AMD Raven Ridge 'Boot Kit' Includes Unnamed Bristol Ridge Chip, Heatsink

Status
Not open for further replies.

InvalidError

Titan
Moderator
I wish AMD/Intel would simply implement their BIOS in such a way that they can run "unknown" chips in some sort of safe-mode (ex.: all non-essential features disabled, voltage and clocks stuck at the lowest power working state available) instead of being out of luck without a compatible chip available for BIOS update.
 

razor512

Distinguished
Jun 16, 2007
2,130
68
19,890
What ever happened to the option to update the bios without the CPU being installed? I believe Asus and gigabyte implemented this on some of their boards in the past.

Beyond that, a safe mode would be useful, most CPUs can easily run at 1 volt, Why not just run the CPU at something like 1V at 500MHz if it is unknown, and then update the bios?

Or better yet, why not have the bios chip in a socket where it can be removed and replaced? Old motherboards did it, and it would be cheaper than AMDs current solution of mailing out a CPU and heatsink, and allowing the user to keep the heatsink, buut return the CPU (because the weight of the heatsink would make return shipping more expensive than the value of the heatsink).

If it was an industry standard to make the chip socketed, then AMD could literally just have a pile of pre-flashed bios ICs, and then simply mail them out to users when they select their motherboard and provide proof of purchase.

Those smaller flash chips cost less than 10 cents each, and can be mailed out in a tiny padded envelope.

 

jchambers2586

Distinguished
Oct 21, 2009
284
0
18,790
asus supports it. others don't I think its a patent thing. I talked to gigabyte they said I would need to send the board back to them so they could update the bios.
 

InvalidError

Titan
Moderator

Key word: 'some', which means everyone else needs a supported CPU. They do this by having a micro-controller hijack a USB port at boot and side-flash the BIOS from a USB key. This means needing said micro-controller, some support logic to re-route USB signals and possibly some logic to isolate the SPI flash from the chipset during that process if necessary.
 

alextheblue

Distinguished
If this was just a clock bump new revision kind of new chip it would probably work. In fact that has happened in the past, recent past even. I strongly suspect there are occasionally serious technical barriers to supporting unknown future chips on a current BIOS (sitting on a shelf), since this has happened to both AMD and Intel from time to time. If it was that easy, their engineers would probably be able to manage it.

Semi-related:

So AMD apparently ships these new APUs with documentation showing an example label to look for, so you know that your motherboard is drop-in compatible with these chips. It's a "AMD Ryzen Desktop 2000 Ready" label. But I had already seen listings on Newegg where some AM4 boards were claimed to be Ryzen Desktop 2000 Ready. I set a couple of models aside for later viewing... a couple days later I started seeing articles about AMD's "Boot Kit" and when I returned to Newegg all traces of that Ryzen 2000 Ready tag are gone. Etailers apparently were experiencing a lot of "my new build won't POST" situations and realized their "2000 ready" boards weren't actually 2000 ready in the intended sense. Whoops. Don't know if that was a miscommunication with the motherboard vendors or what.


Asus has gotten really stingy with their USB Flashback. For AM4 I've only seen it on a couple of their highest end ATX models. Annoying. Personally I wish AMD would integrate this directly into their chipsets (at least B and X chipsets).
 

spdragoo

Expert
Ambassador


Probably add a significant cost to the motherboard. Depending on the original cost of the motherboard, you could even be looking at doubling the cost of a budget board, for a chip that only gets used if and when you update the BIOS.

 

msroadkill612

Distinguished
Jan 31, 2009
202
29
18,710
Whatever, AMD is being jolly British about fixing the problem on some partner's moboS, and they should get kudos for that. Its a solution which focuses on the convenience to the customer, not the cost to them.
 

msroadkill612

Distinguished
Jan 31, 2009
202
29
18,710
PS re my prev point:

e.g. returning the mobo wouldnt work for me. What a hassle if its installed? Yet thats all Gigabyte would (or could really) offer, and Intel ...? Oh you are funny.

Customer empathy is not their strong suite.
 


At one time, not only was the CPU socketted, the FPU was too, along with discreet RAM chips. Then came cost saving methods and an improvement in reliability (no more worries of socket creep) by soldering in everything that wasn't optional. eventually the FPU socket went away completely (with the birth of the Pentium and the demise of the 486 where an integrated FPU became common.) ROM chips eventually got replaced with FLASH which could be updated without any costs outside of time (time includes update time and developer time) and a small amount of storage space.

tl;dr: The standard moved away from socketted chips as much as possible, I seriously doubt that they will ever return.
 

KidHorn

Distinguished
Oct 8, 2009
269
0
18,790
No way am I buying a MB and send it to the manufacturer for a BIOS update. I'll just wait until new MB inventory comes in with the latest BIOS.

