[SOLVED] High Voltage CPU

[medster720]

Hello, I am new here and have been looking around the forums for help with my system. My issue is that my cpu's VID is at 1.41 - 1.45 V at idle and during cpu burns at 100% load it drops to 1.36 despite my BIOS being set to 1.31...

I have tried flashing the BIOS to the most current version. I am not overclocking and I am not running any external programs to regulate any voltage settings or overclocking support. Most of my BIOS settings are set to auto with the exception of cpu voltage as research has shown that the board will overpower the cpu on auto. I am using Core Temp to monitor voltage and the bios confirms the same readings of 1.4 + voltage to cpu

my build is as follows

Thread Ripper 1900x
ASrock x399 fatal1ty pro gaming MoBo
Corsair RMX series RM 1000 PSU

Not sure what other data could be useful to help with this issue. I know that 1.4+ v to the cpu it too high and the frequency is 3.890 even though the limit in the bios is 3.8 flat and under load it will go up to 4.0ghz. I could be wrong about the voltage being too high, this is my first enthusiast build and I have had the build about a year now yet somehow tonight I am just now noticing how high the voltage is.

I can provide any details needed just provided what I thought would be prudent to the issue.

Thanks for any help!

 

[Darkbreeze]

What research are you referring to in regard to "overpowering the CPU on auto"?

If there was a problem with the default core voltage settings I'm fairly certain that AMD or ASRock would have immediately reacted to that by releasing a revised BIOS version that made the necessary adjustments to resolve the issue.

If you are not overclocking there is no reason you should not be able to run the default configuration settings. If you are worried, then I'd look at reducing my clock speed rather than reducing voltage. Reducing voltage too far can cause instability, and that's a much bigger problem in most cases.

 

[medster720]

When doing research on my issue a common solution that was posted was to not use auto settings in BIOS. People are stating that caused the board to over power the CPU. That is based off of responses from users and not anything officially posted.

Leaving cpu voltage and clock speed on auto does not seem to make any difference vs me manually setting it. I have the clock speed set to 3.8Ghz but as we speak it is running at 3.890 GHZ and voltage is at 1.4125 and under load it will go up to 4.0Ghz with a VID of 1.36, so basically under heavy load its operating where it should but at idle or just doing simple tasks like surfing the web its voltage is too high.

 

[Darkbreeze]

1.4v to the CPU is NOT too high. AMD specifically states that 1.45v and lower are the standard operating range for the core voltage on this CPU.

Overclocking the 1900X was an exercise in simplicity. We merely adjusted the data rate to DDR4-3200 and set timings at 14-14-14-34. We increased Vcore to 1.39V, well below AMD's recommended maximum of 1.45V, and adjusted the SoC voltage to 1.1V. This proved stable up to 4 GHz during extended stress tests. However, even in the face of unsafe voltages, we were unable to attain a stable 4.1 GHz overclock to match our efforts with the 12-core Threadripper 1920X. Dialing back the memory frequency didn't help, either. Considering that AMD supposedly selects the top 5% of its dies for Threadripper CPUs, you might assume that the clustered active core arrangement comes into play. We only see a 100 MHz reduction, so it's more likely that our retail sample is simply on the lower end of the bell curve.

It's notable that a 4 GHz overclock might actually hamper the 1900X's performance in lightly threaded workloads, since we lose the benefit of a quad-core 4.2 GHz XFR boost.

https://www.tomshardware.com/reviews/amd-ryzen-threadripper-1900x-cpu,5222-2.html

 

[BFG-9000]

After testing many CPUs for overclocking ability over decades, I can say that generally the chips with highest VIDs run coolest when overclocked, but also have the least overclocking headroom because they run smack into the maximum safe voltage sooner. See, all of the chips of the same model use about the same power, but the higher the VID, the lower the amperage required to reach that power level. Fewer amps stresses the the VRM circuitry less so that runs cooler too. Your 180w Threadripper at 1.4v already draws an astonishing 128.6A and lower VID would make that even worse.

So if not overclocking, higher VID can actually be preferable.

I am not a fan of undervolting either, as that large Google datacenter study showed a surprisingly high error rate with unoverclocked machines at stock voltage.

 

[medster720]

My concern is that AMD states for longevity of the CPU not to exceed 1.35v for long duration. 1.4v+ seems too high for not overclocking even if it can handle it.

@BFG-9000 that actually makes a lot of sense and explains a lot. So if I reading this right maybe try overclocking it slightly to decrease my VID?

 

[BFG-9000]

AMD is confident it will last through the warranty period at not only the VID it gave it, but also at the maximum recommended 1.45v.

VID does not change with anything you do. If AMD felt they needed to, due to competition or whatever, they would raise the VID to the highest their engineers thought was safe for three years so they could rate their chips for a faster speed or to improve yields.

VID is the voltage your chip was actually binned at by AMD. ASRock has likely added some extra voltage on top of this themselves, just as other OEM manufacturers usually do. I have successfully overclocked many Dell/HP/Lenovo OEM machines with chips that normally cannot run at that speed at default voltage, so I know they must be running higher than stock even if monitoring software cannot read any sensors on those (they obviously have no voltage adjustments in the BIOS at all). That would also explain why prebuilt OEM machines seem to have a lower error-rate than most "whitebox" ones built with retail components and not overclocked--despite being made of obviously lower quality components.

In case you didn't know, Pegatron/ASRock is a huge OEM motherboard supplier just like ECS/PCChips. These companies specialize in cheaper boards with usually fewer layers and fewer VRM phases so they can sell them for a few pennies less to the big OEMs. They are very good at compensating for the reduced stability this usually means, usually with some extra voltage. Given the positive results of the Google study, I'd say they more than accomplished that.

Note that OEMs buy their chips from AMD for cheaper without the warranty, so they are warranting the CPUs in their systems themselves. And they are OK with the extra voltage.

 

[Darkbreeze]

AMD would not provide a maximum safe spec of 1.45v if they thought that, in any way, it would come back to haunt them as a liability. And it would, if it was a problem.

 

[InvalidError]

The thing to keep in mind with modern CPUs is that in order to maintain stable internal voltages in the CPU cores and internal logic which are becoming increasingly sensitive to noise on supply rails, they have moved from directly feeding 0.9-1.3V directly to logic to using local low-drop-out regulators throughout the die to isolate local circuitry form variations in die/package-level voltages. That's why modern CPUs require 200-300mV more than CPUs from a few years ago, seemingly breaking the trend of newer CPUs operating on lower voltages.

 

[medster720]

Awesome guys, thanks for all the responses!