Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

RTX3070 only using 30%

Currently running wu 18201. When I checked in on it with afterburner, I saw that my gpu is only running at about 750mhz and 30% of it's power. This isn't the first time I have noticed this on this work unit. I am not being limited by anything I can control. My CPU only shows 5% utilization. 

 

Anyone else have this issue? I am thinking it's triggering something to do with LHR possibly. I don't see anything in the logs to say it's not acting normally. I am usually pulling down 3-4 million ppd. 

F@H LowPPD.PNG

Link to comment
Share on other sites

Link to post
Share on other sites

Thermal throttling perhaps? What are your temps?

OP: My PC cuts off and crashes. Is the 1000W PSU enough?  ME: 1000W is plenty, PSU or system board could be faulty. Get both tested.

OP: I don't think so, both parts were purchased last week.        😒  Why bother asking if you second-guess the advice?

Friends don't let friends use Apple.                                              NOTE: Recently diagnosed autistic. If I ruffle your feathers, please be patient.

Link to comment
Share on other sites

Link to post
Share on other sites

On 4/18/2022 at 7:52 PM, An0maly_76 said:

Thermal throttling perhaps? What are your temps?

As you can see in the top left of the photo, Afterburner is reporting temps of Mid 30's. Precision X1 also states such temps. I am running custom fan curves, and the fans shut off below 40c. So that's on passive cooling. 

Link to comment
Share on other sites

Link to post
Share on other sites

8 hours ago, cowboy2199 said:

As you can see in the top left of the photo, Afterburner is reporting temps of Mid 30's. Precision X1 also states such temps. I am running custom fan curves, and the fans shut off below 40c. So that's on passive cooling. 

That only shows what the sensor sees, and sensors can malfunction. I've seen before in some situations that a temp sensor not in good contact with what it's there to measure, it will read notably lower than actual temp. For example, with no coolant in the radiator, a car's temp gauge can read substantially lower than the engine is really experiencing. I would really suggest getting / borrowing one of those infrared / laser temp sensors to see what your actual temp is.

 

You haven't mentioned CPU. I should think if this is a work machine it would have something fairly to moderately powerful, but what exactly does this machine have for a CPU? Because if the CPU can't keep up with the GPU, it will bottleneck it and slow it down. It doesn't seem so, but here's a theory. If the CPU isn't properly thermal pasted and is throttling, that would cause such a situation. So I would get a bead on your CPU temp, preferably with an IR / laser as previously mentioned.

 

If this is a Ryzen, have you checked to see if PBO is enabled?

 

Also, yes, it is possible that something is triggering the LHR. Probably the mostly likely scenario here. Apparently it's not fool-proof, it has been noted that the LHR can be triggered if the card thinks certain code is mining code. One other thing, if your display is limited to 60hz, it's necessary to enable things like vertical synchronization. This limits the GPU as well and makes it "sandbag" as it were.

Edited by An0maly_76
Revised, more info

OP: My PC cuts off and crashes. Is the 1000W PSU enough?  ME: 1000W is plenty, PSU or system board could be faulty. Get both tested.

OP: I don't think so, both parts were purchased last week.        😒  Why bother asking if you second-guess the advice?

Friends don't let friends use Apple.                                              NOTE: Recently diagnosed autistic. If I ruffle your feathers, please be patient.

Link to comment
Share on other sites

Link to post
Share on other sites

  • 2 weeks later...
On 4/20/2022 at 5:14 AM, An0maly_76 said:

That only shows what the sensor sees, and sensors can malfunction. I've seen before in some situations that a temp sensor not in good contact with what it's there to measure, it will read notably lower than actual temp. For example, with no coolant in the radiator, a car's temp gauge can read substantially lower than the engine is really experiencing. I would really suggest getting / borrowing one of those infrared / laser temp sensors to see what your actual temp is.

 

You haven't mentioned CPU. I should think if this is a work machine it would have something fairly to moderately powerful, but what exactly does this machine have for a CPU? Because if the CPU can't keep up with the GPU, it will bottleneck it and slow it down. It doesn't seem so, but here's a theory. If the CPU isn't properly thermal pasted and is throttling, that would cause such a situation. So I would get a bead on your CPU temp, preferably with an IR / laser as previously mentioned.

 

If this is a Ryzen, have you checked to see if PBO is enabled?

 

Also, yes, it is possible that something is triggering the LHR. Probably the mostly likely scenario here. Apparently it's not fool-proof, it has been noted that the LHR can be triggered if the card thinks certain code is mining code. One other thing, if your display is limited to 60hz, it's necessary to enable things like vertical synchronization. This limits the GPU as well and makes it "sandbag" as it were.

Thanks for the reply. My apologies for not responding sooner. I posted this issue in multiple F@H forums. The one over on EVGA (which is what my card is), is where I focused my t/s.

 

So update... I rolled the driver back, but as soon as I paused F@H to play a game. It resumed the issue again. I reinstalled F@H and same thing. Pause F@H and when I resumed folding, it throttled down. So I left the older driver installed, and now I just quit F@H when I want to play a game, and restart it after I am done playing. When I do this. F@H resumes folding the current WU at 100% of card capable speed. Currently 1950-1965Mhz and 7GHz on memory after +200Mhz. For some reason the memory on this card likes to run at 6800Mhz instead of the advertised 7000, but I can manually set it to that easily and when folding the card doesn't go over 64c. 

 

My CPU is a Ryzen 5600x just fyi..

 

I included a screen shot of Afterburner now, when folding. So didn't really SOLVE my issue, but found a work around for it.

ab running fah.jpg

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share


×