AMD Backtracks: Zen 3 Support on B450 and X470!

Edit: The chart, the line that says Ryzen 4000 APU should be listed as **Assumed, given the parts aren't out yet. I essentially created the table while getting confused as to what exactly AMD has put in the market. The table on the website is a pain to decipher because AMD has to put the long official name in the columns which makes them abnormally shaped.

Over the past week, a lot of press and pointed out that AMD's decision to not support Zen 3 on its 400-series chipsets was rather frustrating, especially when popular mid-range chips like the Ryzen 3 3300X did not have a corresponding B550 series motherboard in the market, forcing new builders to buy B450 instead. The lack of Zen 3 support was basically cutting off the upgrade path to future generations, despite AMD stating it would do its best to support it. This limitation is actually a physical BIOS one, and a technical one.

Today's news is that AMD has heard us, and heard YOU. They are going to go through the motions in order to support Zen 3 on 400 series chipsets. It will require collaboration with the motherboard vendor, and there may be some additional hoops to jump through, but AMD has heard and they will help enable partners and end-users for support. Full details in the video.

0:00 The History
0:52 The Limitations
2:37 The Outrage
3:09 The News!
4:09 The Caveats
5:39 The Recommendation
6:18 The Chipset Table
7:40 Flash / Boot Kits?
8:40 The Press Release
15:08 My Recommendation

Welcome to the TechTechPotato (c) Dr. Ian Cutress
Ramblings about things related to Technology from an analyst

Support the channel by hitting that subscribe button, or using our Amazon affiliate link: [link]

#techtechpotato #AM4 #AMD

[SuperMatthew]

I think that all of this could have been resolved if AMD had informed the board partners beforehand, and said that Zen 3 wouldn't fit on a 16MB BIOS ROM. If that was the case, new motherboard should have been produced with two 16MB ROMs (maybe it's cheaper than a single 32MB ROM), and a jumper that would allow the user to change between Ryzen 1000/2000/Athlon BIOS and Ryzen 3000/4000/Others BIOS. That would have made a lot more sense, and woudn't create this complicated mess with BIOS flashing and motherboard compatibility charts.

[VigneshBalasubramaniam]

Interesting... I didn't think that AMD would change their stance on this, and I understood why they made their original decision. Although it is brilliant for me, I was planning on upgrading my 2700X to the 4000 series flagship on my X470 Taichi Ultimate. Its one of the few affordable boards that has 10GbE (and doesn't sacrifice the aesthetic of a plastic cover over the rear IO like the ASRock X570 Creator).

[signfang]

Wow didn't know Dr. Ian Cutress himself was doing Youtube. Subscribed! Love your articles on Anandtech.

[yanniskouretas]

Oh! a Commodore t-shirt ! legendary ....

[williammurphy]

Nice & concise breakdown! I appreciate it!
Thanks & keep up the good work! Subbed!

[RyTrapp]

For every action, there is a reaction... now I wonder how many nom-enthusiast general consumers are going to end up frustrated by the inevitable clusterfuck that BIOS updates are going to be. We know that the reality is that very few people will be using multiple generation old boards with multiple generation newer CPUs(people say that they're going to do a LOT of shit on the internet... ...and rarely actually do any of it); we already saw this play out with 300-series boards and Ryzen 3000, I haven't actually even heard of anyone doing this since 3rd gen dropped, certainly there's a few out there though. None the less, a LOT of headache for a small vocal minority(last I checked HardwareUnboxed's poll, it had more than 70k votes with only 25% of participants actually having a problem with the lack of 4000-400 compatibility - and this is an enthusiast crowd).
I don't see this being the best decision for their platforms or the general consumer, it's just more opportunity for the non-enthusiast to have another rocky experience with another AMD product; which has been one of their biggest problems in regards to winning back the mainstream for a long time. Linus was even talking about this recently on the WAN show, how much more frustrating just updating a BIOS on a B450 board is compared to Intel.

Of course, no one is concerned about any of that though, it's all only about how this affects "me" obviously. Which is exactly why AMD needs to learn when to say 'no' to the vocal minority, for AMD's and the mainstream consumer's sake.

[duanecook]

I've been holding off commenting on a bunch of videos on this topic, shouting into the wind and all.
I have designed embedded single board computers, and has compiled and flashed coreboot for x86.
With that experience, It is amazing to me that most x86 motherboard don't have a boot-loader recovery method.

All motherboards should have a USB device port that allows it to be plugged into a working computer to write to the UEFI/BIOS flash chip with or without a CPU. This is something has existed on most embedded CPUs for a while.
(I'm not talking about BIOS flashback with USB host port.)

With the possibility to always change the flash, it could be common practice to have modular UEFI/BIOS images built for your specific motherboard and CPU combination. It wouldn't be as convenient as, buy motherboard, put CPU in. But this would allow for true support for all CPUs within the same physical socket.

