AMD Backpedals, Zen 3 Support Coming to B450 and X470

[Ultra Taco]

I just wanted to thank TPU for sparking consumer outrage toward AMD for the second time.

Keeping the multi billion dollar companies honest can only be done if there is journalism that feels the same dissapointment and outrage over a company decieving or going back on it's word, and if there is a community of folks that will stand up to (while remaining seated) this injustice.

It wasn't but last year that TPU made a similar article about how b350 370x motherboards wouldn't support zen 2 / mattise.

There was huge backlash and policy reversed.

If only actual politics worked this justly and swiftly. The FCC just getting it's way again and such.

A meager victory, but still, a victory for the people.

 

[Turmania]

The problem with AMD is that their departments arennot unified it seems and one department moves its legs. The other picks its nose with its fingers. They need to be on same page,their marketing department has always been their weakest department.

 

[InVasMani]

All this outrage at AMD meanwhile Intel deliberately and blatantly lied about Z170. AMD has a fairly reasonable excuse at least on top of zen to zen 2 upgrade path being far better than skylake to kabyfake upgrade path was.

 

[newtekie1]

Wrong, just watch the video

The BIOS size limitation was not a good reason to limit support on all 400 chipset boards as there are plenty of boards out there with 32MB BIOS chips. And it was clearly a 100% excuse to try to increase profits.

In fact, AMD should have never taken it upon themselves to make the decision. The motherboard manufacturers should have been left to figure out how to support the new processors. The ones that used 16MB chips could either flash a BIOS that only supports Zen3 or not support Zen3 at all. The ones that used 32MB chips can just flash the BIOS like normal and move on.

Intel deliberately and blatantly lied about Z170.

How did Intel lie about Z170?

 

[Amite]

Given the slowdown AMD was about to shoot their selves in the foot

 

[hat]

Very well... I find this acceptable. Stay on this path, AMD, as your processors continue to kick ass, and you'll have a lot of happy hobbyists, at least.

Meanwhile, Intel is releasing the same processor on the same manufacturing process on a new motherboard again and everything's still locked down unless you pay big bucks for the K series. They got away with this for a while because their processors were simply better, but now the tables have pretty much turned.

 

[Dyatlov A]

It is very nice from AMD, that is why I will keep with them. So even if they told it will not work, they listened and later told OK we will make it work. Of course it is possible, Intel could do similar things too or like all the new hardware support would be possible for Windows 7, if they would like to do it.

 

[windwhirl]

Can't one just flash an older version if they need to? What's exactly a one-way bios update? Sounds like some randomly made up name to me.

Technically that's probably correct and true and it is a one way ticket, but that doesn't mean you can't then take a one-way ticket backwards. I think what AMD means is once you update it for Zen 3 that's all the BIOS ROM is able to storage information for so Zen/Zen 2 support is lost. That said once on a Zen 3 BIOS ROM update with a Zen 3 CPU I can't foresee a barrier reason why you couldn't flash the BIOS ROM to a Zen/Zen2 BIOS ROM and switch the CPU back out for a Zen/Zen 2 CPU in place of the Zen 3 CPU. That of course is a rather unlikely scenario in the first place. How many people would actually bother to do all of that. It's a one way ticket in a sense in either direction due to the limited ROM storage space to support the assortment of Zen CPU's of a given generation due to the variety of various SKU's between each generation of them.

AMD answered that question to Gamers Nexus' Steve.

Apparently, that kind of operation is also one the most risky ones, with a relatively high chance of bricking the board...

Source:

starting at 11:29.

Also, 300 series boards are NOT supported on Zen 3.

 

[evernessince]

The BIOS size limitation was not a good reason to limit support on all 400 chipset boards as there are plenty of boards out there with 32MB BIOS chips. And it was clearly a 100% excuse to try to increase profits.

In fact, AMD should have never taken it upon themselves to make the decision. The motherboard manufacturers should have been left to figure out how to support the new processors. The ones that used 16MB chips could either flash a BIOS that only supports Zen3 or not support Zen3 at all. The ones that used 32MB chips can just flash the BIOS like normal and move on.



