Video card consuming excessive amount of power?

Toggle sidebar Toggle sidebar
B

Bryman

Distinguished
Apr 23, 2013
20
1
18,525
So... my 7970 is consuming 38 amps (460 watts) when running OCCT... I have the power limit set to +20% but it shouldn't be that high... right? It's not even throttling until it hits 90°C...

If my video card is using more than about 15 amps then it'll start throttling at 90°C... otherwise it waits until it goes up to 97°C and even then it doesn't throttle, it just kicks fans to 100%... but the gpu usage graph goes crazy almost like it's trying to change clock speeds, but isn't for some reason

And only one fan runs unless the fan speed is set to 80% or higher, which is making the rest of my video card very hot... I got the VRM temp up to 133°C and it was still going up.. would of gone up to at least 145-150°C if I let it

It also doesn't let me underclock when running folding@home... it's weird, if I have folding@home running the clock and memory speeds do not decrease but with anything else they do

But the card shouldn't be consuming 460+ watts and not throttle even at +20% power limit... right?

8ZZvlm3.png


(The values are a little off but I'm looking at the PCIe #1 which I believe is accurate, that's my video card)

I'm pretty sure 460 watts is more than a 7970 should take... my video card is pretty messed up
 
Could be monitors, or such at higher res and higher refresh rate. Or just clean install all video drivers like me and it should go back to normal. Mine had the same prob,me for a while, not as bad as you, and I reinstalled the drivers and it fixed it.
 


I ran DDU a few weeks ago and it didn't fix these issues, unfortunately
 


Well, when I was running OCCT I had my video card clocked to it's limits but right now it's at stock overclocked speeds (950mhz core 1425mhz memory) with 1.07 volts

VRM voltage out is fluctuating from about 1.03 to 1.05 and the 12 volt on my video card reads 11.6... you know what, I'll just take a screenshot

vOZBOUG.png


It's not running that hard at all right now... I seem to have gotten one of those folding@home WUs that gives you a third the PPD you usually get, so my video card is barely running

Usually it's about 80°C core temp and 100°C VRM temp with these clock speeds and with fan set to 40%

Also, my video card randomly loses power (screen goes black and computer restarts) about once every 3 days.. I probably damaged the VRMs or something :/

And I get vram issues (Smartpointer: can't dereference a NULL pointer) and (Forces are blowing up!) in folding@home... and about once a week my drivers will crash and they will recover and my memory will be corrupted...

https://www.youtube.com/watch?v=M9kqSDqELEY

Yeah... my video card is pretty messed up :/
 
It is abnormally high, the power draw. I also wonder if the info is accurate, take a reading of the power draw from wall or something. Also the added power limit might cause wrong reading? Idk. Maybe it's a bad card, likely...
 
This is when running OCCT

wUfqKUP.png


Hmm.. is Corsair Link inaccurate or is this inaccurate... because this is a 100 watt difference from what Corsair Link was reporting...

Wait... you have to add the VRM GPU core and the VRM GPU VRAM... right? So... it's having about 390 watts of power go into it

Guess that's not too far off from Corsair Link, it was saying 460

Plus I had my memory clocked to 1740 with 1.675 volts when I was reading Corsair Link... but that shouldn't make more than a 10 watt difference

Yeah, any higher and I get memory corruption... which I guess is understandable

I believe my video card is just messed up really bad... it shouldn't be allowing itself to be consume 400/460 watts of power and not be throttling...

And one of the video card fans doesn't run unless fan speed is 80% or higher

And doesn't let me underclock when running folding@home (not sure if that's normal)

And has vram issues and random shut-offs (probably damaged something with the stress tests)

Edit: Just ran it at about the same memory clock as when I took the screenshot with Corsair Link... core wattage increased about 2 and memory by 4... so only increased power consumption by about 6 watts

So still only about 400 watts for HWmonitor and 460 for Corsair Link
 
Thanks for your help, but I think it's just time for a new video card...

