AMD not supporting Zen3 on older motherboards :(

[Patrik_]

Just now, leadeater said:

As I said you should fall under the Beta support like last round, at least then it's not 'officially supported' by AMD but they are still giving what is necessary to board vendors so they can support you if they wish.

 

B550 should never have been later than 2-3 months after X570, it's going to be a full year or more between the two.

I agree on 100% I really just want to upgrade my zen+ 2600 to Ryzen 4700X so I finally dont have to worry about new cpu for other 4 year atleast. 

[spartaman64]

14 hours ago, JoostinOnline said:

I realized that later, but that just makes it worse.  It went from "We can't do it" to "We won't do it".  At least that's how it seems with the information given.  AMD's biggest selling point was the upgrade paths, at least to me.  With so much changing right now, building didn't feel like a roll of the dice.  It's also worse because they've only just now released the B550 boards.  The three systems I've built in the past 6 months can't be upgrade anymore.

i mean they can probably remove cpu support for other cpus to add support for ryzen 4000 but that would create a massive headache. what if someone buys a ryzen 1700 and a b450 motherboard and it doesnt work and they search it up and it says they need to revert it back to an old bios version so they need to scramble to find someone with a ryzen 3000 or 4000 cpu they can borrow to downgrade their bios. and then later they get a ryzen 4600 and they put it in and then it doesnt work and they find out they need to update the bios. and probably 80% of consumers dont even know what a bios is let alone how to update it

 

also theres motherboard product lineups so theres the regular b450 tomahawk motherboard that dont support ryzen 3000 and need a bios update and you need to buy a ryzen 2000 cpu to perform the update on some of them. and then theres the b450 tomahawk max motherboard that support ryzen 3000 out of the box but not ryzen 4000 and then are they going to release a b450 tomahawk max ultra? that supports ryzen 4000 but not ryzen 1000. so you are going to have 3 different versions of every b450 motherboard just to denote out of the box compatibility 

 

and this is all assuming that clearing out ryzen 1000 makes enough room for ryzen 4000 what if theres more ryzen 4000 products or ryzen 4000 code takes up more space which is pretty likely then they would need to either remove some or all ryzen 2000 support also or only support some ryzen 4000 cpus

[leadeater]

Just now, Patrik_ said:

I agree on 100% I really just want to upgrade my zen+ 2600 to Ryzen 4700X so I finally dont have to worry about new cpu for other 4 year atleast. 

I'm not sure some of the high end 3000 series won't also cover that but if you know new CPUs are coming out why buy the older one, here it'll be because you have to rather than by choice.

 

Like I'm still running a 4930k and I want to upgrade but I haven't hit a situation that has required me to yet, Ryzen 3000 is worlds better than what I have. CPUs can last a very long time without performance problems, so long as you're not trying to do high end edge cases like very high refresh rate (over 140). Pushing money in to GPU has and still is the biggest thing by a lot, maybe that will change over the next 4 years though. Have to account for new consoles and just how much more CPU power Ryzen has brought to gamers.

[leadeater]

3 minutes ago, spartaman64 said:

i mean they can probably remove cpu support for other cpus to add support for ryzen 4000 but that would create a massive headache. what if someone buys a ryzen 1700 and a b450 motherboard and it doesnt work and they search it up and it says they need to revert it back to an old bios version. and then later they get a ryzen 4600 and they put it in and then it doesnt work and they find out they need to update the bios.

 

also theres motherboard product lineups so there regular b450 motherboards that dont support ryzen 3000 and need a bios update and you need to buy a ryzen 2000 cpu to perform the update on some of them. and then theres the b450 max motherboard that support ryzen 3000 out of the box but not ryzen 4000 and then are they going to release a b450 max ultra that supports ryzen 4000 but not ryzen 1000. 

 

and this is all assuming that clearing out ryzen 1000 makes enough room for ryzen 4000 what if theres more ryzen 4000 products or ryzen 4000 code takes up more space which is pretty likely then they would need to either remove some or all ryzen 2000 support also or only support some ryzen 4000 cpus

You forgot about APU support which is a bigger thing on B series chipsets as those are an architecture behind but carry the newer generation product naming. So if you drop Zen+ support then you can put in a Ryzen 3700 but not a Ryzen 3400G.