Back in the day people had to set jumpers on the motherboard, on the hard drives, on the floppy drives. The idea that everything should be plug and play if they happen to have matching connectors has caused this problem.

[tommihommi]

I assume MSI also put a lot of pressure on AMD to backtrack.

[nukedathlonman]

It's a good move for AMD and 400 series owners, and I was willing to accept which ever way AMD went on the subject. So yes, I'm happy for people on 400 series mobo's.


I do have what will seem like an odd question...


I'll do a walk down down memory lane so people know where the question that formed in the back of my mind a while back came from. I jumped into a single core Athlon64 system back in 2004 (an original A64 3200+ "Winchester"). It was based on a Gigabyte nForce4 mobo (first gen PCIe), 1gb (2x 512mb) dual channel memory, and a Gefore 6600 GT... It wasn't too shabby for the time - a solid workhorse. It was around 2007 (maybe 2008) I decided to upgrade it to 4gb RAM and a used Dual Core Opteron 180 (90nm "Denmark") to extend the life of the system for a few more years of service. I didn't go with a 4800+ or the highly prized FX-60 ("Toledo") s939 chips because I found it cheaper and much easier to find the Opty for a reasonable price at the time (FX-60 was highly sought after and commended pretty high prices, even used, while l had a hard time finding a s939 4800+ (AM2 chips where easy to come by). And really, there wasn't much of a difference between the A64 4800+ and Opty 180 (aside from the additional enterprise certifications for the Opty to run 24-7 in a server, I'm pretty sure they were identical chips given they did perform identically in a desktop environment).


Where is this going: I didn't originally flash the BIOS on that system - just plugged the new (used) CPU in and booted it up. The system didn't know what the chip was nor did Windows know what the chip was (although Linux had no troubles identifying it and ran perfectly). So I realized I needed to flash the BIOS to make everything work properly (at least in Windows). I was able to do the flash with the wrong CPU installed with out issue, and yes that fixed everything up correctly.


So, despite the fear of opening pandora's box, I do have to wonder: is it really not possible to have basic operational capability any more when an unrecognized (newer) chip is installed? No one really worry about answering the question - it was only a question that came up in the back of my mind when I upgraded from a 1600 to a 3600 on my current gaming system. If anyone does answer the question, then let's keep it positive and constructive. :-)

[skoopsro]

Too bad they didn't do socketed bios chips across the boards lineup. That would have been the easiest solution for this mess. I have a friend with a x470 mobo that has a socketed bios. I would been so pissed if I was him and there was such an easy solution and amd stuck to their original plan.

[kayaritvards]

Hello Ian! Thanks for another amazing commentary. Is it possible on your side to enable automated subtitles? I have mild hearing loss so it'd be wonderful. Thanks. :)

[DuringDark]

AMD has pushed itself back onto the RGB pedestal many have put them on, good to see they still want to work for consumer goodwill!
BTW, what's the hammer in the background for? Top left

[The_Nihl]

For me, whole this situation blew out of proportion from a jump on conclusion. One slide, which in fact still says that 300 series mobo are not compatible with 3rd gen Ryzen processors, and hallock commentary on Reddit, and outrage was enormous... Im not even sure if AMD actually planned to allow beta AGESA for capable boards, we would just hear about it on 4000 series CPU annoucement or something. Im still not sure if its good idea, I can already see a second hand market flooded with B450 mobos with 4th gen bios....

[squelchedotter]

I think it's pretty clear from the context that they're planning to use a new socket for Zen4.

[gdrriley]

At this point I'm pretty sure my 2700x will get upgraded to a 3900x or a 3950x unless my midrange asrock board gets a bios update.
I was basically forced to B450 because there are so few x570 mATX boards

[RepsUp]

Hope they bring support for high end X370 motherboards as well...

[Space_Reptile]

Thats a Sick Commodore shirt, i approve !

[namyun]

After seeing the "release" of the new 4000-series desktop APUs and thinking about what that compatibility chart means, I have come to the conclusion that AMD does not intend to make the 4000-series desktop APUs a retail product at all, like all the "GE" series desktop APUs before it. The "(OEM Only)" tag on AMD's website is the final nail in the coffin.

For those considering buying a 4000-series desktop APU from the grey-market channels, it seems you'll need an X570 or B550 chipset motherboard. None of the B450 AGESA 1.0.0.6 BIOS updates are advertising support for the new APUs.

[Blzut]

In your table, was excavator support on 300 series intentionally marked as unsupported? Seems like the box for 300 and 400 got flipped there since otherwise the flash kits wouldn't have worked. (Although I'm sure they exist, it seems like most 400 series boards don't officially list support for Excavator.)

[WeirdSeagul]

i reckon some b350/x370 will be support by beta aswell since a couple of the b450/x470 are just refreshes of almost the same board