3900x issue

qtqc1

n00b
Joined
Mar 22, 2016
Messages
53
Hello everyone. I picked up a new 3900x today and installed it on my existing Asus Crosshair hero Vii. I updated the bios before installing the cpu. My idle temps are around 55-60c where I was getting around 35c with my 2600x. When I try and strees the cpu, the pc shuts down. I am using an existing Thermaltake FLOE Ring AIO water cooler. Is it possible the BIOS has the wrong temp readings?
 
Make sure not only center is has paste.
I've seen aio's being used in reviews.

I see 50c idle on custom loop, letting it do it's thing.
and my idle isn't really idle but still not any Real load.
 
Stock 3900x shouldn't be any worse temp wise than any other stock 105w part. This is likely a mounting issue with your cooler.
 
What does CPU-Z show your core voltage is? My board was over-volting my 3900x by quite a bit. I manually dialed in 1.3125 volts.
 
What does CPU-Z show your core voltage is? My board was over-volting my 3900x by quite a bit. I manually dialed in 1.3125 volts.
jayztwocents mentioned this. the bios apply wrong voltages to numerous things, it was asus he mentioned also.

 
jayztwocents mentioned this. the bios apply wrong voltages to numerous things, it was asus he mentioned also.



Yeah, I watched that video even before I put my system together and it was one of the first things I looked for. Sure enough it was way over volted. Another member here tangoseal had the same issue. I believe he is running a MSI board. I don't know if this issue with x570 boards exclusively, or if it's a general bios issue with Ryzen 3000 series that could potentially affect any supported chipset.
 
Yes, it was the stock voltage. put it down and holding good now. Just for comparison, here are my cinebench numbers.
upload_2019-7-12_19-31-46.png
 
Watching this thread for reference. I just got my 3900x ordered today, and it’ll be going in my sig rig.
 
Its r15, I need to download the r20 since its the new standard.
Not sure if anyone else has the vii hero but I'm noticing a weird issue since updating the bios. I updated last week and I noticed that each morning I was finding my computer off. Anyone else experiencing this?
 
Hello everyone. I picked up a new 3900x today and installed it on my existing Asus Crosshair hero Vii. I updated the bios before installing the cpu. My idle temps are around 55-60c where I was getting around 35c with my 2600x. When I try and strees the cpu, the pc shuts down. I am using an existing Thermaltake FLOE Ring AIO water cooler. Is it possible the BIOS has the wrong temp readings?

You're not the only one having the temperature issue. Mine also idles around 45c-55c. I tried reseating the heatsink (H115i Pro and NH-D15)so many times I ran out of the paste. I also tried off-setting the Vcore like so many suggested. It didn't work, either. My motherboard is Gigabyte X470 Gaming 7.
 
This might not be a solution but I had temperatures issues on a new 6700K that I had sent to siliconlottery to de-lid and see what numbers/temps he got. I was using the AIO cooler that I had been using on my older system and after reseating multiple times with no luck I purchased a new cooler since that was the only difference between what I had and what SL tested my chip with. I believe I had the H100 and went to an H110i. I'm not sure what the problem could have been but a different cooler solved my heat issue. I know the die placement for these chips appears to be off center so maybe that's a contributing factor with some of the water blocks/AIOs.
 
Yeah, I watched that video even before I put my system together and it was one of the first things I looked for. Sure enough it was way over volted. Another member here tangoseal had the same issue. I believe he is running a MSI board. I don't know if this issue with x570 boards exclusively, or if it's a general bios issue with Ryzen 3000 series that could potentially affect any supported chipset.

I have a Gigabyte Aorus x570 master. It was cooking my chip but using neg offset I now have it at 1.28v full load. Idle in 30s full load at 68ish. Room is AC cooled. Custom loop cooled cpu and gpu.

williacm

Your AIO probably evaporated most of the coolant out. AIOs just like custom loops lose liquid volume over time. Or your pump was weak or your microfins in the AIO block were clogged up with biogunk.

I dont think die placement has any affect on thermal transfer to be honest. Once I tempered my voltages my thermals were as normal as any typical cpu.
 
Those idle themps seem high. I'm getting about 41c with a Le GRAND MACHO RT and stock settings.
 
Thanks gents, I plan on doing a custom loop in the future, but for now I just wanted to make sure I wasn't pushing a thermal threshold.
 
3900x and H150i Pro with 37c idle and 81c under load with Prime. Thoughts on those temps?