[Patrik_]

1 minute ago, leadeater said:

I'm not sure some of the high end 3000 series won't also cover that but if you know new CPUs are coming out why buy the older one, here it'll be because you have to rather than by choice.

 

Like I'm still running a 4930k and I want to upgrade but I haven't hit a situation that has required me to yet, Ryzen 3000 is worlds better than what I have. CPUs can last a very long time without performance problems, so long as you're not trying to do high end edge cases like very high refresh rate (over 140). Pushing money in to GPU has and still is the biggest thing by a lot, maybe that will change over the next 4 years though. Have to account for new consoles and just how much more CPU power Ryzen has brought to gamers.

I know but Im running 1440p 144hz display so I need to fuel it somehow :DD 

[Andk1987]

Seems a bit more clarification and discussion is happening now.

[IntMD]

Just posted this in the Hardware Unboxed video, but thought I would put it here too with some minor modifications and some additional info...

 

I bought a B450i Aorus Pro wifi in November, along with a 2700X (it was on sale for USD130.00 in Microcenter when I was in the US on holiday) - my personal first ever purchase of an AMD cpu at all.

 

The plan was to skip Zen 2 and then get a Zen 3 chip 4700X or similar. I was also going to wait and see which ITX B550 boards, if any, had the USB C front panel connector and pick up one of those when they were eventually out. In which case, what they are saying wouldn't really affect me in reality, as I'm not going to pick up the only ITX (or rather, DTX) AM4 board with a USB C front panel connector, because;

 

1. It's an Asus board

2. I costs more than a 3800X, and I'm done with spending >£300.00+ on a motherboard these days

 

After this announcement though, I might just wait for AM5 (or whatever Intel are on at the time if they get their shit together) or whatever it gets called. I might also, just pick up a secondhand 3700X when the prices drop sufficiently on those for me to care enough about getting one. Either way, neither the AIBs or AMD will be getting any further money directly from me for any AM4 (at the very least), if it things are going to go the way AMD are currently saying it will.

 

Intel aren't good about stuff like this either, but considering the entire Gigabyte X570 motherboard lineup only has 16mb BIOS chips apparently so I doubt it is unlikely to be higher than that just for Zen 2 & 3 being supported (assuming they even have to drop the Ryzen+ & the AF models compatibility in the BIOS update) or Gigabyte are also screwed, and the MSI Max boards stating what they did I understand why some people are super salty about it. For me, I'll probably just end up new GPU later this year like I had already planned, and save some money on upgrading the CPU & MB.

 

 

[komar]

I will just say to all of this, I'm an owner of x370 asrock killer sli board and a ryzen 5 1500x, bought soon after launch.

I just checked the CPU supported list and there is the 3950 there which is a 16 core CPU, yet my quad core is still more than enough for anything I do.

 

TBH I myself didn't believe that my x370 would get support for the 3000 series since this is 

 

Back in the day I remember AMD promising the AM4 platform to feature support for 3 generations - I have this. >>> I'm a happy AMD customer.

 

 

 

[Bombastinator]

It’s going to be some time before this one is resolved I think.  One thing that isn’t clear is if this is

1:a bios space issue, 

2: a capacity of AMD to produce the code issue, 

3: a 4000 series chips can’t deal with 400 series chipsets at all issue

4: a we just don’t feel like it issue. 

1: might be deal with able for some but not all boards.

2 is marginally reasonable.

3 is quite reasonable

4 is totally unreasonable

 

The truth will out on that one eventually.  but it’s going to take some time.  If the answer is 4, AMD could be back where it started or worse. It’s difficult to know till the 4000 series microcode appears though.  

[Patrik_]

3 hours ago, IntMD said:

Just posted this in the Hardware Unboxed video, but thought I would put it here too with some minor modifications and some additional info...

 

I bought a B450i Aorus Pro wifi in November, along with a 2700X (it was on sale for USD130.00 in Microcenter when I was in the US on holiday) - my personal first ever purchase of an AMD cpu at all.

