"ASRock states that its new firmware has been built to decrease booting times on its motherboards."
It's booting too fast aaaaaaaaaahhh
"Another thing ASRock has announced is that all of its X670E and X670 models will support BIOS Flashback (a feature now native to the AM5 platform)"
I've done that on an MSI AM4 mobo. What does being "native to AM5" change about it? Isn't it still the choice of the manufacturer to support it or not on a particular motherboard?
I see that the article has been fixed. Pity, making flashback native to AM5 would have been nice. That would have implied that all AM5 boards come with some ARM CPU and RAM to get this functionality working, and that AMD has mandated that this functionality is available. Having flashback for all boards (even low end ones) would be great.
I think the bios flashback required a chip in the mobo to activate the feature and not all motherboard supported that. Native means perhaps that is embedded in the chipset already?
The best bios recovery technique isn’t to have a PXE for the BIOS using a separate SoC as some implementations have, but simply to have a dual bios, something that has existed for decades. Using an ARM CPU, RAM and NAND or external USB flash drive to boot a PXE bios recovery is a solution looking for a problem, and traditionally delegated to the server space for management reasons (remote recovery) which isn’t being implemented in consumer boards making the SoC approach moot and overly complex for what should be a simple implementation.
The feature allows BIOS updates to support a new CPU, something dual BIOS cannot help with. It also should recover boards bricked by failed BIOS update.
I'm curious to see their B670E mITX boards. I'm hoping for 2 x pcie5 x4 nvme m.2 slots for storage, 1 pcie5 x16 slot for a discrete video card and decent built-in audio, ethernet and wifi.
So when they say, "Faster" do they mean they disabled some of the controllers like for ethernet, or that they've actually optimized something, like by doing multi-threading?
The key thing missing from this article is what the issue actually was. It was leaked last weekend that the ASRock AM5 boards could require up to 4 minutes for first boot or after a Clear CMOS to initialize RAM. (that was if all 4 slots were occupied). Apparently there was a sticker on the RAM slots on the leaked images that indicated this. Even 2 sticks of RAM would still take 90 seconds-2 minutes I believe.
That is what seems to be the issue ASRock is fixing with the new firmware.
That's pretty short, and my guess is that the interpretation that it refers to RAM training speed is correct. What ASRock is saying is: we have a new BIOS and you can flash it before assembling the system by using flashback.
I really wish they'd just publish specifications to allow the open-source community, such as Coreboot, to develop firmware for these boards. I really miss the days of being able to use Coreboot, and I frankly don't understand why this even needs to be controversial.
To begin with, I didn't call for them making the current BIOSes open source. I realize there may be a ton of reasons why that can be hard, including using code licensed from third parties that they can't easily free up. What I called for was specifications / interface documentation.
That being said, if open-sourcing your crypto implementation makes it insecure, that's a good sign that you're way overdue to throw it away and start over. And if even releasing the specifications of your hardware ruins that security, I wouldn't even let it touch anything actually sensitive.
I'd love to care about something ASRock is doing with motherboards, but I switched to them after using exclusively ASUS boards since around 1999, and everything seemed great until every single one of them died within 3 years of building a PC. I now dread doing favors for people with building PCs after every person I helped with a PC while using an ASRock board has had their motherboard go bad. I loathe those calls from people asking why they're computer isn't working, then trying to eliminate everything relatively easy to replace to find out the MB is the culprit. Endless RAM tests, swapping out PSUs, CPUs, etc. Just want them to create a motherboard that doesn't fail after a few years. Will switch to MSI or go back to ASUS.
I wish for BIOS-es to have "adult mode" without stupid graphics all around. (Beta bioses without stupid graphics for older AMD chipset should be the new norm).
I going to wait till the X670E motherboards boot time is fixed back to normal. Who want to wait a minute to have it finally boot up! Makes buying a 7000 series a joke!
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
26 Comments
Back to Article
nandnandnand - Friday, September 9, 2022 - link
"ASRock states that its new firmware has been built to decrease booting times on its motherboards."It's booting too fast aaaaaaaaaahhh
"Another thing ASRock has announced is that all of its X670E and X670 models will support BIOS Flashback (a feature now native to the AM5 platform)"
I've done that on an MSI AM4 mobo. What does being "native to AM5" change about it? Isn't it still the choice of the manufacturer to support it or not on a particular motherboard?
