Directx weirdness

[Avatarian]

So I have an OC'd Gtx 1060 GB and so I expected my Directx scores to be a bit above the average for the model. And for Directx11 I was right, but for Directx9 and 1o it is far lower than the average for the model and for directx 12 it is slightly lower. I'm not too concerned about the first two but I would like to at least figure out the issues with Directx 12. Thank You.

I'm running a I7-6700k at 4.6 Ghz and my GTX 1060 is OC'd to 2139 Mhz stable at 1.05v above 60C

[Tim (PassMark)]

What results are you seeing for the individual 3D tests?
Are you using the latest drivers for the card?
Have you tried running the test while not overclocked?

There is also this thread that contains many causes of 3D test slowdown, as you're using an Nvidia card Shadowplay has been known to cause issues with the benchmark scores.

[Avatarian]

What results are you seeing for the individual 3D tests?
Are you using the latest drivers for the card?
Have you tried running the test while not overclocked?

There is also this thread that contains many causes of 3D test slowdown, as you're using an Nvidia card Shadowplay has been known to cause issues with the benchmark scores.

Yes I am using the current Nvidia Driver

Ok, so with Full OC I got DX9=105, which is still very low, DX10= 181 which is better than what it was before but still not as high as I'd expect given DX11= 222 which is better than most other gtx 1060s. Finally DX12=34 which like DX 11 is still a bit lower than I'd expect. Finally GPU Compute was 5241

When I set the gpu to regular clock speed of about 1900 Mhz at 1.05v I got DX9=105 yet again which is strange. Then I got DX10=174 DX11= 214 DX12=32 and GPU Compute was 4840 which is a minor decrease I would have expected if the previous oc was stable.

[David (PassMark)]

DX9 test is likely CPU bound as it runs at a low resolution. So the overclock on the GPU won't do much. Doesn't really explain why it is in 100fps range however, unless you have some CPU issue.

What monitor resolution are you running at? The DX12 test prefers to run at 4K. Anything less than this results in a penalty.

The DX12 test is new, so we don't have a lot of results for it yet. I suspect you 32 - 34 fps value is about what is to be expected for this card with the final PerformanceTest V9 release.
(the earlier beta releases might have scores slightly higher, around 36fps)

[Avatarian]

DX9 test is likely CPU bound as it runs at a low resolution. So the overclock on the GPU won't do much. Doesn't really explain why it is in 100fps range however, unless you have some CPU issue.

What monitor resolution are you running at? The DX12 test prefers to run at 4K. Anything less than this results in a penalty.

The DX12 test is new, so we don't have a lot of results for it yet. I suspect you 32 - 34 fps value is about what is to be expected for this card with the final PerformanceTest V9 release.
(the earlier beta releases might have scores slightly higher, around 36fps)

Ok since I'm using DSR that explains less on DX12, that just leaves DX9, I tried putting the cpu at stock and the fps just dropped 2 points. All the CPU tests I ran looked pretty normal, the only odd thing is the prime numbers score is slightly below average at 31, any guesses what may be causing the problem?

[David (PassMark)]

Does your monitor have freesync/G-Sync? As these have been seen to cap frame rates.

[Avatarian]

Does your monitor have freesync/G-Sync? As these have been seen to cap frame rates.

No, it is a tv screen, but it doesn't seem to affect the other tests. Could any nvidia settings be only affecting the DX9 test. It isn't too important at this point, I won't need more than 100 fps on DX9, but I want to make sure there isn't an underlying problem.

[David (PassMark)]

There is another post here with a similar low DX9 scores.
What is the refresh rate of your monitor/TV?
What video card device driver version are you using?

[Avatarian]

The Refresh rate is 60hz so I don't think that is it, and I'm using the current driver. My guess would be according to the other post that there is something about the 10 series that makes it perform lower on DX9 and DX10 tests, I'm not sure this is on Nvidia side which would be bad, or if this is just a problem with your test. Either way thank you so much for the help and I'd love to know if you guys ever get to the bottom of this.

[Canniballistic]

I'm having nearly the exact same experience.

I'm using very similar baselines compared to my own system so they make me think theres something wrong with my card or software and not the GTX 10xx series in general.

In practice as well the card seems to be under performing so I would love to find out what the issue is.

[David (PassMark)]

Canniballistic,

See Tim's initial post above for thing to initially look at.

[Canniballistic]

Canniballistic,

See Tim's initial post above for thing to initially look at.

Thanks.

Yeah I have done a bit of research into it and I believe I've ruled out drivers, overclocking, CPU bottle necking and my monitor from being the cause.

[David (PassMark)]

And what are the actual results?

[Canniballistic]

And what are the actual results?

DX9

• I'm getting 57 fps stock and no improvement with OC
• My baselines range from 161 to 202 fps

DX10

• 183 fps stock with no noticeable improvement OC
• Baselines range from 228 to 279 fps

DX11

• I'm getting 190 fps and 272 OC
• Baselines are 265 to 284 fps

GPU Compute

• I've got 5626 stock and 6010 OC
• Baselines run from 6575 to 7413

CPU benchmarks are comparable except for prime numbers where there is a sizeable performance drop.

[David (PassMark)]

And what hardware do you have?