Low Scores with RX 480

[zentorx]

Guys, maybe you can help out here.

I was running a few benchmarks after I got my RX 480 8GB (reference) to see the performance compared to people who had the same card.

*

But I am getting lower scores than people with the same card (using the same clocks), as an example, this is what I got running Vantage at 1415/2150MHz:

 

AMD Radeon RX 480 video card benchmark result - Intel Core i7-3930K Processor,ASUSTeK COMPUTER INC. SABERTOOTH X79

 

*

A guy on hwbot using the just about the same clocks got over 52k graphics score!

 

That is a difference of 7000 points in the graphics score!

*

I even made a fresh installation of W7 and W10 but it didn't help. I tried using the same drivers as him but it didn't make a difference.

 

This lack of performance is in every benchmark I made, including 3DM 11 and Firestrike.

 

What could be wrong?

Thank's in advance

[buildzoid]

did you disable tesselation?

[speed.fastest]

3DMark Vantage is very CPU bound, if you compare with like 6950X, you had no chance.

[zentorx]

did you disable tesselation?

 

No I didnt' because a friend of mine who has a 1060 gets around 52k graphics score without messing about with the drivers.

Is Vantage this much Nvidia "friendly"?

 

3DMark Vantage is very CPU bound, if you compare with like 6950X, you had no chance.

 

But I'm only talking about the graphics score, my 3930k does not limit the 480 performance

[crustytheclown]

Vantage is CPU bound due to it's low resolution.

If you and your friend want to compare cards try a newer benchmark like Time Spy.

[phobosq]

It is CPU bound, 6950x on air beats 6600k on LN2 by far.

[BurningFormula]

What the others try to explain is that your graphics score is also influenced by your CPU, aswell is the CPU score somewhat influenced by your GPU. the later is not an issue in most cases but the first is very typical for vantage. For reference watch Buildzoids video about his try to reach the #1 wr in vantage with a 7940X and 4 Furys. It was all down to CPU power at his point, as the benchmark could utilize 16 cores and he "just" had 14 under LN2.

The score is influenced so heavily by the CPU because it needs to feed the data fast enough to the gpus for them to max out their performance. The older the benchmark, the fewer the power is needed for the gpu to render a frame and so the faster it needs the new data to keep pushing. If you want a CPU indipendent benchmark, take Unigine Superposition, that has been about not CPU limited at all in my past. You can also use time spy, that is more CPU depending but as its newer, its not that extreme as vantage.

Edited July 25, 2018 by BurningFormula