Destiny 2: PC graphics benchmark

Lokken86 · 15 Nov 2017 at 23:46

Muhammad.Abdelhameed said:
So you are getting the same locked 60fps with less GPU load, great!
Can you test how far a single GPU gains at 4K max settings/unlocked fps?

He didn't say anything about performance and also that v-sync is on so he wouldn't even know if performance went down which it might have if the gpu usage is less.

Muhammad.Abdelhameed · 15 Nov 2017 at 23:49

Lokken86 said:
Wouldn't you want your gpus to be running at 100%? Something is off if they are not being fully utalised.

Because he is using VSync on an 1080Ti SLI rig. His GPUs are not doing much work.

Lokken86 said:
He didn't say anything about performance and also that v-sync is on so he wouldn't even know if performance went down which it might have if the gpu usage is less.

His 60fps before the driver was achieved through 90% GPU utilization, meaning he is almost maxing out his GPU potential. With the new driver his 60fps was achieved with just 65% utilization! meaning at 90% he is probably going to have 80fps or more. Simple math.

Lokken86 · 15 Nov 2017 at 23:49

z10m said:
are you serious.

if you have 60Hz monitor and like tearing than yeah go for it

Mine is 144hz but I don't actually mind tearign anyway haha. I get get it out my head about using v-sync back int he day when it put the fps to 30 if it couldn't do 60 so I always left it off. I don't think that is still the case.

Lokken86 · 15 Nov 2017 at 23:52

Muhammad.Abdelhameed said:
Because he is using VSync on an 1080Ti SLI rig. His GPUs are literally not doing much work.

His 60fps before the driver was achieved through 90% GPU utilization, meaning he is almost maxing out his GPU potential. With the new driver his 60fps was achieved with just 65%! meaning at 90% he is probably going to have 80fps or more. Simple math.

Don't patronize me about maths. Also you are way off. You can't just assume his fps would be exactly 60 if he takes v-sync off with the old driver and then add on another 20fps. He could be getting 200fps before and 150fps now and it would still read only 60fps with v-sync on.

Muhammad.Abdelhameed · 15 Nov 2017 at 23:57

Lokken86 said:
Don't patronize me about maths. Also you are way off. You can't just assume his fps would be exactly 60 if he takes v-sync off with the old driver and then add on another 20fps. He could be getting 200fps before and 150fps now and it would still read only 60fps with v-sync on.

How can he have 200fps with 90% GPU utilization? You are not making any sense! Think about this for a second before you post.

Lokken86 · 15 Nov 2017 at 23:58

Muhammad.Abdelhameed said:
How can he have 200fps with 90% GPU utilization? You are not making any sense! Think about this for a second before you post.

Oh come on mate I'm not working out the exact numbers for no reason when you get jist of what i'm getting atnusing those numbers.

Muhammad.Abdelhameed · 16 Nov 2017 at 00:03

Lokken86 said:
Oh come on mate I'm not working out the exact numbers for no reason when you get jist of what i'm getting atnusing those numbers.

You come on the math is simple really:

OLD:
60fps Vsync'ed = 90% GPU utilization
Without VSync and at 100% GPU utilization you will probably gain about 5~10% more fps

NEW:
60fps Vsync'ed = 65% GPU utilization
Without VSync and at 100% GPU utilization you will probably gain 20~30% more fps

Try this with your GPU in any game you want. It's straightforward really.

Lokken86 · 16 Nov 2017 at 00:09

Muhammad.Abdelhameed said:
You come on the math is simple really:

OLD:
60fps Vsync'ed = 90% GPU utilization
Without VSync and at 100% GPU utilization you will probably gain about 5~10% more fps

NEW:
60fps Vsync'ed =65% GPU utilization
Without VSync and at 100% GPU utilization you will probably gain 20~30% more fps

Try this with your GPU in any game you want. It's straightforward really.

Lol omg. Words like probably and maths don't mix man. His fps could be much higher. I can go play Minecraft at 1000fps and stick v-sync on to get 60fps if you like. Then I can downclock my gpu to 200mhz and still get 60fps. Do see where I'm coming from?

Rilot · 16 Nov 2017 at 07:16

I really think you are completely mis-understanding what he's saying Lokken86
Yes, of course his FPS is 60 with VSYNC on and yes of course you can't say what his FPS is going to be exactly with VSYNC off but you can make an educated guess by looking at the GPU utilisation.
Put it this way. If I was driving a car that could do 0 - 60 in 10 seconds when using 70% of the available power, I could make an educated guess that it would do 0 - 60 in about 8 seconds if I was using 100% of it's power. It's the same deal here.

JediFragger · 16 Nov 2017 at 07:22

edit - nm, I got mixed up with D2

shankly1985 · 16 Nov 2017 at 07:23

Muhammad.Abdelhameed said:
His 60fps before the driver was achieved through 90% GPU utilization, meaning he is almost maxing out his GPU potential. With the new driver his 60fps was achieved with just 65% utilization! meaning at 90% he is probably going to have 80fps or more. Simple math.

