 24GB is the largest SR sticks available, I'm pretty sure they don't make SR 32GB sticks.

 

[john_]

Showing full potential is lying? Umm ..What??
I'm not your personal shopper. What you're asking for isn't a product review then and should probably be an entirely different article about average builds and any performance loss. 24GB is the largest SR sticks available, I'm pretty sure they don't make SR 32GB sticks.

Blah blah blah you keep pretending that you don't read what I am posting, still BS excuses with no links because maybe even in the LIMITED market of 24/48GB RAM you can't find a 8000MHz dimm being as cheap as a 6000MHz dimm.

I am blocking you kid. Post whatever BS you like under my posts. I'll miss nothing of any importance.

 

[Upgrayedd]

Blah blah blah you keep pretending that you don't read what I am posting, still BS excuses with no links because maybe even in the LIMITED market of 24/48GB RAM you can't find a 8000MHz dimm being as cheap as a 6000MHz dimm.

I am blocking you kid. Post whatever BS you like under my posts. I'll miss nothing of any importance.

Ok.. just trying to have a conversation. Don't understand the hostility.

 

[chrcoluk]

Found this (ironically by accident was searching for something else).

An intel rep told someone running their CPU with a higher tjmax is a warranty breach (I assume customer would need to disclose this), which contradicts HUB's claim that only turbo clocks count as out of spec.

TjMAX is set to 115°C by default.

hello. I have a Core i9-13900K and an ASRock Z790 Steel Legend In this combination, TjMAX is set to 115°C by default. Therefore, even if the temperature reaches 100°C, the thermal throttle may not work and the temperature may rise to 115°C. Is there any problem if I continue to use it as it...

community.intel.com

Bad news for ASRock, as it seems they made this mistake on a second board, not just my board. Interestingly ASRock replied to someone who queried them about it stating they did it knowingly as a performance booster.

 

[HOkay]

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.

Reviews that state the test system are not misleading. I can go buy the exact same system & will get more or less the same results. I personally am more interested in seeing what a product can do with potential limitations from other components removed, than a "best value for money" system. Sure there's a limit to that, e.g. no I don't want liquid nitrogen cooling as the cooling solution for a full review (though I wouldn't mind as an extra page at the end for funsies) as that's not a thing I can easily buy & daily drive.

That said, I also want to see value for money reviews just as a point of interest too. I understand most people will probably buy the more value for money system & are more interested in what their system would perform like, but those don't tell me what the performance limits of a product are, they tell me what the performance of a product is in a value for money build.

Then there's content that uses the best possible components except for one which is changed to see the effects, e.g. changing RAM speeds or timings. Those are fascinating & help me decide where I want to land on that price to performance curve for that one component.

My tastes have changed over the years too of course. When I built my first PC it was all about value for money & compromising everywhere. I'm fortunate enough now that I can afford to spend more & get that little bit more performance for that extra chunk of money. I also only game on my home machine these days so the 7800X3D was an easy choice & I fortunately avoided all these latest Intel instability issues.

 

[john_]

Reviews that state the test system are not misleading. I can go buy the exact same system & will get more or less the same results. I personally am more interested in seeing what a product can do with potential limitations from other components removed, than a "best value for money" system. Sure there's a limit to that, e.g. no I don't want liquid nitrogen cooling as the cooling solution for a full review (though I wouldn't mind as an extra page at the end for funsies) as that's not a thing I can easily buy & daily drive.

That said, I also want to see value for money reviews just as a point of interest too. I understand most people will probably buy the more value for money system & are more interested in what their system would perform like, but those don't tell me what the performance limits of a product are, they tell me what the performance of a product is in a value for money build.

Then there's content that uses the best possible components except for one which is changed to see the effects, e.g. changing RAM speeds or timings. Those are fascinating & help me decide where I want to land on that price to performance curve for that one component.

My tastes have changed over the years too of course. When I built my first PC it was all about value for money & compromising everywhere. I'm fortunate enough now that I can afford to spend more & get that little bit more performance for that extra chunk of money. I also only game on my home machine these days so the 7800X3D was an easy choice & I fortunately avoided all these latest Intel instability issues.