This one seems to be failing and for some reason lets itself draw over 400 watts of power without even throttling... and has a bunch of other issues

The video card is over 3 years old... there's much faster ones out now... I was actually looking at getting the 980 TI but was going to wait until the new ones came out, hopefully this video card will last until then... probably should stop running those stress tests and not try to blow up my VRMs with hot temps and excessive current... I doubt 133°C (would of easily gone up to 145-150 if I kept it going) is good for it... actually, the first time I got it up to 130 my screen went black (video card shut off) so definitely too high...
 
Yea it's prob time for a new card... Maybe some power regulation function is broken or something is damaged. Since I've only had my 7970 for 1.3 ish years, I wouldn't know too much problems so far. The newer cards draw less power and produce less heat too and faster. Maybe when pascal and Polaris come out they'll drop prices and getting a used one wouldn't be a bad deal.
 
It may be just the amount of load on the card that is causing the power draw, but the rest are definitely messed up. This guy had a OC i7 and a 7970 drawing ~600w from the wall, still not as much as you... And the temperatures are definitely not safe. The one fan isn't safe either. The power draw may also be due to your memory speed you have, I mean increasing speeds will increase your power draw, but it still shouldn't be 460w. I'm thinking some regulator is broken.
 


When you increase core voltage to max and push the card to the limit with OCCT without any power throttling then yeah, it's insane

I mean... the VRM temp never goes above about 110 unless I purposely get it that high... to get it around 130 I have to get the video card temp above about 90 with 1.3 volts

But yeah, having only one video card fan run on my GPU is perfectly fine for the GPU itself... maybe temps are 75 instead of 70 for the same fan speed.. but it gets the rest of the card very very hot

I had fan speed at 100% and GPU temp was 80 while VRM was 90... then I set it to 80% so only one fan would run, within about 30 seconds GPU temp only increased one while VRM temp increased 10

I tested three different stress test programs... now this was with the same clock speeds and stuff as when it got up to 38 amps but it was when the video card was cooler, which apparently it uses less power when it's cooler

Kombuster only used 28 amps, furmark used 31, OCCT used 33

Yeah, OCCT is insane

Not surprised it's pulling this much power... although I'm not sure why it isn't throttling

The TDP is 250 watts, my video card is consuming about 460

With my video card at 100% fan speed and the side of my computer case off, the video card would get up to about 105-110°C if it didn't start throttling due to temps...

I still find it weird how if my video card is consuming less than 15 amps it'll allow itself to go up to 97°C but if it's consuming more than that then it only lets it go up to 90°C

Idk, something is seriously messed up... 🙁

*sigh*

Actually.. I think I got my VRM temp up to 133°C by running Furmark... it clocked the core down to 500mhz but it still made the VRM very very hot

And my video card shut off at 1.3 volts 500mhz core... I read the voltage becomes very unstable when the VRM gets hot

The first time I ran it I got the VRM temp to 130 and my screen went blank and my computer restarted... then I ran it again and got it up to 133 and the video card stayed on, and then I stopped it before I blew my VRM up

I probably already mentioned this and am just repeating myself
 
So... turns out my video card does power throttle, but it waits 60 seconds after it hits the power limit to do so... and when it does power throttle it looks like this

62Nz5RO.png


Not sure if it's supposed to wait 60 seconds... I can get my video card from 33°C to 88°C at 100% fan speed in 60 seconds or 33°C to 97°C at 0% fan speed in 45-50 seconds

Also I got my power supply to report the amps going up to 40

The difference between Corsair Link and HWmonitor is pretty accurate up to 35 amps... then once Corsair Link reports higher, it starts to spread more apart

From 35 to 40 amps the amps on HWmonitor only increases 3

Of course I know you have to calculate in a 30 watt drop from the power supply to the video card at 40 amps... but even then it's still 20-30 watts off

Also the reason my video card shut off when I saw the VRM temp reach 130 is because the other VRM temp hit 135, which is the point at which my video card shuts off, apparently

The program was only reporting one of the VRM temps and it was the cooler one

Still not sure why my computer randomly restarts sometimes or has vram issues

https://www.youtube.com/watch?v=sVotjOy0k_Y

Stupid Bandicam watermark
 
Also.. sometimes when I minimize and open Chrome, part of the screen is black

z3HduEB.png


I'm assuming that's a video card issue too?