And here is where Nvidia do very well! They have you believe that the benchmark they just release is that improvement in every single part of the game. I very, very much doubt.

shankly1985 · 16 Nov 2017 at 07:31

I set about testing with Destiny 2, using the same benchmark sequence I used in the beta. It's the very first part of the single-player game, which appears to be more demanding than many other areas—limited testing in the EMZ for instance showed 10-20 percent higher framerates. All the flames and other pyrotechnics of the opening sequence appear to be keeping performance in check. In other words, this sequence gives a good baseline for performance improvements, though other parts of the game may show larger or smaller improvements.
By the numbers, GTX 1080 Ti performance improves by 35 percent at 1440p and 41 percent at 4k. The GTX 1070 improves by 30 percent at 1440p and 35 percent at 4k. And the GTX 1060 3GB—a card that really has no right to perform this well at higher resolutions—improves by 27 percent at 1440p and 34 percent at 4k. That's enough to almost reach 30 fps at 4k, which is what console gamers typically get. The new drivers also push the 1070 well above 60 fps at 1440p, and the 1080 Ti now easily handles 4k, whereas previously both were a bit off the mark, particularly with regards to minimum fps.

Besides 1440p and 4k, I also ran tests at 1080p medium and 1080p highest. Surprisingly, there's no improvement at 1080p, and in some cases performance is actually worse. I can see the 1080 Ti framerates not changing at 1080p due to CPU limitations (though I'm using an i7-8700K for the testing), but I figured the 1060 3GB would show improvements at 1080p. It doesn't, but at 1440p and 4k it does.
http://www.pcgamer.com/our-testing-..._source=twitter&utm_campaign=buffer-pcgamertw

Muhammad.Abdelhameed · 16 Nov 2017 at 07:43

shankly1985 said:
And here is where Nvidia do very well! They have you believe that the benchmark they just release is that improvement in every single part of the game. I very, very much doubt.

Yes it is indeed in every part of them game! The problem with NVIDIA GPUs was that max DoF dropped fps hard on them, while it didn"t do the same on AMD GPUs. In fact if you run Destiny 2 with DoF at High instead of Highest, NVIDIA GPUs will be ahead of AMD's. This new driver fixes the situation, and makes the impact of Highest DoF small.

http://www.pcgamer.com/our-testing-..._source=twitter&utm_campaign=buffer-pcgamertw

shankly1985 · 16 Nov 2017 at 07:56

Muhammad.Abdelhameed said:
Yes it is indeed in every part of them game! The problem with NVIDIA GPUs was that max DoF dropped fps hard on them, while it didn"t do the same on AMD GPUs. In fact if you run Destiny 2 with DoF at High instead of Highest, NVIDIA GPUs will be ahead of AMD's. This new driver fixes the situation, and makes the impact of Highest DoF small.

http://www.pcgamer.com/our-testing-..._source=twitter&utm_campaign=buffer-pcgamertw

Read the link, 1440p Ti gets 35% and 41% at 4k

NVidia saying 45% 1440p and 53% at 4k
That is best case, like I said it wont be that high in all cases and is more likely what pcgamer shows in the most demanding part of the game.

Either way it's still a good showing from nvidia.

Muhammad.Abdelhameed · 16 Nov 2017 at 08:18

shankly1985 said:
Read the link, 1440p Ti gets 35% and 41% at 4k

So? These are still very high gains, now NVIDIA looks to be in a considerable lead in Destiny 2 over AMD GPUs, just like the beta.

shankly1985 · 16 Nov 2017 at 08:54

Muhammad.Abdelhameed said:
So? These are still very high gains, now NVIDIA looks to be in a considerable lead in Destiny 2 over AMD GPUs, just like the beta.

Hence "Either way it's still a good showing from nvidia"
What part of that don't you understand?

Gregster · 16 Nov 2017 at 08:56

Muhammad.Abdelhameed said:
So? These are still very high gains, now NVIDIA looks to be in a considerable lead in Destiny 2 over AMD GPUs, just like the beta.

Superb gains in truth and great to see.

Gregster · 16 Nov 2017 at 08:58

shankly1985 said:
Hence "Either way it's still a good showing from nvidia"
What part of that don't you understand?

You are spot on Shankly. If someone is running a Q6600/FX chip or similar, they certainly won't be getting those gains but good that you point it out.

r7slayer · 16 Nov 2017 at 09:10

If i could get a extra 25% out of my Ti on average through D2 im happy. I'm playing at 4k currently and getting locked 72 FPS just in odd places it likes to dipp to just under 70.

chroniclard · 16 Nov 2017 at 09:28

My highly unscientific playing test.....I cant notice any difference. Was getting between 100 and 144 fps, 1440p, 1080, gsync, still getting between 100 and 144 fps. Its exactly as smooth and great as it was before.