How did Intel lie about Z170?

Again watch the video. People are bringing up points that are discussed in it. Steve takes the time to ask the board partners, AMD, and BIOS engineers who make both AMD and Intel BIOS. I can tell you for a fact that according to the engineers that make motherboard BIOS that it is not nearly as simple or devious as you make it sound. Put away your pitchfork and educate yourself.

I'm not saying that AMD pulling support was right, just that it can certainly be justified from a consumer ease of use perspective and from a technical one. Once again, this is explained far more in depth in the video.

 

[ahenriquedsj]

Nice

 

[InVasMani]

How did Intel lie about Z170?

Intel's excuse for Z270/Z370 to support later CPU sku's was pretty much proven patiently false. They had had made some claims and remarks in the past for reasons that they needed to sell people z270/z370 chipsets basically that didn't hold weight and sure if you want optane that's a another subject though I wouldn't doubt if that's also another Intel self cooked artificial limitation.
https://www.pcbuildersclub.com/en/2018/11/intel-core-i9-9900k-runs-on-z170-motherboard-with-5-5-ghz/

 

[dyonoctis]

It's great that it's happening, but some people are acting as if it was eazy peazy. It seems like they didn't get the memo about how restricted this is , zen 2 on B350/X370 was a walk in the cake compared to this.

You will have to send proof that you own a 400 motherboard and bought a zen 3 cpu before seeing any Bios. You are not assured to get it on launch day, and we know nothing about the support that we will get if bugs like destiny 2 not working on zen 2 happens again. In my case I got that uppgrade 1 month later.

Again it's great, but don't downplay the whole thing.

 

[btarunr]

Intel's excuse for Z270/Z370 to support later CPU sku's was pretty much proven patiently false. They had had made some claims and remarks in the past for reasons that they needed to sell people z270/z370 chipsets basically that didn't hold weight and sure if you want optane that's a another subject though I wouldn't doubt if that's also another Intel self cooked artificial limitation.
https://www.pcbuildersclub.com/en/2018/11/intel-core-i9-9900k-runs-on-z170-motherboard-with-5-5-ghz/

Z270 (14 nm) is a newer chipset than Z170 (22 nm). It offers 24 PCIe downstream lanes (vs. 20 on the Z170). Although I agree, Intel's official reason for segmenting 300-series apart from 200-series was teddy poo.

 

[Countryside]

Can you feel the support

 

[Deleted member 190956]

It's funny seeing the fanboys rage that we "stupid users" will mess up what we've been asking all along. Guess they can't stomach that people critical of AMD are getting what they wanted. AMD knew the BIOS limitations and difficulties from the get-go, when they insinuated future support for Zen3. They shouldn't have done so, but it's a good look for them that they didn't insist on screwing their customers over "for their own good". Good job AMD.

 

[Vayra86]

Well, sort of. I mean, the board makers are also going to have to be willing and it's possible and this is pure speculation, that they approached AMD and said hey, we didn't put large enough flash chips for the UEFI on these boards, so we don't want to support them for your next gen CPUs. Obviously the way it is now, it's down to each board makers to decide which boards may or may not get support, which sort of absolves AMD and the board makers.

And AMD has successfully created yet another unclear platform, where you could be getting anything in terms of support. And that is with product lines already out in the wild. *cough* NAVI *cough*

They just never really learn to pre empt this crap properly. Why oh why... Its like they just don't like being awesome.

AMD. They said that it would be a one-way BIOS update for 400 series boards. That is, once you update the BIOS to get Zen 3 compatibility, you can't go back to previous versions, meaning you lose compatibility with the CPUs for which support was removed in the BIOS update.

Wow. Next level.

I'm not saying that AMD pulling support was right, just that it can certainly be justified from a consumer ease of use perspective and from a technical one. Once again, this is explained far more in depth in the video.

That is the world upside down really isn't it. What you expect in the normal world is that when support is added, you get to update through a simple process and are not SOL if it somehow fails. What we have now is an after-the-fact 'oops, yeah its not optimal, but hey, deal with it' situation because AMD was being vague until the very last moment.

