4

I used the GTA V built-in benchmark tests at 1080p vs 720p, and found that the benchmark results had little difference. In fact, the average fps for 1080p was higher than that of 720p. All other settings were kept the same

GTA V Benchmark Results: 1080p
Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 47.884686, 106.038651, 89.360901
Pass 1, 66.367462, 129.559692, 107.134224
Pass 2, 73.644211, 149.183258, 109.369591
Pass 3, 74.202469, 140.992554, 121.042595
Pass 4, 40.994385, 193.316055, 112.253723

GTA V Benchmark Results: 720p
Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 53.231331, 99.779015, 80.330269
Pass 1, 49.893250, 138.234314, 89.887611
Pass 2, 59.239407, 142.057755, 83.320381
Pass 3, 53.007504, 137.594223, 86.780602
Pass 4, 32.280907, 179.472031, 95.087677

I've double checked these results to make sure I didn't get the two mixed up. The 1080p fps performance in terms of minimum, maximum and average were higher than that of the 720p in almost all cases.

This is unexpected, as 1080p would be pushing 2073600 pixels, while 720p only needs to work with 864000 pixels - less than half of the pixels at 1080p.

I have not done tests in other games, but this really bewilders me. Can anyone explain it?

System specs:

  • CPU: Intel i7-6700, 3.4 GHz
  • GPU: AMD RX 480, 8 GB VRAM
  • RAM: 16 GB
  • OS: Windows 10 Home, 64 bit
  • GTA V DirectX 11 mode

I have the latest graphics drivers from AMD at the time of the test. Also, this is a new pre-built computer that was released late 2016 (Dell XPS 8910 SE)

  • 3
    I would have said you're CPU bottlenecked, but that shouldn't be the case with i7-6700. My best assumption is that mid-range/high-end gpu drivers are better optimized for 1080p than they are for 720p, since they're meant to be used at 1080p resolution. I have seen this kind of behaviour as far back as GeForce FX5200 running better at 1024x768 than it did at 800x600, so I wouldn't cross the possibility out. Are your drivers up-to-date? Your CPU frequency is a bit low, but I guess if your cpu is 6700, not 6700k, there's not much you can do about it. – Elise Feb 8 '17 at 23:46
  • @Chippies I downloaded and installed my graphics driver from AMD, and it was the latest version at the time of the test. That sounds like a possible explanation, but how does that happen? I'm just genuinely curious. – Panpaper Feb 9 '17 at 0:02
  • 2
    Rendering the amount of pixels requires raw power, so you'd think that 720p should provide you with more fps than 1080p, but it's more complicated than that. GPU Drivers can be optimized to use "shortcuts" in the rendering process that allow it to squeeze out extra frames and since 1080p is the most popular resolution and your gpu is more than capable at running games at that resolution, there might be a lot of optimizations that apply to 1080p, but don't or aren't as beneficial for 720p. – Elise Feb 9 '17 at 0:13
  • GTA V seems to take into account how much video RAM your graphics card has, so it might be because the lower resolution allowed it to load more into video memory and so display some objects with more detail. – user86571 Feb 9 '17 at 2:16
  • For troubleshooting purposes, you could lower your desktop resolution to 720p, to match the in-game resolution and benchmark it again. Running games at a different resolution than desktop can negatively affect performance. – Elise Feb 9 '17 at 17:01
1

It's always funny to me when I see things like "bottlenecking" in a troubleshooting post, just because every video card is technically bottlenecked by the PCI-E Bus and motherboard before the processor even gets involved. That's not why this is happening.

Believe it or not, your results are precisely as expected. Once you clear 720p, you've effectively got all of the textures that were meant to load being managed. Texture processing is a large part of rendering, and your card has plenty of VRAM to store the textures. GTA5 doesn't use 4k textures, and I'd wager most of them are actually relatively low resolution, laid across high-poly objects.

The high-polygon objects and features like Anti-Aliasing weigh down the GPU, and that's where you're seeing a difference at all. Seeing as your video card is relatively new and comparable to the nVidia 1060 in performance, and GTA 5 is not by any means a new game, the real reason you're not seeing a difference is because you aren't making the card sweat. Your low-end FPS is still faster than most humans can perceive.

I noticed you didn't post your test parameters themselves. You should do so (all graphics settings listed). Turn AA and texture filtering all the way up, just max out every setting, and then run your comparison.

FYI I had a similar experience with the 760GTX; resolution didn't matter, and only pushing AA really tested it. Once I went up to a 1070, GTA 5 could run maxxed for days and barely make my GPU's fans spin.

  • 2
    When people talk about "bottlenecking", you should read it as not the cpu bottlenecking the gpu, but the gpu portion of code being processed much faster by the gpu than the cpu portion of the code. This answer doesn't really explain why they are getting higher fps at 1080p than they do at 720p. When the GPU isn't being pushed to the limit, fps should be similar, if not identical at all resolutions, which is not the case. They are getting almost 25% increase in performance by doubling the rendered pixel count, which seems counter-intuitive. – Elise Feb 9 '17 at 18:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.