Showing results for 
Search instead for 
Did you mean: 


Adept III

Ryzen 9 3900X Boost Frequency

To AMD why your new processors never hit advertised boost speeds in single or multi core performance. I have Ryzen 3900x with Asus X570 Prime Pro board running latest bios and AMD chipset drivers. My single core boost is maxed at 4.2Ghz and multi core at 4Ghz. Doesn't matter if PBO enabled or not. There are 1000's of people complaining about this and you have not even acknowledged this as a problem. Your false advertisement about boost speeds is completely unacceptable. I was always Intel customer but this time I took a chance on your company and trust me this is first and last time i will buy anything from you. You hand selected CPUs for reviewers to get the hype going and stuck your paying customers with garbage. I will be filing a RMA claim and will be exchanging this processor until i get one that performs to advertised specifications.

195 Replies

i run CB20 single core benchmark, only 1 core gets utilized and loaded 100%. rest are either sleeping or less than 1% utilization. Intel point view or not, you can not possibly argue that all other things equal higher frequency means higher performance. AMD didn't say that max boost is only available in a lab under LN conditions with none but one cores enabled. So why am i at fault to expect performance that was sold and advertised to me. However, not to continue this asinine conversation i agree with your point and concede that i do not understand how AMD processors should or would work. Therefore, next time I am going to buy from a company that i understand, i.e. Intel, because i don't need a freaking dissertation to explain what they say vs what really is and why its perfectly fine not to get what you paid for.

what you say is wrong, read reviews please and don't be suprised if intel do the same in the future with an auto-oc

you treat the other of fanboys , but it's just because they don't go your way ...


umm no i don't call everyone fanboy who disagrees with me. if we were disagreeing on aesthetics of AMD logo and i called you fanboy because you didn't agree with my perception of it then yes I'd be biased. But instead you disagreeing with me on infallible fact that 4.3Ghz != 4.6Ghz and never will under no amount of explanation or excuses. That's where i would draw a line between disagreement and fanboyism... Also i did switch from Intel to AMD, part of it was hype about new AMD processors but also part of it was supporting competition because that's the only way to get innovation going. Intel been stagnant for a decade sitting on their outdated architecture and this should definitely give them a kick in the pants. But instead, you seem to want to go the other way, when company fails to deliver on its promises you just say meh its ok, Intel would do the same. So in ten years when both companies are selling you a turd with 100 super turdlets that choke running MS-DOS you'd be the person to blame for that.

the way I see it is simple ...

1) ipc gain = good for gaming

2) lots of cores = good for working (eventually for gaming too)

prices for all of that are really ok if you compare (with what company will you compare ? not a lot of choices), so i'm not too regarding i've got a certain tolerance when i know that i will get what i'm searching in the product (AND I KNEW BEFORE BUYING) , and not focusing on precise numbers, i see the gain comparing to what i had ... yeah i know some like to argue for every little thing and can't see what they have, because that's the way society is

so to conclude , you complains because that something that occupîes you time ... lol (don't be mad at me)


At the end of the day, s_vayner‌ is correct in that the Ryzen 9 3900X should reach 4.6GHz (the advertised boost clock) in single threaded workloads.  AMD has made it clear that not all the cores can reach this boost clock, so it isn't possible to push an all-core overclock to 4.6 GHz the way it has been in previous generations or with Intel.

 But one of the cores, should be able to hit this mark.  And it should do so with stock settings and the stock cooler.  How long it will hold that boost can depend on a number of conditions, but it should get there.  The fact that the CPU falls well short is a problem and grounds for an RMA in my opinion.  Hopefully s_vayner‌ has opened a ticket with AMD as ray_m suggested to resolve the issue.

sure ...he can do that , no one telling the opposite

just trying to tell that in my opinion it's not defective or "bad" , maybe it could even improve with a power plan, bios update or something, as i said , it's not because he didn't saw it that it is not able to

hope he will keep us informed about the conclusions


i did and i will.