 

The plan was to skip Zen 2 and then get a Zen 3 chip 4700X or similar. I was also going to wait and see which ITX B550 boards, if any, had the USB C front panel connector and pick up one of those when they were eventually out. In which case, what they are saying wouldn't really affect me in reality, as I'm not going to pick up the only ITX (or rather, DTX) AM4 board with a USB C front panel connector, because;

 

1. It's an Asus board

2. I costs more than a 3800X, and I'm done with spending >£300.00+ on a motherboard these days

 

After this announcement though, I might just wait for AM5 (or whatever Intel are on at the time if they get their shit together) or whatever it gets called. I might also, just pick up a secondhand 3700X when the prices drop sufficiently on those for me to care enough about getting one. Either way, neither the AIBs or AMD will be getting any further money directly from me for any AM4 (at the very least), if it things are going to go the way AMD are currently saying it will.

 

Intel aren't good about stuff like this either, but considering the entire Gigabyte X570 motherboard lineup only has 16mb BIOS chips apparently so I doubt it is unlikely to be higher than that just for Zen 2 & 3 being supported (assuming they even have to drop the Ryzen+ & the AF models compatibility in the BIOS update) or Gigabyte are also screwed, and the MSI Max boards stating what they did I understand why some people are super salty about it. For me, I'll probably just end up new GPU later this year like I had already planned, and save some money on upgrading the CPU & MB.

 

 

yeah X570 is really expensive ironicly when comes to intel platform all you need to get 5Ghz is just something like Asrock Extreme 4 and you are good, I know that X570 has PCIE 4.0 but almost noone needs it. 

[NumLock21]

6 hours ago, spartaman64 said:

i mean they can probably remove cpu support for other cpus to add support for ryzen 4000 but that would create a massive headache.

Massive annoyance for cpu support and security patches released via bios updates. Updating to the latest bios patched up the security holes, but it can make your CPU not compatible. Use the older bios that supports your CPU, but you don't have the latest security patch.

 

For example, here are the bios for my board. Bios 4101 has the latest patch but it will make my CPU not compatible, where I have to go out and buy a new CPU. Now if I don't want to upgrade my CPU, I'm force to say with bios 3902, but I won't have the security patch found in bios 4101.

I'm also wondering If I update to bios 4101 just for the security patch, and downgrade back to 3902, so I can continue to use my old CPU, will I still have the security patches that was found in bios 4101. Are those security patches hard written into the bios rom chip, such that no matter what version of bios I used, the security patches are permanently there, or they get deleted when I used a bios version that doesn't have it.

 

 

 

[Joe4evr]

1 hour ago, NumLock21 said:

I'm also wondering If I update to bios 4101 just for the security patch, and downgrade back to 3902, so I can continue to use my old CPU, will I still have the security patches that was found in bios 4101. Are those security patches hard written into the bios rom chip, such that no matter what version of bios I used, the security patches are permanently there, or they get deleted when I used a bios version that doesn't have it.

I'm afraid things don't work that way.

[mr moose]

12 hours ago, leadeater said:

Um what? That's literally what you were doing during our conversation and never addressed the issue at all of when AMD were actually able to give out the no support information.

 

Have you not been reading anything I said?  You provide me with the information that we only just got as if that satisfies my gripe that they should have given that information out months if not a year ago.   I asked you to show me where AMD admit they aren't going to support zen3 on earlier boards at any time in the past and you haven;t been able to do that, all you could do it point to the information we only just received that started this whole thread. 

12 hours ago, leadeater said:

Every time you complain they should have said sooner is yet another instance of assuming they could have said sooner.

Your argument is we should have worked it out sooner from the information that was available.  If we the consumer should have been able to work it then what the hell stopped AMD from working it out or even telling us sooner.

 

 I really don't follow your logic. 

12 hours ago, leadeater said:

What if the soonest point in time was 3 days ago, or 10 days ago, or 3 weeks ago. Without knowing when they could have said complaining they didn't is irrelevant.