ET - Sunday, September 11, 2022 - link
I see that the article has been fixed. Pity, making flashback native to AM5 would have been nice. That would have implied that all AM5 boards come with some ARM CPU and RAM to get this functionality working, and that AMD has mandated that this functionality is available. Having flashback for all boards (even low end ones) would be great.tamalero - Sunday, September 11, 2022 - link
I think the bios flashback required a chip in the mobo to activate the feature and not all motherboard supported that. Native means perhaps that is embedded in the chipset already?Samus - Monday, September 12, 2022 - link
The best bios recovery technique isn’t to have a PXE for the BIOS using a separate SoC as some implementations have, but simply to have a dual bios, something that has existed for decades. Using an ARM CPU, RAM and NAND or external USB flash drive to boot a PXE bios recovery is a solution looking for a problem, and traditionally delegated to the server space for management reasons (remote recovery) which isn’t being implemented in consumer boards making the SoC approach moot and overly complex for what should be a simple implementation.RobATiOyP - Tuesday, September 13, 2022 - link
The feature allows BIOS updates to support a new CPU, something dual BIOS cannot help with. It also should recover boards bricked by failed BIOS update.Threska - Tuesday, September 13, 2022 - link
Considering the RMA process at some companies, one less avenue of failure is good.HardwareDufus - Friday, September 9, 2022 - link
I'm curious to see their B670E mITX boards. I'm hoping for 2 x pcie5 x4 nvme m.2 slots for storage, 1 pcie5 x16 slot for a discrete video card and decent built-in audio, ethernet and wifi.ballsystemlord - Friday, September 9, 2022 - link
So when they say, "Faster" do they mean they disabled some of the controllers like for ethernet, or that they've actually optimized something, like by doing multi-threading?Khanan - Saturday, September 10, 2022 - link
Faster in this case means optimized bios rom so it can initialize Agesa or itself faster.ballsystemlord - Monday, September 12, 2022 - link
That did not really answer the question.Khanan - Thursday, September 15, 2022 - link
It does. And the direct answer is no. Ofc there is nothing disabled, just optimizations for stuff to initialize faster.Devo2007 - Saturday, September 10, 2022 - link
The key thing missing from this article is what the issue actually was. It was leaked last weekend that the ASRock AM5 boards could require up to 4 minutes for first boot or after a Clear CMOS to initialize RAM. (that was if all 4 slots were occupied). Apparently there was a sticker on the RAM slots on the leaked images that indicated this. Even 2 sticks of RAM would still take 90 seconds-2 minutes I believe.That is what seems to be the issue ASRock is fixing with the new firmware.
Hul8 - Saturday, September 10, 2022 - link
I doubt this article is about DDR5 training times.ASRock is stating "faster POST" which implies on every boot - not just when you've cleared the CMOS or installed a new CPU.
ET - Sunday, September 11, 2022 - link
Thankfully at least one site (Hot Hardware) cared to provide a link to the ASRock press release (https://www.asrock.com/news/index.us.asp?iD=4966).That's pretty short, and my guess is that the interpretation that it refers to RAM training speed is correct. What ASRock is saying is: we have a new BIOS and you can flash it before assembling the system by using flashback.
Dolda2000 - Sunday, September 11, 2022 - link
I really wish they'd just publish specifications to allow the open-source community, such as Coreboot, to develop firmware for these boards. I really miss the days of being able to use Coreboot, and I frankly don't understand why this even needs to be controversial.ballsystemlord - Monday, September 12, 2022 - link
I second the motion.Khanan - Thursday, September 15, 2022 - link
I disagree, it is not open source because of safety concerns and TPM as far as I know.ballsystemlord - Friday, September 16, 2022 - link
So telling someone about how secure your product is makes it insecure?Even if that were true, than that would mean that we should all be designing our own TPMs because MS knowing makes it insecure.
Also, the OP, Dolda2000, advocated for open specs, not opensourcing the whole thing.
Dolda2000 - Saturday, September 17, 2022 - link
To begin with, I didn't call for them making the current BIOSes open source. I realize there may be a ton of reasons why that can be hard, including using code licensed from third parties that they can't easily free up. What I called for was specifications / interface documentation.That being said, if open-sourcing your crypto implementation makes it insecure, that's a good sign that you're way overdue to throw it away and start over. And if even releasing the specifications of your hardware ruins that security, I wouldn't even let it touch anything actually sensitive.
BioHazardous - Sunday, September 11, 2022 - link
I'd love to care about something ASRock is doing with motherboards, but I switched to them after using exclusively ASUS boards since around 1999, and everything seemed great until every single one of them died within 3 years of building a PC. I now dread doing favors for people with building PCs after every person I helped with a PC while using an ASRock board has had their motherboard go bad. I loathe those calls from people asking why they're computer isn't working, then trying to eliminate everything relatively easy to replace to find out the MB is the culprit. Endless RAM tests, swapping out PSUs, CPUs, etc. Just want them to create a motherboard that doesn't fail after a few years. Will switch to MSI or go back to ASUS.PeachNCream - Monday, September 12, 2022 - link
Protip: Stop being a tech support serf for other people.Threska - Monday, September 12, 2022 - link
That would be Reddits job.Khanan - Thursday, September 15, 2022 - link
Your comment is just anectodal, I don’t think quality of Asrock is that bad.Jake7 - Monday, September 12, 2022 - link
They should spend some cycles to support fwupd.orgXenon2 - Friday, October 7, 2022 - link
I wish for BIOS-es to have "adult mode" without stupid graphics all around. (Beta bioses without stupid graphics for older AMD chipset should be the new norm).Harry_Wild - Friday, October 14, 2022 - link
I going to wait till the X670E motherboards boot time is fixed back to normal. Who want to wait a minute to have it finally boot up! Makes buying a 7000 series a joke!