I assume that is with no avx correct? I get similar temps with my CPU OC'ed to 4.4. AIDA64 does run AVX but for some reason I have to disable it on prime as it really is more of a burn test then load test lol, atleast disabling avx2 helps. With avx 2 its even hard to cool at stock lol, because I am not really using more voltage then stock. Its barely running 1.328v at full load.
 
So I returned my 3900X and picked up a 3800X. AIDA64 with AVX2 on reports 79C at stock.
Oddly, lowering Vcore causes significant performance loss. When Vcore was manually lowered to 1.275v in the bios, Geekbench scores were lower by 1000 points than when the Vcore was left at auto in the bios. As I increase the Vcore, the bench scores also climbed.
 
So I returned my 3900X and picked up a 3800X. AIDA64 with AVX2 on reports 79C at stock.
Oddly, lowering Vcore causes significant performance loss. When Vcore was manually lowered to 1.275v in the bios, Geekbench scores were lower by 1000 points than when the Vcore was left at auto in the bios. As I increase the Vcore, the bench scores also climbed.

You may have inadvertently starved the CPU of voltage it needed for boost clocks. It isn't uncommon to see these chips pull 1.5v for short periods while boosting a single core to the max advertised speeds. If you set the voltage to manual, then it won't use more than you've specified.
 
You may have inadvertently starved the CPU of voltage it needed for boost clocks. It isn't uncommon to see these chips pull 1.5v for short periods while boosting a single core to the max advertised speeds. If you set the voltage to manual, then it won't use more than you've specified.

I’ve noticed the performance scale with voltage too. My issue isn’t that it needs up to 1.5 to boost but that mine pulls 1.45 at all times when left at auto even while doing nothing but staring at CPUZ. In fact, the only time it drops below that is courtesy of vdroop at full load. I’ve followed all the AMD guidelines by installing the latest chipset driver, setting the power options to Ryzen Balanced and using a single monitoring application.
 
I’ve noticed the performance scale with voltage too. My issue isn’t that it needs up to 1.5 to boost but that mine pulls 1.45 at all times when left at auto even while doing nothing but staring at CPUZ. In fact, the only time it drops below that is courtesy of vdroop at full load. I’ve followed all the AMD guidelines by installing the latest chipset driver, setting the power options to Ryzen Balanced and using a single monitoring application.
There is nothing wrong with pulling 1.4-1.5 volts when the CPU isn't doing real work.

If you have Corsair iCue make sure you turn it off, it causes high idle on all Ryzen CPU's.

On my 3800x sytem on the X470 Gaming Pro Carbon I see below 1V w/ Ryzen Balanced power plan occasionally during idle. The CPU makes clock adjustments at such a fast rate, that any program running in the background will re-awake the cores.

You have to remember that Zen2 flips between power states so fast, that most reporting software like HWINFO64 may not even notice it.

But again, there is no problem with 1.4-1.5 during low workload. As long as you aren't seeing high voltages under MT workload you are fine.

Finally, i've noticed that this CPU does not respond at all to PBO at all. It's either a AEGSA bug issue, or AMD has tweaked these CPU's so good that PBO is now irrelevant. I see slightly higher temps on the 3800x compared to the 2700x, and the CPU is being pushed to 4.550GHZ without me even touching any PBO settings.

Here are my CPU-Z results - https://valid.x86.fr/v8dpi7
 
There is nothing wrong with pulling 1.4-1.5 volts when the CPU isn't doing real work.

If you have Corsair iCue make sure you turn it off, it causes high idle on all Ryzen CPU's.

On my 3800x sytem on the X470 Gaming Pro Carbon I see below 1V w/ Ryzen Balanced power plan occasionally during idle. The CPU makes clock adjustments at such a fast rate, that any program running in the background will re-awake the cores.

You have to remember that Zen2 flips between power states so fast, that most reporting software like HWINFO64 may not even notice it.

But again, there is no problem with 1.4-1.5 during low workload. As long as you aren't seeing high voltages under MT workload you are fine.

Finally, i've noticed that this CPU does not respond at all to PBO at all. It's either a AEGSA bug issue, or AMD has tweaked these CPU's so good that PBO is now irrelevant. I see slightly higher temps on the 3800x compared to the 2700x, and the CPU is being pushed to 4.550GHZ without me even touching any PBO settings.

Here are my CPU-Z results - https://valid.x86.fr/v8dpi7

According to AMD themselves, that voltage is ok PERIODICALLY to achieve boost. This is not the case with my 3900x. It’s constant regardless of boost or load. In fact AMD has a link to submit such behavior specifically because it isn’t normal.
 
