Question GPU only draws about 120 W during heavy load, is this normal?

[DexDacozed]

Hello

So my GPU, RX 5700, only draws usually about 110 - 130 watts during gaming. In FurMark, it peaks at 115 watts. I'm wondering if this is normal, the RX 5700 apparently draws over 200 watts during heavy gaming, Tom's Hardware itself recorded a peak of 202 Watts in FurMark on a 1600x900 monitor. I'm running a 1920x1080 monitor for reference.

Now, I've been having problems with black screen crashes under heavy load and I'm wondering if the GPU isn't able to draw enough power for itself or something like that.


CPU: AMD Ryzen 5 3600
GPU: AMD Radeon RX 5700 DUAL EVO OC
RAM: HyperX (4x8GB), DDR4 3200Mhz
M.2: WD Blue SN500
SSD: WD Blue 500GB
MOBO: Asus Prime B450-Plus
PSU: Corsair RM750x

Thanks for reading!

 

[rcald2000]

Question: Is this a new build or existing? Which GPU did you have on this system before the RX 5700? Was the primary storage moved from an Intel build to this AMD system? If so, was the OS reinstalled, or was the SSD simply moved from one system to the other, without reinstalling the OS? If this GPU was installed into an existing system, did you run DDU (Display Drivers Uninstaller) in safe mode, and download and install a fresh copy of the AMD Adrenaline drivers? Have you made any changes with power limit or overclocking the GPU.

I do not have hands-on experience with the RX 5700, but I assume that a power draw less than the listed 180 watt TDP is not normal. The black screen crashes under load are not normal.

 

[rcald2000]

Addendum What's interesting is that I just looked up the customer reviews of your card on Amazon, and the first review is what I've pasted below. Honestly I believe that simply running DDU in safe mode, and reinstalling the drivers would solve the problem. The review below is from a time period before AMD fixed their drivers.

https://www.amazon.com/ASUS-Overclo...-RX5700-O8G-EVO/dp/B07WRMT3BH#customerReviews

3.0 out of 5 stars NIce card, but if your case doesn't cool well enough it'll black-screen crash.
Reviewed in the United States on February 29, 2020
Verified Purchase
I installed this in an older XPS8500 where the case cooling was great for its time but cannot keep up with modern graphics cards. I have to take the side of the PC case off and lay it down flat to be able to play any game work playing for more than 5-10 minutes.

Rather than buying a different card I bought a different PC case. Wish me luck.

 

[DexDacozed]

Addendum What's interesting is that I just looked up the customer reviews of your card on Amazon, and the first review is what I've pasted below. Honestly I believe that simply running DDU in safe mode, and reinstalling the drivers would solve the problem. The review below is from a time period before AMD fixed their drivers.

https://www.amazon.com/ASUS-Overclo...-RX5700-O8G-EVO/dp/B07WRMT3BH#customerReviews

3.0 out of 5 stars NIce card, but if your case doesn't cool well enough it'll black-screen crash.
Reviewed in the United States on February 29, 2020
Verified Purchase
I installed this in an older XPS8500 where the case cooling was great for its time but cannot keep up with modern graphics cards. I have to take the side of the PC case off and lay it down flat to be able to play any game work playing for more than 5-10 minutes.

Rather than buying a different card I bought a different PC case. Wish me luck.

Hi, thanks for the response!

I actually have uninstalled everything already with DDU, then got the latest drivers. It's a completely new build, and this problem has persisted from the beginning. The temperatures are completely fine when it crashes. What I'm most interested in right now if I could somehow get more GPU to draw more watts and see if that'd clear it. It crashes with default volts, clocks, and power limits. Changing any of those seem to do nothing regarding the crash.

 

[rcald2000]

RAM: Did you simply enable XMP, or did you manually set the timings? In addition to the 24 pin PCIe connector, you also have the 8 pin connector from the power supply plugged into the motherboard? I would recheck CPU temps under load, but if you're certain that they are fine, then you could take a photo of the inside of your build, upload to imgur and share the link here. But if those things check-out then I assume that there is a problem with the GPU, and I would refund/exchange for a different model of (MSI?) RX 5700 XT, or the newer RTX 3060 Ti, if you can find one. Just make sure that you can actually find a replacement card that is comparable in performance, since many GPUs are currently out of stock.

 

[DexDacozed]

1. I have enabled D.O.C.P/XMP, but I've turned it off for testing the crash and it still crashes.
2. Yes, both of those connectors are attached.
3. I have logged data from the crash with HWiNFO, and at the time of the crash CPU temp was 61C and GPU was 80C. I can show the logs if you want.
4. Here's the build:

   https://imgur.com/1aBDKNI

   View: https://i.imgur.com/1aBDKNI.png
   . Not the best picture, I can take a better one if really needed.

Yeah, I'm thinking of RMA'ing the GPU pretty soon.

 

[rcald2000]

There are no guarantees, but were I in your situation, I would return the card to the retailer, since that's usually an easier process than doing a manufacturer RMA. And then I would purchase a different brand and/or model of RX 5700 XT, or maybe try to snag a 3060 Ti, this weekend, if possible. Good luck, and please let me know how you make out.