• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

New AGESA update to address CPU's not boosting to their advertised max

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Zerileous

Senior Member
Joined
Jun 21, 2002
It appears the samples used by GN (at least on their Gigabyte board) are now reaching their max boost speed with single core loads. Performance seems to improve 1-2% for their 3900X.


I wonder if others will have similar results.
 
How much of a real world difference does that even make? I get it though. They made a claim and fell short of that claim so there this fix to meet that claim, even if its only 1-2%.

Unboxed Hardware showed a lot of it also depends on your hardware. He also talked about his chip able to hit the boost clocks with certain boards. He also had a fan who has a chip that couldn't hit the boost clocks. The fan came and visited him, used his hardware and hit the boost clocks.

Also goes over the new bios release.
 
Is this out? I know beta biases were there.

Though it really doesnt effect performance, the fact that a lot (most?) We not able to hit the rated boost clocks on the box/specs page even though the "nominal" conditions were met was a bit disconcerting.
 
Before launch AMD were discussing enabling it for older chipsets if mobos could be validated. In the end they decided against ANY older chipset having that functionality, but the code was present to enable it. Some mobo manufacturers apparently exposed it to users. AMD didn't like that so they're going to hard kill it.
 
Apparently the power requirements were such that AMD didn't trust the previous generations of boards in general, and picking a handful (if that many) that were acceptable was deemed a bad idea on several fronts. Probably the right call, honestly. The socket works for the Gen 2 chips (mostly), and PCIe 4.0 NVMEs and GPUs aren't exactly tearing up the shelves by flying off in quantity. Yet. :D
 
ASRock released BIOS series with ABBA AGESA. I will check it on the X570 PG TB3/ITX soon but tbh I have mixed feelings about this motherboard and I didn't know I will ever say that but AORUS X570 I Pro WiFi is simply better. I just see that all reviewers skipped some things worth to mention and the PG TB3/ITX receives top awards anyway.
 
ASRock released BIOS series with ABBA AGESA. I will check it on the X570 PG TB3/ITX soon but tbh I have mixed feelings about this motherboard and I didn't know I will ever say that but AORUS X570 I Pro WiFi is simply better. I just see that all reviewers skipped some things worth to mention and the PG TB3/ITX receives top awards anyway.
Do tell... :)
 
Joe, I know that you reviewed this motherboard (also AORUS X570 I Pro WiFi to which I was comparing some things) but I wasn't talking about your review, or maybe mostly about others.
What I meant are things like:
- ASRock still hasn't fixed mobo freezing when you hold Del key too early trying to enter BIOS
- The motherboard has DDR4-4533+ support but in real I couldn't run memory past that while Extreme4 (rated at 4666) or Gigabyte (rated at 4400) runs at 4800+
- Each failed memory OC ends on clearing CMOS as the board can't recover automatically while Extreme4 could
- SB fan is noisy at standard settings and at 5.5k RPM chipset has 65°C+
- The motherboard has only one M.2 and it's on the back so most M.2 PCIe 4 drives (most have heatsinks) can't be installed in smaller cases or without a hole under the motherboard
- Support for coolers is really limited. I like the fact it supports Intel coolers as they're more popular but out of 6 coolers that I have, only one could be installed and it's Noctua D9L with 1 fan as with 2 it won't fit too. Even AIO couldn't be installed or it was blocking one memory slot. The main issue are too tall VRM/chipset heatsinks which are blocking every larger cooler. Even Scythe Big Shiruken 3 designed for ITX mobos couldn't be installed as chipset heatsink is too tall.

What can I say ... most of that stuff is not included in most reviews or reviewer mentions about it and later says that everything is perfect ;)


Btw. I had no problems to reach max boost clock on ASRock and Gigabyte motherboards that I have. At least with Ryzen 3600 and 3700X all was fine before ABBA AGESA.
 
Last edited:
Interesting...

* I never ran into the delete key issue. I only press the delete button when the keyboard lights up and that worked for me.
* Surprised on the memory... wonder if it has anything to do with all the crap between the socket and DIMMs...
* It is the first chipset fan I heard too...
* Correct... though many ITX boards are like this (M.2 on the back, no heatsink possible). Or, its stacked on top of the PCH heatsink... bleh either way.
* I noticed the issue with space as well. The DRAM ran into the H150iI have.

I mentioned a few of these things in my review at least! :)


EDIT: Not one of the boards I used allowed my 3700X to reach max boost. It was always a meager .25/.5 multiplier low. Due to a death in the family last week, I haven't had time to test the new BIOS'.
 
Last edited:
Thanks for linking that article, PolRoger.

So I'm feeling much better about being able to get my Ryzen 2700 to 4050 mhz. on all cores with 1.36 vcore. That gives me a 20% increase on Cinebench R20 scores over leaving everything on default and for me ,"Keep it simple, stupid!" works. This article also helped me understand why overclocking Zen 2 chips is yielding much improvement over stock.
 
Great article, it's conclusion actually landed on a question that had been floating around in my mind since Der8auer's instructions which indicated adjusting the polling frequency of HWiNFO64 for his test.

But simply put, unless a user is polling this quick, the user will not see the momentary peaks in turbo frequency if they are on the boundary of supporting it. The downside of this is that polling this quick puts an artificial load on the system, and means any concurrent running benchmark will be inadequate.
...
It all leads to a question – if a core hits a turbo frequency but you are unable to detect it, does that count?
 
Back