AMD to address Ryzen 3000 boost clock issues with BIOS update

onetheycallEric

Posts: 225   +47
Staff
A hot potato: In response to the growing chorus of Ryzen 3000 customers who aren't seeing advertised boost speeds with the latest Ryzen processors, AMD will look to address the problem by pushing out a BIOS update to motherboard partners. Confusion over boost block behavior and sporadic availability for certain SKUs have been something of a blemish on AMD's otherwise impressive Ryzen 3000 launch. Here's hoping that AMD can put one of those issues to rest, sooner rather than later.

While Ryzen 3000 has been a tremendous success for AMD, that success hasn't come without a bit of controversy. Shortly after launch, there's been growing discussion surrounding the boost clock behavior of Ryzen 3000 series.

Early reviews for the chips demonstrated that the processors weren't hitting their specified boost speeds. What's more, boost speeds were limited to only one core, which AMD later confirmed. AMD went on to update its Ryzen product pages to clarify that point and attempt to mitigate further confusion. However, the crux of the issue seems to be that users are reporting that their chips aren't hitting advertised boost frequencies at all -- with any core.

That's an important detail that overclocker and YouTuber Der8auer emphasized when he commissioned a survey that some 2,700 users responded to. The survey found that only 5.6% of users with the 12-core Ryzen 9 3900X were hitting advertised maximum boost clocks. The Ryzen 5 3600 fared better, with nearly half (49.8%) of users reporting the processor was achieving maximum boost speeds. Still, the numbers don't exactly paint a pretty picture.

AMD has acknowledged the issue via Twitter, saying it has identified a firmware-level problem it aims to rectify with a future BIOS update.

“AMD is pleased with the strong momentum of 3rd Gen AMD Ryzen™ processors in the PC enthusiast and gaming communities. We closely monitor community feedback on our products and understand that some 3rd Gen AMD Ryzen users are reporting boost clock speeds below the expected processor boost frequency. While processor boost frequency is dependent on many variables including workload, system design, and cooling solution, we have closely reviewed the feedback from our customers and have identified an issue in our firmware that reduces boost frequency in some situations. We are in the process of preparing a BIOS update for our motherboard partners that addresses that issue and includes additional boost performance optimizations. We will provide an update on September 10 to the community regarding the availability of the BIOS.”

We'll have to wait until next week to see what AMD has in store for a fix, but hopefully it'll bring about better -- and more consistent -- performance in respect to Ryzen 3000 boost clocks.

Permalink to story.

 
The crappy clocks of this CPU's its the main reason why I didn't move from my 2700X, I will wait till they improve the node and hopefully they can hit at least 4.5Ghz all core with Ryzen 4000 :)
 
The crappy clocks of this CPU's its the main reason why I didn't move from my 2700X, I will wait till they improve the node and hopefully they can hit at least 4.5Ghz all core with Ryzen 4000 :)
The clocks are not why you would want to upgrade, the overall performance is the main reason. Most boards that don't reach the advertised clocks are just 25-50MHz below anyway with few exceptions.

Although I doubt that you need to upgrade your 2700x even if they improve the clock speeds. You are better off skipping this gen and waiting for the 4000 series.
 
We will provide an update on September 10 to the community regarding the availability of the BIOS.

And if it doesn't help widely, I probably know AMD's response: if you are not happy with our CPU's boost clocks, we'll bring you new one through our RMA (like we did it with seg fault under Linux).

Anyway, it's good news for PC hardware related sites ;)
 
The crappy clocks of this CPU's its the main reason why I didn't move from my 2700X, I will wait till they improve the node and hopefully they can hit at least 4.5Ghz all core with Ryzen 4000 :)

That is the exact CPU I upgraded from to my current 3700X (only because 3900Xs are nowhere to be found lol). The clock is only slightly higher then the 2700X yet the speed increase was immediately noticeable for me. The last time an upgrade actually gave me a noticeable improvement like this was going from a Phenom II x4 to a 3570K. I didn't get that going from a 3570K to a 5820K. In fact no CPU upgrade since then has made my system feel more responsive (save for my 2700X when I was heavily multi-threading but that's rather niche).

A good chunk of applications I click on now just load, instantly.
 
I think the insane temperature of 3000 series may have something to do with their poor boost clocks. They shoot to 70-80C instantly under any significant load, it reminds me of the GTX 480s. I think it may be from all the cores being on a single tiny chiplet, and the lack of surface area to transfer heat along wiht the asymetrical design not working well with coolers that were built around the old monolithic design.

Keeping temps low is key for AMD's boost, but few coolers can do it. Even coolers with a single lapped copper plate struggle. I'll bet this is what AMD improves on with the 4000 series to get clock rates up.
 
No one cares but here goes. I bought a 8600k few yrs ago. It would not boot until I disabled multi thread. I contacted Intel and they would not replace the chip. Pissed me off so much I will never buy intel again. That is short version so I don't bore everyone.
 
Core with Ryzen 4000 will be nothing short of 5.2GHz, its possible with the shrinking die.
Neah, 5.2GHz will not be possible on 7nm+ no matter how much we want it to happen. The more tightly everything is packed in, the harder it will be to get high clock speeds.

At most we'll see single core boost clocks of 4.9GHz for the best binned CPUs (about a 5% increase) and that's fine because Zen3 should also improve IPC a bit (I'm hoping that it will be better than 5%). AMD just needs to make Zen3 around 10% or 15% better than Zen2 (overall performance IPC+clocks+other optimisations) and they definitely have some key areas they could improve to get good results.

I wonder if 5nm will drop boost clocks back down to 4.2GHz again for AMD and Zen4 (kinda like how it happened for Intel with 10nm).
 
Last edited:
I think the insane temperature of 3000 series may have something to do with their poor boost clocks. They shoot to 70-80C instantly under any significant load, it reminds me of the GTX 480s. I think it may be from all the cores being on a single tiny chiplet, and the lack of surface area to transfer heat along wiht the asymetrical design not working well with coolers that were built around the old monolithic design.

Keeping temps low is key for AMD's boost, but few coolers can do it. Even coolers with a single lapped copper plate struggle. I'll bet this is what AMD improves on with the 4000 series to get clock rates up.
Something worth checking is the thickness of the CPU silicon as der8auer found that the extra thickness in the 9900K was one of the main reasons why that CPU was getting so hot.
Another thought I have is your note that the CPU uses an assymetrical design - with some parts pushed out towards the edge, we'd have to be a lot more careful with how the thermal paste is being spread out.
 
Last edited:
We will provide an update on September 10 to the community regarding the availability of the BIOS.

And if it doesn't help widely, I probably know AMD's response: if you are not happy with our CPU's boost clocks, we'll bring you new one through our RMA (like we did it with seg fault under Linux).

Anyway, it's good news for PC hardware related sites ;)

It looks like the difference between the 2700X between 3700X (besides 40w less power requirement) is only about 10-15% performance increase. For my situation, it's not worth the upgrade.
 
From 4375 to 4392 I gained 17 mhz. it wasnt worth the effort really but its closer to the stated 4400 mhz. not at all worried about 8 mhz
 
Back