According to AMD themselves, that voltage is ok PERIODICALLY to achieve boost. This is not the case with my 3900x. It’s constant regardless of boost or load. In fact AMD has a link to submit such behavior specifically because it isn’t normal.

See if there is any software running on your system that monitors CPU stats. Like I said, iCue was causing this for me and I had to disable it.

Still, if you have any software putting any demand for processing it's going to cause the voltage to go up, period.
 
See if there is any software running on your system that monitors CPU stats. Like I said, iCue was causing this for me and I had to disable it.

Still, if you have any software putting any demand for processing it's going to cause the voltage to go up, period.

The only monitoring tool i'm using is CPU-Z. I've closed iCUE, I've closed MSI Afterburner/Riva Tuner, I've ended processes that were still active with any monitoring apps that may have still been running after I closed out the main program. I don't have Ryzen Master running. There is no load on the CPU and it is a steady 1.448v
 
The only monitoring tool i'm using is CPU-Z. I've closed iCUE, I've closed MSI Afterburner/Riva Tuner, I've ended processes that were still active with any monitoring apps that may have still been running after I closed out the main program. I don't have Ryzen Master running. There is no load on the CPU and it is a steady 1.448v

A no load idle is ok. It can be 1.5v at idle and be safe. Do a load and it should drop to 1.35v or less if it s correctly being applied.

Remember its about amperage not volts. 1.45v full load is going to generate more heat, even if at 105watts its pulling less current. Processors are weird when it comes to power mathematics.

Because one would think to reach a TDP of 105watts you would want to run 105volts LOL that would be 1 amp of current, but instead we feed processors 1.45v and its like 75 amps at 105 watts tdp. Then we drop the voltage to 1.3 and its like 80 amps to reach 105w tdp. We go even lower and say 1.25v @ 105watts TDP and the current jumps even more to 84 amps.

However we find that the more power phases we can divide the load across we can keep those amps at bay by splitting the work across 8 or 10 or 12 or even now 16 phases.

It is interesting that processors increase in heat as more voltage is applied as we still approach TDP but it turns out that less voltage applied is reduced thermals and I believe we should be factoring in resistance of the silicon and copper and all the materials that comprises the processor. It turns out that lower voltage decreases thermals because of the whole R being thrown into the equation. So it isnt so much the amperage that is causing the heat but the resistance to that amperage at higher voltages that seems to be culprit.

Reduced voltages at load is what matters. Not idle. Idle is not pulling much current, thus resistance doesnt generate electromotive friction, thereby a higher voltage under load is the most important thing to curtail.

A good analogy might be a water pipe, where water sits in the pipe @ 50psi but its not flowing anywhere. As soon as the valve is open the pressure drops to 40psi but the current is very high through the downstream pipes and faucets. I cant think of an analogy for the thermals right now but I am sure I can come up with one later.

So voltage is high at idle because the electron pressure potential (something I made up for the sake of analogy) is higher until the valve (processor pipelines) lets the electrons flow (current) through the silicon. You will see a voltage reduction under load even if the current is actually high.

Someone save me before this ends in disaster hahaha
 
Last edited:
The only monitoring tool i'm using is CPU-Z. I've closed iCUE, I've closed MSI Afterburner/Riva Tuner, I've ended processes that were still active with any monitoring apps that may have still been running after I closed out the main program. I don't have Ryzen Master running. There is no load on the CPU and it is a steady 1.448v

Same here on my 3900x & Aorus Master.... I was just going to leave this CPU at stock and just focus on memory overclocking; but it is out of control from a voltage perspective. I've tried all the supposed "fixes", changing cpu volts from auto to "normal" (this did reduce voltage by a tiny bit, ~.02), windows power plan while better still ends up with the same behavior. (Power Save plan does fix the issue, but performance goes in the toilet.) My cpu and water temps stays relatively cool compared to what my threadripper 2950x used to output. I Guess I'll have to do a manual overclock and reduce the voltage that way; which I kindof didn't want to do. (I'd rather run stock and have higher boost frequencies for lightly threaded stuff) Its too bad they didn't have this figured out at launch; Precision Boost features on my TR 2950x worked great, but for whatever reason its completely borked on ryzen 3000.

One of the strangest things about it is that idle seemed to work right after installing the ryzen x570 chipset driver and powerplan; but after a few reboots; I now never see idle voltages and clocks anymore on the ryzen balanced power plan.
 