If the soonest time was 3 days ago then why are you arguing consumers should have worked it out last year?

 

12 hours ago, leadeater said:

You want them to have said sooner, an emotion. You don't know when they could have actually said anything, missing factual information.

I believe they did not just decide to can support right now, and frankly the assumption they didn't know sooner is so thin I can't even imagine trying to use it as a legitimate argument.

 

12 hours ago, leadeater said:

We also know nothing of the size and complexity of the Zen 3 AGESA code nor the impact of trying to package it with Zen 2 without Zen/Zen+ and all the other extra pitfalls that may come along with it, like the mish mash of APUs and how those are supported or customers updating BIOS version with CPUs install that would then not be supported and haven't purchased a Zen 3 or some other supported replacement.

No we don't, but AMD do.

12 hours ago, leadeater said:

All we have is a statement from AMD saying the decision was made based on existing BIOS size limitations, everything else is pure guessing and speculation, time frames included.

Which according to you,  we consumers should have been guessing and speculating this lack of support based on last years issues, rather than the marketing from AMD.

 

12 hours ago, leadeater said:

And you already know my opinion on the matter, we had fair warning support might be a problem during the Ryzen 3000 launch.

 

I know you keep saying that,  from what you have said now it appears you think consumers should have worked it out before AMD.

 

 

[leadeater]

13 minutes ago, mr moose said:

If the soonest time was 3 days ago then why are you arguing consumers should have worked it out last year?

No my point was we were forewarned that these issues existed so buying older chipsets on boards with limited BIOS sizes could and is known to be a problem. You could have either gone with X570, waited, or purchased under the knowledge that future support might actually be a problem.

 

13 minutes ago, mr moose said:

I believe they did not just decide to can support right now, and frankly the assumption they didn't know sooner is so thin I can't even imagine trying to use it as a legitimate argument.

You have no idea how long AMD were working on trying to make support possible and from the communications with board partners with media they apparently had no idea themselves support was being dropped until we were told.

 

If you don't know when the decision was made and how much sooner it could have been made then you have no basis to complain about how much sooner they could have said anything. Wanting to be told sooner is totally different than having been able to have been told sooner.

 

So unless you have some actual factual information to base the complaint off and have a good idea of actually when they decided to drop support or found the engineering and product support was too much to do it the soonest they could have said was the 7th, so we were told as soon as AMD were ready to say. Obviously sooner is better.

 

13 minutes ago, mr moose said:

I asked you to show me where AMD admit they aren't going to support zen3 on earlier boards at any time in the past and you haven;t been able to do that, all you could do it point to the information we only just received that started this whole thread. 

What I was talking about at that time was Ryzen 3000, 3rd Gen, and the release situation around that. As we established at the time you were confused between that and Zen 3, I provided you information to what I was talking about and that situation as to why this was a known problem. The Zen 3 evidence is the OP and was also provided to you.

 

So no you had all the required evidence supplied to you you just lacked the understanding at the time to differentiate the two topics, Ryzen 3000/3rd Gen and Zen 3/Ryzen 4000.

 

If I'm talking about Ryzen 3000 and you're thinking I'm talking about Zen 3 then that is clearly going to be a problem.

[mr moose]

1 minute ago, leadeater said:

No my point was we were forewarned that these issues existed so buying older chipsets on boards with limited BIOS sizes could and is known to be a problem. You could have either gone with X570, waited, or purchased under the knowledge that future support might actually be a problem.

 

You have no idea how long AMD were working on trying to make support possible and from the communications with board partners with media they apparently had no idea themselves support was being dropped until we were told.

 

If you don't know when the decision was made and how much sooner it could have been made then you have no basis to complain about how much sooner they could have said anything. Wanting to be told sooner is totally different than having been able to have been told sooner.

 

So unless you have some actual factual information to base the complaint off and have a good idea of actually when they decided to drop support or found the engineer and product support was too much to do it the soonest they could have said was the 7th, so we were told as soon as AMD were ready to say. Obviously sooner is better.