You seem to be telling him he doesn't have a problem.

This is from the initial post of this thread.

" I have Ryzen 3900x with Asus X570 Prime Pro board running latest bios and AMD chipset drivers. My single core boost is maxed at 4.2mhz and multi core at 4Mhz."

Then you chimed in with...

"the difference now is my 3700x OC itself to 4.3Ghz (ALL CORES ! , 4.4 for 3700x for example IS ON ONE CORE , you should read reviews before buying "

Perhaps you should read his complaint thoroughly.  He already said he isn't getting to 4.6GHz when just a single core is in use.

You follow that up with...

"i don't know why you don't understand : 4,6 is on 1 core 1 thread ..

I have 3700x and never got over 4300mhz ... and that's ok because windows always use more than 1 core so it's a context

maybe if you can deactivate all core except 1 you will get you're 4.6ghz ..."

You don't seem to understand.  He has been talking about single threaded performance the entire time.  And a processor should be able to hit the advertised boost clock single threaded with stock settings on the stock cooler, end of story.

sorry just updated my last answer before reading yours,

in fact i've been lost between facts and what seems a lot of frustration from his part, but i did answer on the single thread thing

when i say mine didn't got over 4300 it's in single core too, the 3700x should hit 4400 ... for my case, i searched and found that on all cores i reach what others reach and what i'm can expect so , because "at the end of the day" it's a thermal and voltage thing, i don't see a reason why that couldn't be possible to reach higher than that on one core... , and as i never use it on one core (i don't even now if possible without deactivate cores ) i'm not too regarding as i think (for me ) the important thing was to know what to expect for the "all core" frequencies when i bought it


Now imagine if yours only got to 4GHz, 400MHz shy, just like OP. I think you'd be a bit more worried you got a dud too.

imagine you get a dolpin for chrismas .. yeah ... facts...


I was trying to suggest perhaps putting yourself in their position to see how you might react, but I don't think you got it, you reply with nonsense. Are you just here to troll?

there is no 400mhz shy , there is "all core" frequencies and "single core" frequencies , but that seem difficult to understand from what i see... so i hope you're happy with your dolphin


there is no 400mhz shy , there is "all core" frequencies and "single core" frequencies , but that seem difficult to understand from what i see... so i hope you're happy with your dolphin


"there is no 400mhz shy"

Oh but there is.  The OP said his CPU boosts to a single core frequency of 4.2 GHz, which is 400 MHz shy of the advertised 4.6 GHz by simple arithmetic and basic definition of words.

"because "at the end of the day" it's a thermal and voltage thing"

At the end of the day a processor should reach the advertised speeds (single and multi) at stock settings with the provided cooler.

"as i think (for me ) the important thing was to know what to expect for the "all core" frequencies when i bought it"

And (for the OP) the important thing is at his CPU works the way AMD said it was supposed to. 

again ... nothing that haven't been already said=>... for the one core boost , bios, agesa, power plan etc etc etc

not mentionning his problem with PBO etc (motherboard , bios ...)

no reason his cpu is different from another and can't achieve 4.6 under certain conditions

even, you can't expect a 12 core cooled on air (i presume it is not watercooled here for sure) do better except in peaks ... just be realistic ... 4.6 (theorical) on one core = 4.1 on all core (maybe 4ghz at some moment under heavy load and poor cooling system) for a 12 core ...

my 3700x = 4300 all core = 4100 if temperature too hot ... that's called "throttling"

can someone help me to explain ? lol

somebody already spoke about the difference of numbers between to different motherboard ... so just be patient , you're only problem is not seeing this 4.6 doesn't mean it's working bad


Just to continue to beat the dead horse here.

"you can't expect a 12 core cooled on air (i presume it is not watercooled here for sure) do better except in peaks ... just be realistic ... 4.6 (theorical) "

You can expect exactly that, because that is what the manufacturer stated, directly on the product specification sheet.  4.6 isn't theoretical, it is the manufacturer's guaranteed boost clock for single core printed directly in the product specifications.  If it doesn't hit it, then the product is "not as advertised". 