A no load idle is ok. It can be 1.5v at idle and be safe. Do a load and it should drop to 1.35v or less if it s correctly being applied.

Remember its about amperage not volts. 1.45v full load is going to generate more heat, even if at 105watts its pulling less current. Processors are weird when it comes to power mathematics.

Because one would think to reach a TDP of 105watts you would want to run 105volts LOL that would be 1 amp of current, but instead we feed processors 1.45v and its like 75 amps at 105 watts tdp. Then we drop the voltage to 1.3 and its like 80 amps to reach 105w tdp. We go even lower and say 1.25v @ 105watts TDP and the current jumps even more to 84 amps.

However we find that the more power phases we can divide the load across we can keep those amps at bay by splitting the work across 8 or 10 or 12 or even now 16 phases.

It is interesting that processors increase in heat as more voltage is applied as we still approach TDP but it turns out that less voltage applied is reduced thermals and I believe we should be factoring in resistance of the silicon and copper and all the materials that comprises the processor. It turns out that lower voltage decreases thermals because of the whole R being thrown into the equation. So it isnt so much the amperage that is causing the heat but the resistance to that amperage at higher voltages that seems to be culprit.

Reduced voltages at load is what matters. Not idle. Idle is not pulling much current, thus resistance doesnt generate electromotive friction, thereby a higher voltage under load is the most important thing to curtail.

That's reassuring to hear; but do you also agree that there is still definitely an issue? Der Bauer's latest video talks about how ryzen 3000 is incredibly efficient at idle; and can even acheive idle at 0.2 volts; yet at stock settings in windows are seeing high idle frequencies associated with abnormal amounts of heat for an idling processor doing no work.
 
That's reassuring to hear; but do you also agree that there is still definitely an issue? Der Bauer's latest video talks about how ryzen 3000 is incredibly efficient at idle; and can even acheive idle at 0.2 volts; yet at stock settings in windows are seeing high idle frequencies associated with abnormal amounts of heat for an idling processor doing no work.

I absolutely agree
I'm not sure what to do to get to those levels. Mine idles at 0.9v and sits at 35c. But I have a waterblock and a -0.04v offset
 
So not to start a whole new thread but I recently upgraded to a 3900X, and it ranges to approx 0.92v at idle to boosting to about 1.5v, my temps typically seem to range at 35c-37c at idle using the AMD Wraith and Ryzen power plan. Does this seem to be in the correct area where they should be?
 
So not to start a whole new thread but I recently upgraded to a 3900X, and it ranges to approx 0.92v at idle to boosting to about 1.5v, my temps typically seem to range at 35c-37c at idle using the AMD Wraith and Ryzen power plan. Does this seem to be in the correct area where they should be?

Yes
 
So not to start a whole new thread but I recently upgraded to a 3900X, and it ranges to approx 0.92v at idle to boosting to about 1.5v, my temps typically seem to range at 35c-37c at idle using the AMD Wraith and Ryzen power plan. Does this seem to be in the correct area where they should be?
That sounds like many other's experiences; although many think that boost voltage is too high; and it is currently being researched by AMD. Are you running a fresh copy of windows? I'll be interested to hear if once you get steam and all of your normal programs running if your chip still drops down to .92v idle as often. Once I installed steam, onedrive and a few other items; my cpu never idles (unless I set a power saving plan) and pings around at high clocks all the time with voltages rarely below 1.44v temps are fine, as I'm on 3x360 radiator WC setup but there is a lot of excess power being pulled from the wall. a true idle at lower volts would have a temp of 35C, but when its pinging around in the 1.4v+ range I'm seeing low to mid 40s constantly. load temps reach 79C
 
A no load idle is ok. It can be 1.5v at idle and be safe. Do a load and it should drop to 1.35v or less if it s correctly being applied.

Remember its about amperage not volts. 1.45v full load is going to generate more heat, even if at 105watts its pulling less current. Processors are weird when it comes to power mathematics.

Because one would think to reach a TDP of 105watts you would want to run 105volts LOL that would be 1 amp of current, but instead we feed processors 1.45v and its like 75 amps at 105 watts tdp. Then we drop the voltage to 1.3 and its like 80 amps to reach 105w tdp. We go even lower and say 1.25v @ 105watts TDP and the current jumps even more to 84 amps.

However we find that the more power phases we can divide the load across we can keep those amps at bay by splitting the work across 8 or 10 or 12 or even now 16 phases.