I think AMD should have offered two packages. One with a crappy low end AM4 chip for new systems and one without for upgrades. Maybe price them $10 apart.
 

InvalidError

Titan
Moderator

Hm, no. You can get USB-capable micro-controllers for less than $2. Basic Arduino boards cost as little as ~$5 shipped and are already over-powered for flash-rescue purposes. In bulk, we're talking maybe $2 to build the rescue programmer directly on-board. If you integrate the micro-controller directly into the chipset as btmedic likely meant, the cost goes down to something like 0.1 square millimeter (~1.5M transistors) on the chipset die.

The biggest obstacle to putting the rescue controller on-chip is the need to use mask-programmable memory due to eeprom/flash memory not being compatible with high speed logic fabrication processes, which means hard-wired one-size-fits-all code that can't be tailored to individual motherboards.
 

ghettogamer

Prominent
Mar 3, 2017
92
0
660
still a great effort in AMD's part the richer blue team never even bothered to reach out when they put buyers in the same situation when they "refreshed" their sockets with new cpu's in the past
 

spagafus

Prominent
Jul 6, 2017
6
0
510
Why bother with cooler and paste if it's only meant to be used to POST and update the BIOS?

Edit: To answer my own question - the system may not POST without a CPU fan unless that happens to be disabled in the BIOS, which is unlikely for a new MB. I hadn't thought it through for my initial question.
 


If they grabbed an off the shelf part for this update CPU/APU, it will need it or it will shut down before the job is done. I doubt they'd make a custom part for the job. i.e. a seriously underclocked APU (well, clocked fast enough to do the update, but no faster AND with a locked multiplier if a multiplier is even present) with just adequate graphics to access a UEFI BIOS if need be.
 




This actually was a thing for at least many HP Desktops for socket LGA775, the BIOS said something like "A unknown CPU detected" and it ran at 800Mhz, booted into Windows and flashed the BIOS and tada it worked.
 

bigdragon

Distinguished
Oct 19, 2011
1,107
547
20,160
I like that AMD is being proactive about this. Intel wasn't proactive back when I bought into the x79 platform. Thankfully, ASRock was able to overnight a replacement BIOS chip. Boards without a socketed chip had to be sent in for flashing. AMD's flash-at-home solution is much nicer than having to ship the motherboard somewhere.
 

Brian_R170

Honorable
Jun 24, 2014
288
2
10,785
I agree, kudos to AMD for offering it, but I doubt many motherboard buyers will either want or be able to take advantage of it. Most buyers will just return the board to the retailer just because they don't want to take the extra time and effort. Looks like most of the remaining buyers will be forced by AMD to send the board back to the manufacturer for service because the manufacturer offers the service.
 
AMD had no choice but to do this. it's not like intel with Kaby lake CPU's where they had a newer chip set out with it that was guaranteed to work with a newer CPU without a BIOS update. it still doesn't matter how long you wait there is no way to know when mobo manufacturers will ship any with the latest BIOS and know how old the stock is from a warehouse.
 

InvalidError

Titan
Moderator

Some manufacturers include the BIOS version on the box's serial number label, just need to find a (r)etailer who can confirm what version is going to ship or update the BIOS before shipping. Of course, that means potentially having to deal with the BS of the seller not actually checking or updating before shipping.
 

spdragoo

Expert
Ambassador


Not quite true. None of the 100-series motherboards worked "out of the box" (i.e. sans BIOS update) with Kaby Lake. They all required a BIOS update to work with Kaby Lake (even if the vendor applied the BIOS before actually selling it). Only the 200-series motherboards (which were designed from the start to work with Kaby Lake) didn't need a BIOS update to work with Kaby Lake.
 

InvalidError

Titan
Moderator

That is BS: the vendor or manufacturer updating the board to Kaby Lake BIOS version prior to shipping is exactly the same as the end-user using a Skylake chip to update the BIOS himself prior to plopping in a Kaby Lake chip. If a board advertised as updated to be compatible by the manufacturer or vendor doesn't work, it means either that the seller lied about it or he f'd something up.
 

spdragoo

Expert
Ambassador


Isn't that what I said, though? Whether it's the vendor applying the patch or the end-user applying the patch, someone ...

Oh..

Ok, when I said, "vendor", I meant "Newegg", "MicroCenter", "Best Buy", etc., NOT "ASUS", "ASRock", etc. The local MicroCenter tends to apply BIOS patches before selling to the customers so that the customers don't have to do it themselves, but they've also got the spare CPUs sitting around to do it.

We're on the same page, though: if a BIOS update is required for Kaby Lake to work on a board, it doesn't matter who applies the patch (OEM, vendor, end-user), it's still a situation of "Skylake works out-of-the-box, but Kaby Lake requires a BIOS update or it won't work".
 
Status
Not open for further replies.