 

What I was talking about at that time was Ryzen 3000, 3rd Gen, and the release situation around that. As we established at the time you were confused between that and Zen 3, I provided you information to what I was talking about and that situation as to why this was a known problem. The Zen 3 evidence is the OP and was also provided to you.

 

So no you had all the required evidence supplied to you you just lacked the understanding at the time to differentiate the to topics, Ryzen 3000/3rd Gen and Zen 3/Ryzen 4000.

 

 

So your entire argument sums up as:

 

consumers were forwarned (by observing resolved issues) and should have made precautions to buy motherboards with big enough bios so ensure future support, but AMD couldn't get their shit together to do the  same thing.

 

Lets ignore the fact that bios size is irrelevant if AMD aren't going to support boards that do have a big enough bios anyway.

 

 

Your Entire argument relies on consumers knowing more about AMD products than AMD does.

 

 

 

[leadeater]

9 minutes ago, mr moose said:

consumers were forwarned (by observing resolved issues) and should have made precautions to buy motherboards with big enough bios so ensure future support, but AMD couldn't get their shit together to do the  same thing.

The issues were not resolved, Beta support was provided by AMD to board vendors that wished to allow processor support on chipsets that AMD themselves don't list as supported or label as Beta only. And there are still boards that cannot accept all Ryzen CPUs and boards without BIOS options to reolve that. Unless there is full and official support on all boards and on all chipsets then it is not resolved, workaround at most.

 

This same method could be done for B450 and X470, maybe it can't. It probably can but AMD are saying they won't. Doesn't change the fact we knew about these problems.

 

9 minutes ago, mr moose said:

Your Entire argument relies on consumers knowing more about AMD products than AMD does.

No it does not, it relies on you paying attention and applying a bit of forethought with that knowledge. If you know the problems existed before assuming it won't happen again is naive.

 

And the problem is B550 being released so late so enough time can pass for people to forget or get complacent. And yes they should have made the decision about support during the 500 series announcements so that it could have been communicated, but they hadn't so it wasn't.

[Bombastinator]

6 hours ago, Patrik_ said:

yeah X570 is really expensive ironicly when comes to intel platform all you need to get 5Ghz is just something like Asrock Extreme 4 and you are good, I know that X570 has PCIE 4.0 but almost noone needs it. 

It’s not about ghz. Ghz is an indicator of speed but it is not speed itself.   Actual speed is what matters.  If an AMD cpu is faster at 4.0 ghz than. An intel is at 5.0ghz the AMD is faster.

[mr moose]

Just now, leadeater said:

The issues were not resolved,

 

But we can run 3000's series CPU's on even A320's,  looks resolved to me.  How can you argue that the ability to do what the issue said you couldn't means the wasn't resolved?

 

Just now, leadeater said:

Beta support was provided by AMD to board vendors that wished to allow processor support on chipsets that AMD themselves don't list as supported or label as Beta only. And there are still boards that cannot accept all Ryzen CPUs and boards without BIOS options to reolve that. Unless there is full and official support on all boards and on all chipsets then it is not resolved, workaround at most.

Ahh, the 100% issue, unless it is 100%  we can pretend it is a 0%.  In the eyes of a consumers there were issues and most of them were resolved, AMD didn't claim this meant anything and kept on with business as usual, therefore to conclude that meant future CPUs weren't going to be supported requires assumptions.  It is more logical in the absence of official information, that the consumer will see a company working to make all CPU's supported on all chipsets, all the issues show is that there were issues, they do not provide the consumer with any information regarding future support.

 

 

Just now, leadeater said:

This same method could be done for B450 and X470, maybe it can't. It probably can but AMD are saying they won't. Doesn't change the fact we knew about these problems.

And that is the crux of the issue, no consumer can work out what AMD will or won't support if that support in based on information the consumer doesn't have.  The only conclusion any consumer can make regarding future support is based on what the company tells them.  In this case AMD said nothing on the back of massive marketing basically claiming support for everything.

 

Just now, leadeater said:

No it does not, it relies on you paying attention and applying a bit of forethought with that knowledge. If you know the problems existed before assuming it won't happen again is naive.

