GTX 1080 - low FPS compared to common benchmark tests

[Apophis2k]

So my problem is getting sub 60 FPS in games like GTA V, Anno 2205 at full HD + Ultra settings (MSAA turned off, for GTA V grass settings also normal). GTA V: everything else maxed out including draw distance and extended draw distance. Frame scaling off.

According to benchmarks, the FPS should be averaging 120 for Anno 2205 and over 100 with GTA V. In both games, I'm getting at best 90 FPS but most of the time 60-70, occasionally dripping below 50. VSYNC off for both games, from NVIDIA control panel it is set for the application to decide.

The only game that runs like it's "supposed to" is Battlefield 4, getting constant 120-150 FPS on ultra @ 1080p.

Here's my setup:

GPU: KFA (aka Galax) GTX 1080 EX OC
CPU: i5 3570k @ 4,5 GHz (gaming temps around 60°C, IBT maximum stress test max temp 76°C after 20 runs)
RAM: 8GB 1600 MHz Kingston DDR3
Mobo: MSI Z-77A GD55 (had to flash the bios to make my GTX 1080 work)
PSU: Corsair TX750 750W

I have the latest game ready drivers by NVIDIA (relesed Oct 28th) and did the clean install. Before installing my GTX 1080 drivers I used Display Driver Uninstaller in safe mode to clean up old NVIDIA drivers.

I'd expect a lot more from a GTX 1080 given that it should be able to run 4K decently with some settings toned down. Playing GTA V the processor usage occasionally hits 100% but it doesn't always correlate with FPS drops, the game still runs smoothly. What could be wrong?

Thank you!

From NVIDIA control panel settings I turned on 'prefer maximum performance'

 

[sizzling]

I do suspect its your cpu. The benchmarks are showing anything less than a 4790k or 6700k paired with a 1080 can bottleneck the 1080 in many modern titles.

 

[Apophis2k]

http://www.tomshardware.co.uk/answers/id-3223461/gtx-1080-low-fps-games.html

Check this post, he has a 6700K and low fps in CS:GO...

 

[Illyria]

This may sound stupid but it sounds like you have v-sync on, you should turn it off if you want more POWER!!!

 

[sizzling]

He had low fps because he had a power saving mode, I don't see your point?

 

[Illyria]

He had low fps because he had a power saving mode, I don't see your point?

Oh i am stupid then

 

[sizzling]

No your not, you are just learning. We all started knowing nothing about this stuff at one time.

 

[Illyria]

No your not, you are just learning. We all started knowing nothing about this stuff at one time.

I mean im stupid because i didnt read all of the question

 

[Apophis2k]

I don't think the cpu (OCed to 4,5 GHz) can explain a 50% drop in fps compared to benchmarks. I'm clueless what to do!

 

[sizzling]

It could, compared to a modern i5 the amount of work per clock cycle that it can do will be much less.

 

[RobCrezz]

It could, compared to a modern i5 the amount of work per clock cycle that it can do will be much less.

No it wont. a 3570k @ 4.5ghz is easily a match for a i5 6500.


OP. Have you checked your CPU and GPU utilization in the other games? Have you tried setting the GPU to max performance under the power options in nvidia control panel?

 

[dontlistentome]

I did a quick search on "i5 3570k" and I see a review from 2012. Regardless of how much it's overclocked, I'd be surprised if a four year old CPU could be expected to drive a brand new top of the line videocard in the same way a new top of line CPU would. A simple way to check is to compare benchmarks of CPUs only. FInd benchmarks of newer CPUs where the i5 3570k was also tested and see the difference.

 

[sizzling]

It could, compared to a modern i5 the amount of work per clock cycle that it can do will be much less.

No it wont. a 3570k @ 4.5ghz is easily a match for a i5 6500.


OP. Have you checked your CPU and GPU utilization in the other games? Have you tried setting the GPU to max performance under the power options in nvidia control panel?

I highly doubt it would keep up and a i5 6500 will bottleneck a 1080, even an OC'd 6600k can in some games

 

[Apophis2k]

Seems that I have found the (most likely) solution myself.

I believe it's just the missing info on how the benchmark tests have been done. In most of the benchmark tests I've found they only mentioned "max detail, 4xMSAA, 1080p" and no mention whatsoever of the advanced settings - that is very misleading. So I turned them (extended shadows distance and ext. distance scaling off) with the other settings left 'on'. This resulted in a rough average FPS of 120ish at 1080p. Also in most of the tests they didn't mention what grass setting was used. The grass on anything higher than normal is known to bring any GPU to it's knees - it's probably bugged?

The advanced settings have a huge impact on how the game runs. There's not much benefit from benchmark tests and comparisons if all of the settings aren't mentioned in detail so everyone can try the exact same settings themselves.

What comes to the CPU and GPU utilization asked by RobCrezz, my GPU utilization often reaches 100% and my CPU hovers around 70-95%, sometimes hitting 100% on all cores but that doesn't always correlate with the times the FPS drops. Weird. Seems like the only thing that significantly effects the performance of the game with my setup is the advanced graphics settings.

All that being said, I still suspect there probably still is a CPU bottleneck with this game (advanced graphics) using my current CPU, even overclocked. Waiting for Zen/Kaby Lake.

Thank you everyone who took the time to look into this and hope it helps other people as well!