This is the very thing that has always plagued AMD. It never looks really professional. And then they seem to recover, promising everyone it will be better this time... and devolve into the same shit once more. Its a sign of company culture...

 

[oxidized]

I think that what AMD meant is that you will not be able to go back to an older BIOS version the "easy way", that is, using the BIOS/UEFI setup program or the firmware updater that usually comes already bundled with the BIOS. You can always take the chip out and flash it with an EEPROM programmer or similar tool if you have the skill for that.

Why you wouldn't be able to do it the easy way? I do not know the internals of a BIOS update process, so this is just my guess and take it with a truckload of salt:

1-The updater checks the version number and doesn't allow you to flash an older version. Probably the most simple way to block it. Getting past that limitation could be as easy as changing the new/old firmware filename or as complicated as reverse-engineering the whole thing.
2-The update/downgrade process works in such a way that it would end up being truncated half-way through if you were using a Zen 3 processor, which would leave you with a bricked board.

If there is someone around who may actually know, it might be @R-T-B ...

AMD answered that question to Gamers Nexus' Steve.


Apparently, that kind of operation is also one the most risky ones, with a relatively high chance of bricking the board...

Source:

starting at 11:29.

Also, 300 series boards are NOT supported on Zen 3.

Well let's wait from mobo manufacturers about bios flashback capabilities, it might be possible to flash older bioses through those features.

 

[Frick]

If you want to buy a Ferrari buy a Ferrari don't try making a 700BHP mini which is basically what's going on here people with B450 motherboards (Mini's) wanting the same power as a X570/B550 (Ferrari)...

More like if you want a car with driver assistance tech don't buy car without it. It's not like a fancy motherboard actually does something unless you use the extra features.

 

[medi01]

I was critical of AMD for AM4 promises not delivered, but they did indeed have objective reasons, other than greed, to not support older mainboards.

There is so much that can go wrong and will go wrong.

WIth this "step back" nobody would blame AMD once shit starts happening.

 

[dyonoctis]

Now that the issue has been fixed on 400, a fews 300 owner's on reddit are suddenly stepping out of the shadow and want to join the fun as well.

 

[Chrispy_]

I'm sure it'll make a lot of people happy, and it also means that AMD are continuing to offer support for the latest processors in older AM4 boards.

But this move has come at the expense of older chips. It's likely to impact anyone who had a Bristol Ridge APU and changed motherboard (upgrade or replacing a dead one) and it will make things difficult in the used market where you can no longer guarantee that a 400-series board has an older BIOS you need.

Overall, it's a good thing in my opinion but we shouldn't be calling it a victory - it's just that the noisy minority (enthusiasts) have influenced AMD to do what enthusiasts want, at the cost of making things worse for the silent majority.

Zen3 on 400-series wasn't AMD being lazy, it was a fork in the road, and they got shouted at by enthusiasts for picking the anti-enthusiast fork. Whilst we as a demographic have benefitted, others have now lost out - including the demographic that has the least disposable income to deal with the problems this direction will cause. I hope ya'll remember that when helping people in the forums over 400-series compatibility issues a year or two from now.

 

[Kapone33]

The squeaky gear gets the oil? On a BIOS made for PCIe 4.0 specs in the CPUs but let's put them in B450 boards? AMD lied to us? From A320 to X470 how many CPUs can you insert? PCIe 4.0 for $200 ($300 Cad) a 3300 and B550 will do that and oh they made some B550 boards fully compatible with the 1 or 2 PCIE 3.0 NVME drive you have in your current AM4 and you wouldn't even need to worry about buying a new Windows key. The objective fact is the 32MB BIOS is made for PCIe 4.0 and Ryzen3 CPUs are PCIe 4.0. You could put a Corvette engine in a Cavalier, that does not mean that the Cavalier will last unless you do structural changes to accept all of the added horsepower. It is the same thing here.