 

And the problem is B550 being released so late so enough time can pass for people to forget or get complacent. And yes they should have made the decision about support during the 500 series announcements so that it could have been communicated, but they hasn't so it wasn't.

 

How can any consumer work out what will be supported when they don't know why the B550 was delayed, or why the bios were too small, or why many of them worked after being told they wouldn't,  then you have the problem of AMD not actually being specific about any of this.

 

Sorry but I can't accept that as even fair let alone realistic.

 

[leadeater]

2 minutes ago, mr moose said:

But we can run 3000's series CPU's on even A320's,  looks resolved to me. 

*Some boards, not all boards. It's not resolved. Not unless you purposely blind yourself and assume one or a few board that can mean EVERY board with that chipset can. Can ALL of them? No.

 

Honestly you're operating on wishful thinking, I really don't care. What I care about was the quoted statement.

 

13 hours ago, mr moose said:

I am not naive to companies doing the dirty on their customers, I just won't make/accept accusations based of hearsay or emotion.

Maybe actually do what you say rather than be selective about it. Everything you are saying is nothing but emotion and hearsay, everything.

[tehxf]

13 hours ago, Bombastinator said:

There are no levels. “Better” “worse” and “the same” are not the only options.  That’s a simplification prone to error.
It’s known as binary thinking.  
Seeing this mistake use to manipulate people in American politics a lot lately.

 

I believe that AMD as a rule has been slightly less bad about some things, much less bad about some things, just as bad about other things, and somewhat worse about a couple others.

 

so perhaps less bad though less than perfect, but it depends on how you weight what.  It’s complicated enough that binary “just as bad” isn’t accurate.

No, as a business they are the same. That is the point i am making. Since when have these giants genuinely cared about the end user... they dont. So changing chipsets should not be a surprise to anyone...

[Neftex]

1 minute ago, leadeater said:

*Some boards, not all boards. It's not resolved. Not unless you purposely blind yourself and assume one or a few board that can mean EVERY board with that chipset can. Can ALL of them? No.

not all of them can but its not AMD deciding they cant. got it? just support the chipset and let mobo makers decide where its applicable. from mobo manufacturers it seems that they would be able to make it work and bios size is NOT the deciding factor

[mr moose]

Just now, leadeater said:

*Some boards, not all boards. It's not resolved. Not unless you purposely blind yourself and assume one or a few board that can mean EVERY board with that chipset can. Can ALL of them? No.

It still tells the consumer that AMD is working at supporting all they can,  So when it comes to future CPU's unless AMD say they aren't going to be supported,  then the only conclusion you can draw without making assumptions is that AMD fully intend to support them on everything.  Especially the newer B450 boards.

 

And this is still ignoring the fact that If bios size was going to be a real problem then AMD should have known last year. Unless you are telling me that you worked out something about their new design that they didn't?

 

 

[leadeater]

5 minutes ago, Neftex said:

not all of them can but its not AMD deciding they cant. got it? just support the chipset and let mobo makers decide where its applicable. from mobo manufacturers it seems that they would be able to make it work and bios size is NOT the deciding factor

For AMD it was, it probably wasn't the only reason but it definitely was a reason. Why doing so is a problem has been explained by myself and others, so if you go back to the other page and have a read what was said about it that would help.

 

At no time have I said they shouldn't do it, it's just actually not totally simple nor without problems and things board partners do affect AMD because they have support their customer purchases, the CPUs.

[leadeater]

5 minutes ago, mr moose said:

And this is still ignoring the fact that If bios size was going to be a real problem then AMD should have known last year. Unless you are telling me that you worked out something about their new design that they didn't?

Ah yes so knowing about the problems means they immediately know the solution to the problem or have made a decision about it immediately, sure. That's totally how things work......

[Blademaster91]

10 minutes ago, leadeater said:

Ah yes so knowing about the problems means they immediately know the solution to the problem or have made a decision about it immediately, sure. That's totally how things work......

Yes because thats how it worked before? Even the high end B450 boards except for a few MSI max boards only have a 16MB bios chip, so they should've seen it coming, or at least communicated better with board partners so they aren't putting up false advertising.