"you're only problem is not seeing this 4.6 doesn't mean it's working bad"

It is.  the core performance boost should look like this.

Yet, the OPs sample doesn't ever break 4.2 GHz.  That means on less threaded applications (especially 4T or less) he has lost substantial performance.

"4.6 (theorical) on one core = 4.1 on all core (maybe 4ghz at some moment under heavy load and poor cooling system) for a 12 core"

First of all, we aren't discussing multicore performance, as all AMD promises for that is (3.8 GHz) and the OP was seeing 4 GHz, hey look at that, we hit the spec!  Secondly, what are you basing that statement off of?  Could you produce the relevant data please that supports your assertion about 12-core performance?

you're just repeating what's already been said, you think i didn't see ? doesn't change my response , so i think i won't answer anymore here because we don't advance , just go read reviews and try finding something different ... your graphic doesn't mean anything, it just show what it's mean to be under the right conditions, and here in this case conditions are surely not fulfilled

at least i don't see anything really strange between my 3700x behaviour and this 3900x , at the end , it's surely a good reason for you to rma  , but it will be difficult to find better

in fact you just read word for word the marketing text and expect something the marketing has surely thought about, for the rest , read reviews if you don't want bad suprises .. you surely aren't so naive ... that doesn't mean there is no problem with something in his configuration but i would be really surprised if there is something about the cpu ... again, drivers etc etc .. .blablabla lol

did you notice that you graphic don't comes from amd but techpowerup and techpowerup surely "tested" this in certain conditions ?


If there is nothing wrong, why did ray_m earlier in this thread contact the OP to work on his "issue"?   Clearly, even AMD thinks the behavior is not normal.

The "up to" was altered by AMD after this issue was brought to their attention.. Amd changed their website because of it..

Do you see anything about might hit this boost speed, or might hit this base frequency?  I sure dont.. That is false advertising/false representation..  Amd is lying and misleading consumers..


The ad doesn't say "you will most certainly hit 4.5".

It just says the CPU has a max boost of 4.5, that's the design limit of the CPU. 

You are really whining about 0.1/0.2 MHz? 



max boost.. so it should hit it....   They arent even hitting their base clocks.. but people dont notice that because they are looking at the big end.. The first thing I noticed is my 3800x is not running its base clock either..low base clock.PNG

mmmh seems I was wrong about PBO , i deactivate it this morning , and my cpu still goes up to 4275 mhz

i think i may confound it with the "core boost" option

otherwise, i tried some tweaking and tests too with cinebench

it appears PBO must be "enabled" not "auto" to work properly on my board

and it helped gain 60 points in CB15 (PBO appears to maintain a 4125mhz stable instead of going down to 4050)

if I "disable" PBO idling "seems" better too

i also did a test on the TPU II option ... and then got in windows and upped the core clock to 4.4Ghz => it didn't do what i thought it would do but i saw 2 cores reguraly go up to 4.4Ghz then (the thing is only one is supposed to be , so something was too high and it crashed as i launched a game ... ) regarding this little tests i really think all of this is just a matter of tweaking

another test , i fixed core clock to 4.250 with 1.5V , was stable no problem but I saw some temps go pretty high then (70° core and 86° cpu package) ...

even with my AIO (corsair H115i pro with 2000 rpm fans)

...still a matter of tweaking , options, drivers to my view


Don't set manual 1.5v stay below 1.35v. Auto is okay to see up to 1.5v

yup , it was just for a little test to be sure to be stable as i read lot of people can do 4350 with 1.3-1.4V (the question is : is this voltage on load ?? maybe will have to change the load line calibration then) , when i saw temps it just confirmed me why pbo wasn't maintaining at this freq/voltage for long time


yea, but this was published after thousands started finding the issue.. just amd trying to save face...   This is just another rushed to market product from amd with poor and improper support..

