AMD Updates Ryzen Product Pages to Elaborate on "Max Boost Clocks"

[biffzinker]

If that isn't the best... I've got a dud. It takes over 1.3V for me to run 4.25 GHz on a 3700X...temps break 90C in AIDA64 stress test (default) with Corsair H150i, lol.

1.41V for a stable 4.175 GHz

 

[hzy4]

Mine 3700x runs 1.23v at 4.1Ghz but could not get stable at 1.35v at 4.3Ghz, so I figured to run auto until the drivers get tweaked and more info about safe voltage for 24/7 is available

 

[Xpect]

.... for a CPU where spec says 4400...

Well, no, Spec says 3600 MHz. Boost is Just an added Bonus if the CPU evaluates it can go higher.

 

[EarthDog]

Well, no, Spec says 3600 MHz. Boost is Just an added Bonus if the CPU evaluates it can go higher.

That isn't how it works though. If it says on a package that it should boost to X.XX GHz, then it should boost to that value. Many are finding it does not, even with 'nominal' conditions.

 

[TheMadDutchDude]

Mine break into the mid 70s or so, but I’m on custom water.

The only time I’ll ever see the full boost is if I get it under the cold stuff, which won’t happen.

My buddy in the UK has one that does 4.35 all core at 1.3v.

 

[Xpect]

That isn't how it works though. If it says on a package that it should boost to X.XX GHz, then it should boost to that value. Many are finding it does not, even with 'nominal' conditions.

It says on every IT Shop here in Germany "Clock 3.6GHz, Boost up to 4.4GHz"

In my opinion and if you read the exact wording for those products, Boost Clock is Just a Bonus. Advertising, Sure, but always as "up to". Guaranteed is only Stock Clock.

 

[EarthDog]

It says on every IT Shop here in Germany "Clock 3.6GHz, Boost up to 4.4GHz"

In my opinion and if you read the exact wording for those products, Boost Clock is Just a Bonus. Advertising, Sure, but always as "up to". Guaranteed is only Stock Clock.

It says MAXIMUM on their product page and box. And people with 'nominal' setups (as AMD described it themselves) are not able to reach it. It really seems that people who are reaching the max boost, there isn't a rhyme or reason to it. Great motheboard and cooling... still doesn't reach it... MEH parts, a couple have... it makes no sense. THAT is mainly my issue with it.

Why can't I reach my 3700X's max? I have a 3x120mm AIO, tried it across 6 motherboards ranging from entry level to mid-range and still, it's short.

 

[hzy4]

All x570 MB are garantied to support the 12 core chips so its unlikekly tha mb fault that 3700x cannot hit 4.4 boost

 

[Nordic]

I haven’t seen mine hit it’s max frequency, and it’s under custom water cooling.

What waterblock and radiator do you use? I have been thinking about installing my custom watercooling loop again. I was getting temps of 40c on my 4.8ghz@1.5v 2500k and my 7970 overclocked to 1425mhz. I was thinking that the same setup could potentially cool my 3900x to about the same temperature.

 

[EarthDog]

All x570 MB are garantied to support the 12 core chips so its unlikekly tha mb fault that 3700x cannot hit 4.4 boost

Sure.

I don't think it's a power issue in the first place though.

I was thinking that the same setup could potentially cool my 3900x to about the same temperature.

Heh... not going to happen, bud. 40C at full load, even on a 2500K (but at 4.8 GHz and 1.5V) sounds impossible or your load was something jenky like gaming and not a proper stress test.

 

[TheMadDutchDude]

What waterblock and radiator do you use? I have been thinking about installing my custom watercooling loop again. I was getting temps of 40c on my 4.8ghz@1.5v 2500k and my 7970 overclocked to 1425mhz. I was thinking that the same setup could potentially cool my 3900x to about the same temperature.

I’m running an EK Supremacy EVO, Hardware Labs 360 slim something or other and a 240 EK PE rad.

 

[bug]

It says on every IT Shop here in Germany "Clock 3.6GHz, Boost up to 4.4GHz"

In my opinion and if you read the exact wording for those products, Boost Clock is Just a Bonus. Advertising, Sure, but always as "up to". Guaranteed is only Stock Clock.

You have to realize this is an opinion shared only by you and probably a few other people that have never seen a CPU.

 

[Nordic]

Heh... not going to happen, bud. 40C at full load, even on a 2500K (but at 4.8 GHz and 1.5V) sounds impossible or your load was something jenky like gaming and not a proper stress test.

I stress tested it with prime95 and running the gpu full bore. I also was running boinc and folding@home 24/7 then, at least not when not using the computer. 40c max load was what I got.

The same cooling loop with only a 3930k couldn't keep that cpu below 80c when overclocked to 4.6ghz.

