RyZen 3000 Boost Issue: What's your take?

[TheLostSwede]

The only thing that's missing is Cool 'n Quiet, rest is there. For Elite the only difference is that you can't set clock control to .01 as it accepts 100.01 as 101 Mhz.

I can't find High Frequency Support under the memory settings, but maybe that's somewhere else?

Below are the settings I've changed. Only just changed the Global C-state Control option though.

 

[Chomiq]

I can't find High Frequency Support under the memory settings, but maybe that's somewhere else?

Below are the settings I've changed. Only just changed the Global C-state Control option though.

High Freq Sup shows up once you enable XMP.

 

[TheLostSwede]

High Freq Sup shows up once you enable XMP.

Ah, that'd be why, can't run XMP so...

 

[biffzinker]

For Elite the only difference is that you can't set clock control to .01 as it accepts 100.01 as 101 Mhz.

Any attempt at adjusting from 100 MHz to 101 MHz results in a black screen with a blinking cursor.

The only way out is to hold power button for 30 seconds then bridge the Clr_CMOS pins.

 

[Chomiq]

Any attempt at adjusting from 100 MHz to 101 MHz results in a black screen with a blinking cursor.

I didn't even attempt to boot at 101, "accepts" means that it registers value as 101 instead of 100.01. Once I've seen that I said "nope, nope, nope" and set it to 100. If it registers 100.01 as 100.01 it should be fine, otherwise stick to 100 instead of Auto.

 

[TheLostSwede]

Any attempt at adjusting from 100 MHz to 101 MHz results in a black screen with a blinking cursor.

Odd, I was running at 101MHz at one point, but I'm not going to say it was 100% stable at that point, more like 98%...

 

[Liviu Cojocaru]

The highest boost I've seen on my 3600X was 4376mhz...I just OC'ed to 4250 on all cores with 1.3v for now, will try to lower the voltage and see where can I keep it

 

[biffzinker]

I didn't even attempt to boot at 101, "accepts" means that it registers value as 101 instead of 100.01.

My board doesn't allow a decimal value of .01. The only changes you can make are Auto, 100, and 101-3.

 

[TheLostSwede]

My board doesn't allow a decimal value of .01. The only changes you can make are Auto, 100, and 101-3.

Maybe the clock gen doesn't allow for it? Or ASRock didn't implement that fine grained settings in the UEFI.

Can't say enabling Global C-states makes any difference whatsoever. At least not that I can see.
Boosts are just as random and unnoticeable as before.

 

[johnny-r]

I still disagree, I don't believe its the CPU:

I used the same app, CPU-Z and Aida64 reports 4199.xx MHZ per core...

 

[4rk4noid]

Base Clock 3.8GHz Max Boost Clock 4.6GHz

I don't understand why so much talk about this. Is it working bellow base clock? Everyone knows max boost clocks depends on a few conditions, one of them is temperature.

 

[Apocalypsee]

Base Clock 3.8GHz Max Boost Clock 4.6GHz

I don't understand why so much talk about this. Is it working bellow base clock? Everyone knows max boost clocks depends on a few conditions, one of them is temperature.

Welcome to TPU. Its just how people talk to each other in this board.

 

[Roddey]

Welcome to TPU. Its just how people talk to each other in this board.

Ya I can just see me sitting here in a coat and long johns just to get those extra fps this winter! Or running benchmarks for fun! Can't wait!

 

[Nordic]

Base Clock 3.8GHz Max Boost Clock 4.6GHz

I don't understand why so much talk about this. Is it working bellow base clock? Everyone knows max boost clocks depends on a few conditions, one of them is temperature.

Temperature is irrelevant to the problem. I still don't get max boost clocks. I cannot create conditions where I reach max boost clocks. Temperature, voltage, or workload are all irrelevant variables because I can not get more than 4525mhz.

 

[Hellfire]

So I've only had my rig up for the day so far it's great.

First thing I did was change auto voltage to 1.25v in the bios, auto had it 1.47c and idle temp of 46c.

1.25 dropped the CPU package to 36c.

Stock it was running Upto 4608mhz on some cores under load with more than one different core hitting it.

Day one, I've got 4350mhz all core boost stable at 1.25v but I bumped to 1.27 for headroom.

My aim is to try to get 4400mhz on all cores.

 

[Deleted member 185158]

Base Clock 3.8GHz Max Boost Clock 4.6GHz

I don't understand why so much talk about this. Is it working bellow base clock? Everyone knows max boost clocks depends on a few conditions, one of them is temperature.

What it is that most people believe that the Cpu should Max boost ALL cores to 4.6ghz. When in fact that is only the single core boost feature that reaches or ever has reached, in the past even on other AMD platforms. But insisted it is that all cores must reach this here max frequency.
In addition, to that note, It is unrecognized that the base frequency is actually a p-state where boost frequencies are not, therefor it should be considered a boost range frequency which means the difference between x and y systems.

You can tweak PBO for a higher all core frequency, but has it's limits in a variety of ways. Even if you could pull off the 4.6ghz all core boost, it's going to take 1.5v and some chilling to get there. That is part of design with top boards, Ryzen chips and LN2.