I could run DDU again... but I just ran it about a month ago...
 


I don't have a Hard Drive

Anyway... I removed 15.12 and installed 14.9... I'll see if that fixes the extremely low PPD on 13000s projects on folding@home and the other issues

Well, the one fan not running and power throttling issues are permanent... but besides that

My video card hasn't crashed on me in about 4 days... Well.. It did when I ran 2 kombusters and heaven and furmark and OCCT at the same time... but under normal usage it hasn't crashed in a few days

I know I need a new video card... I was hoping to wait until the new generations come out but I'm not sure my card will last that long/I can put up with the issues until then

Although my video card doesn't seem to be getting worse, so that's good

I haven't really gamed on this card much lately, but I do run folding@home 24/7 (except if I game or do anything video card intensive)
 
So I found out that the low PPD on 13000 projects was due to them taking more cpu usage and I just needed to reduce the cpu folding to 6 threads instead of 7

I also figured out that my video card does not start throttling after 60 seconds when it hits the power limit... it just takes about that long for the VRM temp to hit 110

And I found another one of my many video card glitches.... when folding@home is running I can't underclock my video card ALSO it disables throttling when video card temps get high

My video card usually starts throttling when VRM temp hits 110 or gpu temp stays above 97 for more than a few seconds... (first it tries to force fans to 100% to cool the video card down without throttling)

But... with folding@home running

https://www.youtube.com/watch?v=MEPNzrBoyBM


https://www.youtube.com/watch?v=l_D0ezitJbY


The video card shut itself off when the VRM temp hit 135 and the core temp hit 101

It's a good thing it shut off... VRM temp would of EASILY gone to 160+ if it didn't

Without power limits my video card gets really hot :S

And this is only about 30 amps... my video card can get up to about 35 without folding@home running and just the stress test running


Core temp would easily hit about 110 eventually if the video card never shut off... luckily it does shut itself off

Kind of quickly though.. my 7850 doesn't even force fans to 100% until 102... but my 7950 forces fans at 97 and shuts the video card off at 101
 
Anyways I wish I thought of this sooner, but it couldn't possibly consume more than 375 watts without burning up your PCIE power connectors. The ghz edition has 2 8 pin or 8+6 pin on reference. That means it can only draw some 300-375 watts, but still...
Your GPU is definitely weird when I look at it
 
Nope, the PSU reports 40 amps and my video card reports 35 amps

But it only runs for about a minute

Also... the hotter the gpu the more power consumption.. it starts out at about 30 amps at the video card, and when the temp hits 97 it's at about 35 amps

Anyway... I think my video card is failing even more.. fans decided to go all the way down even at 100%... did this multiple times... my video card shut off at about 105

https://www.youtube.com/watch?v=V0n3guwbmcU

My video card fan decided to run AFTER the video card shut off... *sigh*
 
So... I set fan speed to 100% and after a few seconds the fans became inaudible... I felt underneath the card and felt a small amount of air being pulled in... the fans were just spinning really slow... I slowed it down even more with my finger and the fans instantly went back up to 100%



I tried slowing down the fans again with my finger to get the fan speed back up... didn't quite work very well this time

https://www.youtube.com/watch?v=MQDtgIHuoFI


So.. stopping the fans does help get them started again... but apparently it has to be stopped fast and just for a split second for it to work

https://www.youtube.com/watch?v=8iqs0Eq2fog



Anyway... I took the 7970 out... the core temp kept hitting 104 and shutting off on me... I popped in my 7850

Sooooo, don't have to deal with that glitchy 7970 anymore... The new Nvidia and AMD cards should be coming out real soon.. I'll buy one of those
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom