• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Identifying GFX card bottlenecks

harr88

Junior Member
My computer is rather old, a Q6600(Aug 2007) running a Nvidia 660Ti in PCIe 1.1.

I don't really have any huge problems with the light gaming I do, but with all the deals lately I've been tempted to purchase a new one, but rather than just impulse buying (bc it's new and shiny) I'd like to see if my gfx card is actually being bottle-necked by my cpu/mobo/ram.

So I'm not sure about any standardized benchmark results, something that's free and I can compare my settings to others 660Ti's (using same settings and resolutions) to determine if my mobo/cpu/ram are impacting my gfx card performance significantly?

Thanks.
 
What resolution do you play at? You should be mostly fine but I would imagine your system is going hold you back some.

Sorry I do not know of anywhere that you can find info to compare against your system other than asking around.
 
Change the resolution lower and see if that impacts framerate. If it does not, then your CPU is bottlenecking. Otherwise your GPU is bottlenecking.
 
Some games are CPU limited, some are GPU limited and some are both.

The best thing to do is look at GameGPU that reviews a wide variety of games and search for the games you play/intend to play.

I can tell you that in most modern games, your CPU is a major bottleneck since at stock speeds it's slower than Phenom II X4 955 3.2ghz by 25%+.

http--www.gamegpu.ru-images-stories-Test_GPU-Simulator-Need_for_Speed_Rivals-test-nfs_proz.jpg


http--www.gamegpu.ru-images-stories-Test_GPU-Action-Assassins_Creed_4_Black_Flag-test-ac4_proz.jpg



http--www.gamegpu.ru-images-stories-Test_GPU-Action-ARMA_III-test-a3_proz.jpg


http--www.gamegpu.ru-images-stories-Test_GPU-strategy-Total_War_ROME_II_Patch_2-rome2_p2_proz.jpg


Blacklist%20proz.jpg


http--www.gamegpu.ru-images-stories-Test_GPU-Action-Battlefield_4-test-bf4_proz_2.jpg


Your GPU is much slower but still in your case I'd get a $30-40 cooler and overclock it to 3.2-3.4ghz. Otherwise, I'd consider doing a platform upgrade to be honest. A GPU substantially faster than 660ti is at least $300 USD.
 
identifying bottlenecks in the CPU can be done with programs like MSI Afterburner.
run a benchmark like Unigine Heaven in windowed mode and along side it MSI Afterburner with the GPU usage graph next to it. you can also look at the Perofrmance tab of the Task Manager to see CPU usage.
the GPU Usage during Heaven should be pegged at 100%. if you see it fluctuating, it means your CPU is too weak to feed the GPU with information.

example: here is BF4 with a CPU bottlenecking an R9 280X

jyh2mnimz2dm.png
 
yes, I made a similar suggestion earlier today, but I added that you must turn off vsync. vsync artificially dampens load on GPUs so you need it off to get an accurate measurement. 🙂

identifying bottlenecks in the CPU can be done with programs like MSI Afterburner.
run a benchmark like Unigine Heaven in windowed mode and along side it MSI Afterburner with the GPU usage graph next to it. you can also look at the Perofrmance tab of the Task Manager to see CPU usage.
the GPU Usage during Heaven should be pegged at 100%. if you see it fluctuating, it means your CPU is too weak to feed the GPU with information.

example: here is BF4 with a CPU bottlenecking an R9 280X

jyh2mnimz2dm.png
 
GPU-Z/Afterburner is the way to go. If you aren't getting above 80% usage most of the time (you do need to test a variety of scenes and scenarios however as it really varies) you are almost certainly held back by the CPU.

Downclocking the CPU and seeing what difference it makes to the games average FPS can also tell you how much impact the CPU is having, because its rarely a linear scaling factor. Looking at GPU usage is the quickest approach but plotting lots of different CPU speeds and drawing a curve through them is more accurate.
 
Back
Top