GPU Over usage and Voltage Limit

[CoronaBlue]

The Problem: Last night, while playing Dragon Age: Origins, I noticed that the fans on my graphics card were being really loud. I alt-tabbed out to take a look at MSI Afterburner, and my GPU usage was around 90%. This is strange to me, because I'm playing a five year old game that normally doesn't even use more than 60% of my GPU. The temperature was also hovering around 60 C. Shouldn't a higher GPU usage require more power dissipation? All of my games that require 100% GPU usage cause it to run at 75 C, at least.
 
In addition, I noticed that the measurement called "Voltage Limit" was oscillating between one and zero (like a square wave, but not periodic). I couldn't figure out exactly what this was measuring, but I've learned enough from my electrical engineering studies to know that anything called "voltage limit" is probably not something that I want turning on. To test, I played a bit of Warcraft III to see if I could reproduce any of these results. I didn't notice the fans getting loud, but there was a spike in the GPU usage that seemed to be associated with the "voltage limit" suddenly hitting one, and then falling back down.
 
I know that I need to play other games to see if this happens again, but I want to make sure it is safe to do that. It also probably wouldn't hurt to revert back to a previous driver. I only just recently updated to 344.60. I'm not sure if that could be the issue, but I might as well try.
 
So is this my GPU saying GG, or is it too early to really tell? I'm willing to attribute the GPU usage anomaly (if it is in fact an anomaly) to Dragon Age: Origins being really weird on PC, but the voltage limit worries me.
 
OS: Windows 7 64 bit.
 
Specs:
CPU: Intel Core i5-4670K 3.4GHz Quad-Core Processor
CPU Cooler: Noctua NH-L9i
Motherboard: MSI Z87-G43 ATX LGA1150 Motherboard
Memory: Corsair Vengeance 8GB (2 x 4GB) DDR3-1866 Memory
Storage: Samsung 840 EVO 120GB 2.5" Solid State Disk
Storage: Western Digital Caviar Black 1TB 3.5" 7200RPM Internal Hard Drive
Video Card: Gigabyte GeForce GTX 770 2GB Video Card
Case: Corsair 200R ATX Mid Tower Case
Power Supply: Corsair RM 750
Optical Drive: LG UH12NS30 Blu-Ray Reader, DVD/CD Writer
 
BIOS Version: American Megatrends Inc. V1.3, 6/6/2013

[lx_srs]

I don't know how many times I have to say this, but people need to set the font colour to auto for those of us running the night theme.

[Birdhunter]

GPUs like the gtx 770 up- and downclock themselfs according to their needs (loads) and up to a specific limit. These limits can be Temperature, Voltage, etc..

Voltage Limit = 1 means, your GPU has clocked itself higher and reached its max. allowed Voltage.

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/61310-nvidia-geforce-gtx-780-review-5.html

In other words: Your GPU was running under higher loads and therefore clocked itself for maximum power.

I would search if DA:O is poorly opmtimized, since this can lead to ridiculous loads, even though they would not be needed (like Watch_Dogs did).

Edit:

FYI: Since you have a GIGABYTE card, you might want to use their version of Afterburner, called OC Guru II. Not necessary, but I would always recommend software from the same manufacturer first, since it's mostly the most compatible one and reduces the chance of misreadings.

http://www.gigabyte.ch/support-downloads/utility.aspx?cg=3

[CoronaBlue]

GPUs like the gtx 770 up- and downclock themselfs according to their needs (loads) and up to a specific limit. These limits can be Temperature, Voltage, etc..

Voltage Limit = 1 means, your GPU has clocked itself higher and reached its max. allowed Voltage.

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/61310-nvidia-geforce-gtx-780-review-5.html

In other words: Your GPU was running under higher loads and therefore clocked itself for maximum power.

I would search if DA:O is poorly opmtimized, since this can lead to ridiculous loads, even though they would not be needed (like Watch_Dogs did).

Edit:

FYI: Since you have a GIGABYTE card, you might want to use their version of Afterburner, called OC Guru II. Not necessary, but I would always recommend software from the same manufacturer first, since it's mostly the most compatible one and reduces the chance of misreadings.

http://www.gigabyte.ch/support-downloads/utility.aspx?cg=3

Thank you for the informative post. Am I wrong in thinking that a higher load should necessarily lead to a higher running temperature?

 

Dragon Age: Origins runs strangely for a lot of people it seems. It is known that the game has a really bad memory leak. I enabled the Large Address Aware flag to resolve this issue. I've also read that it doesn't properly utilize multiple cores, so that might be an issue. I get stuttering for no apparent reason. I haven't seen anyone say that it is poorly optimized, but I would definitely contend that it has its fair share of issues.

[Birdhunter]

I couldn't get it to run properly on both of my R9 290X, only on a single one, and even then not on max settings (lag, screen tearing, the lot). While playing, I didn't notice the memory-leak, but then again, I'm running 4GB vram, so there's more headroom then on a 2gb card, and I did so for mostly under one hour.

Regarding Temperature:

The higher Temperature of a chip is mainly caused by higher voltage (physics basic U=R*I and P=U*I), which in turn is necessary to keep it stable at higher clocks, which are raised when it needs to calculate more/faster, if it wants.

So its basically:

Load->Clock->Voltage->Heat

However, sometimes a program can cause the clocks beeing raised, even though it's not necessary, and your GPU runs higher clocks on 20% load, or lowers the clock because of the heat and then runs slow, despite/because 100% load.

In other words:

When monitoring your GPU, you want to compare Clock, Voltage, Load and Temperatures to see, if there's an anomalie. If one goes High and the other ones stay low, thats when somethings fishY.

TL,DR:

Higher Loads MOSTLY lead to higher Temperatures, but in reverse, higher Temperatures are not always caused by higher GPU-loads. Like everything PC, it depends on a lot of factors and therefore shouldn't be generalized, but looked at on a case to case base.