I wonder if a better CPU cooler like one of the top counted dual fan jobs will allow the CPU to run properly.


The listed boost frequency was never all-core so that should come as a surprise. No matter how many times people like Steve from Gamers nexus try to squash that misconception, people seem to still expect that...

What should be expected is single core tasks reach, and maintain, this advertised boost speed, which is what isn't happening. I'm not going to be one that says its the processor or AMD lied, because its neither as the speed will actually be maintained if you set it with a manual OC.

Journeyman III

4mhz is indeed quite slow, it is the speed of my first cpu, big endian based too, the z80a from zilog (the ax register was named A and mov ld), well don't tell me oh everybody understood anyway, 1ghz and 1mhz well....

Adept III

I've filled out a service request myself.  According to CPUZ at the moment, all cores are pegged at 4.4Ghz, idling at 1.48v.  Previously when monitoring with CPUZ idle voltage would dip down int the 0.3v range at times, and boost frequency would jump to 4.5-4.55Ghz. 

Ryzen Master is showing the cores idling and sleeping properly, but still at 1.48v.  And when monitoring with Windows Task manager, the CPU doesn't boost above 4.37Ghz.  Which is pretty similar to what Ryzen Master shows in single thread work loads.


asus just released a update of the ryzen power plan, now idle goes to 0.9V

even , injecting 1.5V with low intensity is not a problem

if you ever did some oc , load line calibration is used to maintain the voltage under load when intensity is needed

if you watch consumption it's about 3-5W even if voltage is 1.5V and/or freq at +4Ghz

under load the voltage decrease with intensity increasing => 1.3V

so these are only analysing tools or drivers that need to be updated

Journeyman III

Ill put it out there as follow.  I've used my Ryzen 9 3900x in my Asus Crosshair VII Hero.  It boosted 4600mhz single core without sn issue.  All core boost was between 4.075 to 4.15.  Unfortunately I ran into issue with my C7H board and had to look for a replacement and settled for a MSI MEG X570 ACE.  This board is great, make no mistake,  I immediately noticed all core boost dropped to 3.95ghz.  I did not check single core boost as Im more of a manual overclocker.  So right now Im running all core 4.36ghz with actusl loaded vcore at 1.292v.

The issue with these cpu's is not the cpu but the motherboard.  I believe Agesa will give you your boost clocks but there is a problem with Agesa  From what I understand there is an Agesa on the way.

Adept III

This is a suggestion for AMD to improve their technical service. I filled out support request as suggested in the forum and got email back asking me to send pictures of my pc, motherboard, bios version, driver versions, windows version etc. Basically i have to describe every piece of my pc hardware and software for them to make any determination. So instead why don't you (AMD) create a "Report" function in your own AMD Master software that would gather all that info and send it to you? It can also run some sort of test to show all necessary parameters you looking to validate i.e. temps, boost, frequency whatever. I think this will really help out to expedite this process. Thanks,

Hello s_vayner‌,

I did find some anecdotal data you could try.  It seems a lot of users are having this boost clock issue with the 3900X, when using 3600MHz RAM or higher.

Have you tried running your RAM at a lower speed?  Say 3200 or 3400MHz?

Do your boost clocks improve?

i have not. will give it a go tonight. i always loaded XMP profile and left it there.

Some mobos are sending too much voltage when setting RAM (and therefore FCLCK) to 1800MHz (DDR) which is causing issues. VDDG set too high when it only needs ~0.95v and also SoC only needs 1.1v when they are being set 1.1v and 1.2v respectively.

How does it feel to have to Pay AMD to be a beta test mule?

Journeyman III

I have similar issue on Asus x570 prime pro 

Could be related to this MB bios. Getting 4.2 max boost instead of 4.4

Journeyman III 

Also check the document and search for X570 prime pro and It looks that more people have issue with this mb getting the right max. boost clocks