I had a monsta 240mm long 80mm thick radiator with 32mm panaflows and 25mm spacers between the fans and the radiator. The fans ran at full speed. Swiftech waterpump and pump with an alphacool full cover gpu waterblock. The radiator is so thick that I used it as the resivor. I have pictures somewhere. I bet I even made a thread about it back then around 2013.

A 3900x doesn't use the amount of power as a 3930k, but does have a smaller node. Since I am running it at stock, and letting it overclock itself I should get some great temps.

 

[EarthDog]

I stress tested it with prime95 and running the gpu full bore. I also was running boinc and folding@home 24/7 then, at least not when not using the computer. 40c max load was what I got.

The same cooling loop with only a 3930k couldn't keep that cpu below 80c when overclocked to 4.6ghz.

I had a monsta 240mm long 80mm thick radiator with 32mm panaflows and 25mm spacers between the fans and the radiator. The fans ran at full speed. Swiftech waterpump and pump with an alphacool full cover gpu waterblock. The radiator is so thick that I used it as the resivor. I have pictures somewhere. I bet I even made a thread about it back then around 2013.

A 3900x doesn't use the amount of power as a 3930k, but does have a smaller node. Since I am running it at stock, and letting it overclock itself I should get some great temps.

Sure.

Anyway, there's no chance for these chips to get anywhere close to those temps. Soooooooo many different variables involved, but no way.

 

[Nordic]

Ok, I get it, you're bitter.
Try this: fire up SuperPi (single threaded) and see whether you reach the advertised speeds. If you don't, then yes, AMD has shafted you.

I apologize that it took me this long to get back to you, but here are my results. I reset hwinfo's sensor results right before running superpi, and took the screenshot as soon as it was over. Looks like during this run I peaked at 4400.3mhz on core 5. Core 5 is my fastest core in my first CCX.

 

[EarthDog]

Looks like during this run I peaked at 4400.3mhz on core 5. Core 5 is my fastest core in my first CCX.

Doesn't the 3900x boost to 4.6GHz?

 

[Nordic]

Doesn't the 3900x boost to 4.6GHz?

It is supposed to. I have seen my 3900x boost as high as 4525mhz max, and it doesn't sustain that speed for more than a second. I have only gotten to 4525mhz while idle that I have seen. Bug specifically called out superpi so I shared my results under that workload.

 

[londiste]

Isn't highest on that screenshot 4450.3 MHz on Core 3?

 

[Nordic]

I didn't see that. The results are better then I thought.

 

[EarthDog]

Isn't highest on that screenshot 4450.3 MHz on Core 3?

Ohhhh, now only 150 Mhz away from what it should be doing!

 

[Nordic]

Even if I was getting 4.6ghz, it doesn't sustain those speed very long. I averaged about 4080mhz on my fastest core while running superpi. I reset hwinfos recorded values, immediately clicked start on superpi. As soon as superpi was done, I took the screenshot.

The problem isn't just the inability to reach the boost clock but the inability to sustain a higher clockspeed than 4.2ghz.

 

[bug]

I apologize that it took me this long to get back to you, but here are my results. I reset hwinfo's sensor results right before running superpi, and took the screenshot as soon as it was over. Looks like during this run I peaked at 4400.3mhz on core 5. Core 5 is my fastest core in my first CCX.

First of all, no apology needed, I wasn't tasking anyone, I was just offering a way to test boost under the best circumstances.
Second of all, bummer. It seems like AMD really pulled numbers from where the Sun don't shine

 

[medi01]

I'm an old fashioned simple lad.when I buy something according to the box, I expect it to perform like it. You write 4.5 ghz boost I expect it to do so with all cores.I don't buy an 8 core processor just so 1 core reaches advertised speeds.it's shame and scam.

Well, then so is Intel's CPU (and even more so, as frequency drops much sharper with more threads)

 

[Nordic]

Well, then so is Intel's CPU (and even more so, as frequency drops much sharper with more threads)

Given proper cooling, I don't think there isn't an intel CPU that won't do the boost frequency on all cores if you set it to do that in the bios. That is technically overclocking, but it does work. AMD's boost is like this magical thing that sometimes boosts the cpu higher than we can overclock it ourselves. Intel's boost is deterministic. AMD's is something else.

EDIT: I think it was debauer who said he is unsure if AMD even knows why their boost is behaving the way it is.

 

[medi01]

Given proper cooling, I don't think there isn't an intel CPU that won't do the boost frequency on all cores if you set it to do that in the bios. That is technically overclocking, but it does work. AMD's boost is like this magical thing that sometimes boosts the cpu higher than we can overclock it ourselves. Intel's boost is deterministic. AMD's is something else.

EDIT: I think it was debauer who said he is unsure if AMD even knows why their boost is behaving the way it is.

By default (and without excessive power consumption) AMD boosts to 96% of the advertised max on all cores.
While Intel boosts to 89%.

Cough.