Ryzen 1700X throttling at 55C?

[ToothlessFury]

Hello everyone,

 

I have recently purchased a new Ryzen based build.
And I would like to ask for your best guess or advise on what might be wrong.

 

First problem is:

While stress testing the system using the OCCT software the screen freezes from time to time, but the stress test continues and when it unfreezes the time indicator jumps ahead.

There is a straight line in the monitoring graph for the duration of the freeze (no data input).

Hwinfo which was logging all the values jumps for the duration of the freeze too (30s and more gaps).

Is this normal?

 

Second problem is:

The CPU starts to "thermal" throttle at 55C cca 8 minutes into the test.

This is very clearly seen on the provided measurement graphs, that were captured while the stress test was running.

It does not do that in AIDA stress test.

What could cause this throttling? The case has more than enough airflow.

 

I have tried:

Reseating the CPU cooler

 

Measurements:
https://imgur.com/a/N6Zpi

 

The system is in stock settings (no OC).

System info (everything stock, no OC):

* CPU: Ryzen 1700X
* MB: ASUS ROG STRIX X370-F
* RAM: Patriot Viper 16GB 3000Mhz CL16 (2 sticks)
* PSU: Corsair RM550X
* GPU: NVIDIA GT710
* SSD: Patriot flare 120GB SSD (System)
* HDD: Multiple WD blue/red HDDs
* CPU Cooler: Arctic freezer 33
* System: Windows 10 64 bit (latest available build)

 

Is it a badly seated cooler? Might it be VRM throttling? Or is it indicating a bad PSU?

I have already posted a problem with random black screens with the same steup which was not yet solved.

[DocSwag]

What voltages is it running at?

 

EDIT: didn't see the voltage graphs, perhaps manually set voltage to 1.25v?

[Ace McPlane]

14 minutes ago, ToothlessFury said:

Hello everyone,

 

I have recently purchased a new Ryzen based build.
And I would like to ask for your best guess or advise on what might be wrong.

 

First problem is:

While stress testing the system using the OCCT software the screen freezes from time to time, but the stress test continues and when it unfreezes the time indicator jumps ahead.

There is a straight line in the monitoring graph for the duration of the freeze (no data input).

Hwinfo which was logging all the values jumps for the duration of the freeze too (30s and more gaps).

Is this normal?

 

Second problem is:

The CPU starts to "thermal" throttle at 55C cca 8 minutes into the test. This is very clearly seen on these graphs, that were captured while the stress test was running.

It does not do that in AIDA stress test.

What could cause this throttling? The case has more than enough airflow.

 

I have tried:

Reseating the CPU cooler

 

Measurements:
https://imgur.com/a/N6Zpi

 

The system is in stock settings (no OC).

System info (everything stock, no OC):

* CPU: Ryzen 1700X
* MB: ASUS ROG STRIX X370-F
* RAM: Patriot Viper 16GB 3000Mhz CL16 (2 sticks)
* PSU: Corsair RM550X
* GPU: NVIDIA GT710
* SSD: Patriot flare 120GB SSD (System)
* HDD: Multiple WD blue/red HDDs
* CPU Cooler: Arctic freezer 33
* System: Windows 10 64 bit (latest available build)

 

Is it a badly seated cooler? Might it be VRM throttling? Or is it indicating a bad PSU?

I have already posted a problem with random black screens with the same steup which was not yet solved.

1. Reset your CMOS.

2. Go into your bios and confirm that your CPU frequency and CPU core voltage are at stock settings.

3. Try stressing again.

 

I'm a little confused about your graphs though.

What were you doing or running and the end of those graphs?

It's looks like your CPU voltage went over 1.4 at the end there.. without an overclock? That shouldn't be.

I wouldn't even go over 1.3 overclocking unless I had a custom water loop.

 

 

 

[rio2]

The temporary GUI freezing during OCCT is normal, because it takes up all the system resources and leave none to the GUI and input handling, but about about the throttling I honestly have no idea, but measuring the VRM temperatures might be worth giving a shot.

[rio2]

4 minutes ago, DocSwag said:

What voltages is it running at?

 

EDIT: didn't see the voltage graphs, perhaps manually set voltage to 1.25v?

Manually setting the voltage when not overclocking is a terrible idea, because the chips automatically adjust the voltage at stock given what turbo clock it is running as at and if single core turbo kicks in that voltage might not be enough and your system will crash and yes Ryzen chips usually go way over 1.25v during single core turbo especially during XFR turbo.

[DocSwag]

2 minutes ago, rio2 said:

Manually setting the voltage when not overclocking is a terrible idea, because the chips automatically adjust the voltage at stock given what turbo clock it is running as at and if single core turbo kicks in that voltage might not be enough and your system will crash and yes Ryzen chips usually go way over 1.25v during single core turbo especially during XFR turbo.

Pretty sure turbo is supposed to not go that high if it doesn't have the voltage.

[ToothlessFury]

16 minutes ago, stateofpsychosis said:

1. Reset your CMOS.

2. Go into your bios and confirm that your CPU frequency and CPU core voltage are at stock settings.

3. Try stressing again.

 

I'm a little confused about your graphs though.

What were you doing or running and the end of those graphs?

It's looks like your CPU voltage went over 1.4 at the end there.. without an overclock? That shouldn't be.

I wouldn't even go over 1.3 overclocking unless I had a custom water loop.

 

 

 

Ok, will try reseting the CMOS and test again.

But looking into BIOS, everything is set to auto. 

Nothing was touched there.

 

After the stress test ended I went to watch the meassured values. 