 

[newtekie1]

Again watch the video. People are bringing up points that are discussed in it. Steve takes the time to ask the board partners, AMD, and BIOS engineers who make both AMD and Intel BIOS. I can tell you for a fact that according to the engineers that make motherboard BIOS that it is not nearly as simple or devious as you make it sound. Put away your pitchfork and educate yourself.

I watched it. Most of his points don't even make sense. He's talking about things that are motherboard makers decisions to make, not AMDs. AMD's decision to just outright not support the 400 series chipsets was wrong and done from a purely business standpoint(for them).

Steve talks in that video like flashing BIOS is this huge scary thing that most people don't know how to do. But if you have an older motherboard and are buying a newer CPU for it, which is what this whole fiasco is about, then you expect to have to flash your BIOS to support the new CPU. It's a common thing.

He talks about how removing older CPU support would be like "pulling the rug out from under people", literally his words. But we've already seen that, and it isn't an issue. There are plenty of 16MB boards that have lost support for older processors already. Right now, on the 32MB ROM boards the issue is anything older than the 3000 CPUs can't address more than 16MB. That's fine, then you make 3000 the stepping stone. If you want 4000 CPU support, you need to use a 3000 series CPU to flash the BIOS first. Then the board will only support 3000 and 4000 CPUs. What are the chances someone will downgrade afterwards anyway? I'd say 99% of people upgrade their CPU to the new one and then never touch it again.

He talks people might buy a B450 motherboard new that has the latest "Zen3 Only" BIOS on it, but wanting to use that board with an older processors. Ok, we have methods to handle that too. Just like the stickers that are put on motherboard boxes now that let users know the motherboard already has the BIOS to support 3000 series CPUs. They put a sticker on the motherboard that says "This Board Only Supports Ryzen 4000 CPUs". Problem solved.

And at the end of the day, if you watched the video, the most important thing that he says is "B450 Can Support Zen3". That's all you need to know. The issues that arise by the smaller BIOS on a lot of boards have already had workaround developed. I do not buy this "we did it for consumer ease of use" crap. They did it because they want to sell chipsets and make more money, period!

Intel's excuse for Z270/Z370 to support later CPU sku's was pretty much proven patiently false. They had had made some claims and remarks in the past for reasons that they needed to sell people z270/z370 chipsets basically that didn't hold weight and sure if you want optane that's a another subject though I wouldn't doubt if that's also another Intel self cooked artificial limitation.
https://www.pcbuildersclub.com/en/2018/11/intel-core-i9-9900k-runs-on-z170-motherboard-with-5-5-ghz/

What? First off, Z170 supports all the same CPU that Z270 supports. I think your issue is with the requirement of Z370 for CoffeLake. And there are technical reasons for that. Not specifically with the chipset, but the socket. This has been gone over many times. The number of power delivery pins on the socket need to be increased to support more CPU cores, and the power draw that goes along with them. It is technically possible to get a Coffeelake CPU working on Z170/Z270 motherboards. The issue is reliability over years of use. You can overload power delivery for short amounts of time, and all seems fine, but over long periods of time bad things can happen. That was the issue. The chipset distinction was done to make it extremely easy to tell what motherboard supported new vs old processors.

 

[Countryside]

Now that the issue has been fixed on 400, a fews 300 owner's on reddit are suddenly stepping out of the shadow and want to join the fun as well.

Not gonna happen, my x370 board is EOL, Asus hasn't even released bios with agesa 1004, last bios update was 7 months ago.

 

[TechLurker]

In line with the drama to some extent, I expect that AMD will no longer guarantee more than 2 series (not necessarily generations) worth of CPUs per socket; any compatibility after is just because AMD was able to make it work and isn't because they promised. If a CPU is likely to need a slight revision and a slightly newer socket, I expect them to maybe pull a Socket+ (as they did with AM2+ and AM3+) and still allow compatibility with older CPUs while restarting the upgrade path for at least another 2 series. I do still expect AMD to also end up dropping old CPUs over time with such a scheme (especially if it stretches say, 5-6 series on a Socket+), even if all of them are standardized to 32MB BIOS.