For the most part here, we have a lot of intermediate overclockers going for best performance but forgetting the wide range of possible overclocking outcomes a lot of times dependent on enthusiasm where you'd see enthusiasts using extreme cooling methods.
With a wide range low to high end boards, those with high end boards lack a little realization the board was meant for extreme cooling, and those with mid range boards lack a little realization they will never get to desired speeds, setting the bar too high on ambient air and AIO coolers.

 

[londiste]

Straw man argument. Almost nobody insists it should boost on all cores.
It would be OK if it boosted to the spec max boost on one core but in many/most cases it does not.

I cannot find any way to make my 3600X boost to 4.4 GHz. I see 4.3 GHz max and that is for a split second, with any single-core load it consistently runs at 4.2GHz.

 

[Deleted member 185158]

Straw man argument. Almost nobody insists it should boost on all cores.
It would be OK if it boosted to the spec max boost on one core but in many/most cases it does not.

I cannot find any way to make my 3600X boost to 4.4 GHz. I see 4.3 GHz max and that is for a split second, with any single-core load it consistently runs at 4.2GHz.

How are you running that 3600 Ryzen on a Z370 board? Update you system specs would ya?

 

[londiste]

How are you running that 3600 Ryzen on a Z370 board? Update you system specs would ya?

I have more than one system. In addition to the primary one with i5 I have an internet box with B450/2400G and a test box with B350/3600X.
3600X is on an Gigabyte GA-AB350N Gaming WIFI with the latest BIOS - F42a (AGESA 1.0.0.3 ABB).

 

[Nordic]

What it is that most people believe that the Cpu should Max boost ALL cores to 4.6ghz. When in fact that is only the single core boost feature that reaches or ever has reached, in the past even on other AMD platforms. But insisted it is that all cores must reach this here max frequency.

While it would be nice to have it possible to have a higher all core boost, this is at least a variable I can manipulate given voltage and temperature. I have been able to increase my average clockspeed with voltage and better cooling.

It is as if you haven't kept up with the thread or the problem. I can not currently reach 4600mhz single core under any circumstances. Neither can 94% of 3900x owners according to debauer.

 

[TheLostSwede]

Base Clock 3.8GHz Max Boost Clock 4.6GHz

I don't understand why so much talk about this. Is it working bellow base clock? Everyone knows max boost clocks depends on a few conditions, one of them is temperature.

1. In this case, it has nothing to do with temperature.
2. People are experience all sorts of random "hard" upper limits, regardless of cooling or settings.
3. Some of us, like me, has actually gotten past the hard limit thanks to an updated UEFI, so it's highly unlikely to be a CPU issue.
4. PBO/PBO+ does nothing in most cases.
5. The overbearing silence from AMD isn't helping, which leads to more discussion...

 

[Hellfire]

While it would be nice to have it possible to have a higher all core boost, this is at least a variable I can manipulate given voltage and temperature. I have been able to increase my average clockspeed with voltage and better cooling.

It is as if you haven't kept up with the thread or the problem. I can not currently reach 4600mhz single core under any circumstances. Neither can 94% of 3900x owners according to debauer.

94% did you calculate that or pull it out your butt. I've had no problems with 4600 on some cores.

 

[TheLostSwede]

What it is that most people believe that the Cpu should Max boost ALL cores to 4.6ghz. When in fact that is only the single core boost feature that reaches or ever has reached, in the past even on other AMD platforms. But insisted it is that all cores must reach this here max frequency.
In addition, to that note, It is unrecognized that the base frequency is actually a p-state where boost frequencies are not, therefor it should be considered a boost range frequency which means the difference between x and y systems.

You can tweak PBO for a higher all core frequency, but has it's limits in a variety of ways. Even if you could pull off the 4.6ghz all core boost, it's going to take 1.5v and some chilling to get there. That is part of design with top boards, Ryzen chips and LN2.

For the most part here, we have a lot of intermediate overclockers going for best performance but forgetting the wide range of possible overclocking outcomes a lot of times dependent on enthusiasm where you'd see enthusiasts using extreme cooling methods.
With a wide range low to high end boards, those with high end boards lack a little realization the board was meant for extreme cooling, and those with mid range boards lack a little realization they will never get to desired speeds, setting the bar too high on ambient air and AIO coolers.

I think you've read some other forum than TPU, as I don't think anyone here is expecting all core boost to 4.6GHz, but when not even one core gets within 100MHz of boost, something is weird.

Read my points above-

 

[Nordic]

94% did you calculate that or pull it out your butt. I've had no problems with 4600 on some cores.

I am quoting a survey conducted by debauer from a video earlier in this thread.

Edit: because I am certain that you will not go back and find the video, because it is clear you aren't actively keeping up with the thread, here is the video.

Edit 2: thelostswede beat my edit.

 

[TheLostSwede]

94% did you calculate that or pull it out your butt. I've had no problems with 4600 on some cores.

Your reading comprehension needs improvement.
He said according to der8auer.
Please see.