It is interesting that processors increase in heat as more voltage is applied as we still approach TDP but it turns out that less voltage applied is reduced thermals and I believe we should be factoring in resistance of the silicon and copper and all the materials that comprises the processor. It turns out that lower voltage decreases thermals because of the whole R being thrown into the equation. So it isnt so much the amperage that is causing the heat but the resistance to that amperage at higher voltages that seems to be culprit.

Reduced voltages at load is what matters. Not idle. Idle is not pulling much current, thus resistance doesnt generate electromotive friction, thereby a higher voltage under load is the most important thing to curtail.

A good analogy might be a water pipe, where water sits in the pipe @ 50psi but its not flowing anywhere. As soon as the valve is open the pressure drops to 40psi but the current is very high through the downstream pipes and faucets. I cant think of an analogy for the thermals right now but I am sure I can come up with one later.

So voltage is high at idle because the electron pressure potential (something I made up for the sake of analogy) is higher until the valve (processor pipelines) lets the electrons flow (current) through the silicon. You will see a voltage reduction under load even if the current is actually high.

Someone save me before this ends in disaster hahaha

I get that it won't kill my CPU right away, but you can absolutely ruin a CPU from too much voltage even in the absence of excessive current draw. It's not something i'm worried about in the short term, but it is something I would like for AMD/Board manufacturers to address. I'm all for throwing efficiency out the window for performance but in this case, it's throwing efficiency out the window in a no load situation which really only serves to use more power than necessary and add more heat soak to my AIO
 
I get that it won't kill my CPU right away, but you can absolutely ruin a CPU from too much voltage even in the absence of excessive current draw. It's not something i'm worried about in the short term, but it is something I would like for AMD/Board manufacturers to address. I'm all for throwing efficiency out the window for performance but in this case, it's throwing efficiency out the window in a no load situation which really only serves to use more power than necessary and add more heat soak to my AIO

I absolutely disagree. If what you said were true, why does AMD include an 'AMD Ryzen High Performance' power plan that keeps the CPU state to 100%, meaning a constant voltage all the time between 1.4-1.5v. (No different than just running the 'Windows High Performance' power plan)

Again, the voltage in these low-load situations is NOT an issue.
 
That sounds like many other's experiences; although many think that boost voltage is too high; and it is currently being researched by AMD. Are you running a fresh copy of windows? I'll be interested to hear if once you get steam and all of your normal programs running if your chip still drops down to .92v idle as often. Once I installed steam, onedrive and a few other items; my cpu never idles (unless I set a power saving plan) and pings around at high clocks all the time with voltages rarely below 1.44v temps are fine, as I'm on 3x360 radiator WC setup but there is a lot of excess power being pulled from the wall. a true idle at lower volts would have a temp of 35C, but when its pinging around in the 1.4v+ range I'm seeing low to mid 40s constantly. load temps reach 79C

No my current installation of Windows 10 is about 4 years old with all kinds of stuff installed. I'm still dropping down to .92v very frequently, even with Chrome open and Steam running, plus various other programs like Afterburner, Icue and G Hub.
 
I absolutely disagree. If what you said were true, why does AMD include an 'AMD Ryzen High Performance' power plan that keeps the CPU state to 100%, meaning a constant voltage all the time between 1.4-1.5v. (No different than just running the 'Windows High Performance' power plan)

Again, the voltage in these low-load situations is NOT an issue.

What exactly are you absolutely disagreeing with?

I didn't say 1.5v is going to destroy the CPU, I said excessive voltage can easily destroy a CPU even in the absence of excessive current. Why do you think ESD is dangerous to components?

At the end of the day, AMD wants people to submit cases where this is happening, so you can disagree all you want, but seems like the designers of the CPU agree that it isn't normal.
 
Last edited:
What exactly are you absolutely disagreeing with?

I didn't say 1.5v is going to destroy the CPU, I said excessive voltage can easily destroy a CPU even in the absence of excessive current. Why do you think ESD is dangerous to components?

At the end of the day, AMD wants people to submit cases where this is happening, so you can disagree all you want, but seems like the designers of the CPU agree that it isn't normal.

Well ESD, when Electro Statics do indeed discharge, they are in the range of tens of thousands of volts.
 
Well ESD, when Electro Statics do indeed discharge, they are in the range of tens of thousands of volts.

And thankfully, virtually no current, or we'd all be dead ;)

Either way, Ryzen Balanced power profile shouldn't be running at nearly 1.5v under a no load condition
 
Back
Top