So it would have been almost idle.

 

Even now watching the idle voltage it jumps between 1.2V to 1.42V max meassured is 1.47

[Ace McPlane]

15 minutes ago, ToothlessFury said:

Ok, will try reseting the CMOS and test again.

But looking into BIOS, everything is set to auto. 

Nothing was touched there.

 

After the stress test ended I went to watch the meassured values. 

So it would have been almost idle.

 

Even now watching the idle voltage it jumps between 1.2V to 1.42V max meassured is 1.47

Okay, are you sure you didn't install some kind of auto overclocking software in windows?

You can overclock in windows outside of the bios too and those stupid auto-overclocks always add ridiculous amounts of voltage, so check that.

What utilities did you specifically install while setting up your rig?

Often these types of things are right there along side your motherboard drivers on the manufacturers page so you could have installed one without knowing.

[ToothlessFury]

5 minutes ago, stateofpsychosis said:

Okay, are you sure you didn't install some kind of auto overclocking software in windows?

You can overclock in windows outside of the bios too and those stupid auto-overclocks always add ridiculous amounts of voltage, so check that.

What utilities did you specifically install while setting up your rig?

Often these types of things are right there along side your motherboard drivers on the manufacturers page so you could have installed one without knowing.

That is very unlikely. It is a fresh install of windows.

The only thing that was installed were drivers (from official vendor websites - not ASUS - no utility software, not even the sound software from ASUS), hwinfo and a few basic programs for work.

 

But the strange thing is that i have noticed that my ryzen 7 is running at 3.5Ghz after the last BIOS update? With core boost to 3.8Ghz.

It also visible on the measurement graph.

[Ace McPlane]

5 minutes ago, ToothlessFury said:

That is very unlikely. It is a fresh install of windows.

The only thing that was installed were drivers (from official vendor websites - not ASUS - no utility software, not even the sound software from ASUS), hwinfo and a few basic programs for work.

 

But the strange thing is that i have noticed that my ryzen 7 is running at 3.5Ghz after the last BIOS update? With core boost to 3.8Ghz.

It also visible on the measurement graph.

That's about stock: https://www.amd.com/en/products/cpu/amd-ryzen-7-1700x?&gclid=Cj0KCQiAuP7UBRDiARIsAFpxiRJhf-q68k9ekOom96bw1HRQU2ZLhfoHdg3B8gtJn4sSz0J18wO4WIAaAmdgEALw_wcB

 

[ToothlessFury]

3 minutes ago, stateofpsychosis said:

That's about stock: https://www.amd.com/en/products/cpu/amd-ryzen-7-1700x?&gclid=Cj0KCQiAuP7UBRDiARIsAFpxiRJhf-q68k9ekOom96bw1HRQU2ZLhfoHdg3B8gtJn4sSz0J18wO4WIAaAmdgEALw_wcB

 

The core boost yes, but the base clock? It has been like 3394 Mhz before the update and now it is 3492 Mhz

[Ace McPlane]

 

7 minutes ago, ToothlessFury said:

That is very unlikely. It is a fresh install of windows.

The only thing that was installed were drivers (from official vendor websites - not ASUS - no utility software, not even the sound software from ASUS), hwinfo and a few basic programs for work.

 

But the strange thing is that i have noticed that my ryzen 7 is running at 3.5Ghz after the last BIOS update? With core boost to 3.8Ghz.

It also visible on the measurement graph.

Apparently these high voltages are normal and they're due to XFR

 

https://www.reddit.com/r/Amd/comments/6dt3qm/1700x_voltage_extremely_high_whats_normal/

 

Are you using the stock cooler?

If so, I'd recommend just upgrading the cooler if you're thermal throttling.

 

 

[Ace McPlane]

Just now, ToothlessFury said:

The core boost yes, but the base clock? It has been like 3394 Mhz before the update and now it is 3492 Mhz

That's not a big enough difference to really matter anyways.

[ToothlessFury]

3 hours ago, stateofpsychosis said:

 

Apparently these high voltages are normal and they're due to XFR

 

https://www.reddit.com/r/Amd/comments/6dt3qm/1700x_voltage_extremely_high_whats_normal/

 

Are you using the stock cooler?

If so, I'd recommend just upgrading the cooler if you're thermal throttling.

 

 

The cooler is written int he System info. It is Artic freezer 33.

[Ace McPlane]

13 minutes ago, ToothlessFury said:

The coolers is written int he System info. It is Artic freezer 33.

Oh, that should be enough.

 

To be honest, I don't even think it really is thermal throttling.

You said AIDA64 didn't do it right?

I'd trust AIDA64 over OCCT, but I guess wait until more people comment and see what they say.

[0x1e]

3 hours ago, ToothlessFury said:

The coolers is written int he System info. It is Artic freezer 33.

I would manually set the cpu fan at 100%, add an offset Vcore of 0.10 - 0.25 (it's a little high, but it's just for testing).
I would then add a fan blowing air at the VRMs while testing.
If all goes good then I would back off on the Vcore a little and test again. Then again and again until either it throttles, or was never the Vcore to begin with.
Then the VRM fan, then the CPU cooler.

Go overboard then divide and conquer.

[AndrewN]

Hi!

Did you solve your problem?

I have the same problem with throttling and freeze while rendering in 3ds max (vray). I have the same MB and CPU. I thougth about PSU issue, but it works perfect at other PC (Intel).

I'm sure that this problem is related with thermal package, because it appers only in these moments - http://joxi.ru/ZrJp9WZs9aWDB2