You want reviews with no compromised specs and liquid nitrogen in a page or two.
I want to know the expected performance and have a page or two showing a no compromised specs scenario.
You want a "What if everything is perfect" review, I want a "What the vast majority of consumers will get" review.

I am also fortunate enough to have solved all my financial problems. But I wouldn't pay 30-50% more money to get that 1-5% extra performance. And that's the way the vast majority of consumers think and reviews should be targeting the vast majority first, keep a couple of pages for the enthusiasts. We where looking to get that 1-5% extra performance 15-20 years ago, when that extra performance mattered. Today we have hardware that is way overpower for the majority of tasks and that 1-5% extra performance is only needed for benchmark scores that almost none cares about. Overclocking, benchmark charts, perfect conditions scores, those where much more important and interesting in the past.

P.S. Now that I am thinking of it, maybe reviews should be done with those cut down almost butchered windows versions where half or more of background tasks are killed to not take any CPU time. I mean, I stopped caring about benchmarks when I have seen a Windows XP version so much cut down that was looking like Windows 95. How about that? Should reviews go one step further and use Windows versions made especially for benchmarking? You talk about no potential limitations reviews. Why focus only on hardware? Maybe we should also use software that helps that "no limitations" scenario.

 

[HOkay]

You want reviews with no compromised specs and liquid nitrogen in a page or two.
I want to know the expected performance and have a page or two showing a no compromised specs scenario.
You want a "What if everything is perfect" review, I want a "What the vast majority of consumers will get" review.

I am also fortunate enough to have solved all my financial problems. But I wouldn't pay 30-50% more money to get that 1-5% extra performance. And that's the way the vast majority of consumers think and reviews should be targeting the vast majority first, keep a couple of pages for the enthusiasts. We where looking to get that 1-5% extra performance 15-20 years ago, when that extra performance mattered. Today we have hardware that is way overpower for the majority of tasks and that 1-5% extra performance is only needed for benchmark scores that almost none cares about. Overclocking, benchmark charts, perfect conditions scores, those where much more important and interesting in the past.

P.S. Now that I am thinking of it, maybe reviews should be done with those cut down almost butchered windows versions where half or more of background tasks are killed to not take any CPU time. I mean, I stopped caring about benchmarks when I have seen a Windows XP version so much cut down that was looking like Windows 95. How about that? Should reviews go one step further and use Windows versions made especially for benchmarking? You talk about no potential limitations reviews. Why focus only on hardware? Maybe we should also use software that helps that "no limitations" scenario.

One reason I want to see the full performance of a component is because I strongly prefer a high frame rate when I'm playing on desktop, as in 90fps as a minimum, but 144Hz+ ideally. Some game engines don't like going that fast so it's useful for me to see for example a CPU with minimal limits from other components to see what sort of performance it can push vs the competition so I can decide if it's worth it to me. I want to know what % faster a CPU is in games in a best case scenario to help me extrapolate & apply that % to games I play that I know are CPU limited. I used to play an MMO which just needed the fastest possible single threaded performance for example.

Both types of review are needed as they serve different purposes, & everyone should be checking multiple sources anyway. Ideally every review would include both, but that's double the work so it's unlikely to happen. There's plenty of reviewers doing the every man builds for people to look at if they prefer to just see that data.

Re. Whether the extra performance matters - I agree, it doesn't really matter, 5% even 10% difference is a very similar experience, but I want to be right up at the top end anyway, just because I want to make sure I have the best possible experience.

Re. Cut down Windows install, my preferred is a fresh Windows install which most reviewers do I think, however if a review did use a cut down version I'd have no problem with that if they stated exactly what they've cut out. That might even push me to cut down my Windows install since my desktop PC at home is purely for gaming as I have a company laptop for work.

Also I just want to say thanks for a calm response with some good points for us to debate!

 

[OkieDan]

I watch/read CPU reviews to see limits of CPU, not the limits of a motherboard. I use motherboard reviews to pick a motherboard after I've decided on the CPU I want, paying special attention to VRM quality. There could be a better way to go about it but it seems like the best way for me.

There's often times vastly different VRM quality differences in mid-range boards so picking a mid range board for a CPU review doesn't necessarily help anyone out. Perhaps anyone wanting mid and low teir bords used to review high end CPUs should consider looking at motherboard roundups instead of CPU reviews.