Post your GFXBench Metal results

[pastrychef]

Seeing as how Apple has pretty much abandoned OpenGL and moved on to Metal, I feel old benchmarks like Unigine Heaven and Valley don't accurately represent what works best with macOS anymore. Instead, I think a benchmark that compares Metal performance would be far more useful for macOS users.

GFXBench Metal is free from the Mac App Store.

Here are my results:
CPU: i7-8700K @5GHz
GPU: MSI RX Vega 56 Air Boost
Test ran at 4K resolution



Edit:
Much better results with RadeonVega56AirCoolingFansFix.kext.

 

[pastrychef]

From my HP Elite 8300:
CPU: i7-3770 @3.40GHz
GPU: Nvidia GTX 745
Test ran at 1080p resolution

 

[pastrychef]

From my Dell Latitude 3450
CPU: i5-5200u @2.2GHz
GPU: Intel HD5500
Test ran at 1600x900 resolution

 

[zipb]

Here are the results of my RX480 @ 4K. First time it ran full blast in OSX, with fans revving up.

 

[ainz]

This is a 4GB GTX980 running on 10.13.4 with latest NVIDIA .30.107 on 4K desktop...



Pleasantly surprised at how this compares against the AMD results posted above... was expecting worse.

 

[crazyallstar31]

Here are my results:
CPU: i7-4790K @4.8GHz
GPU: Sapphire Rx Vega 64 Air
Test ran at 4K resolution
MacOS 10.14.1 (18A389)

 

[pastrychef]

Apple has depreciated OpenGL and OpenCL.
https://developer.apple.com/macos/whats-new/#deprecationofopenglandopencl

 

[dfritchie]

Looking at the results posted here, and my own testing shows all on screen test are capped at 60 fps. In my mind this is a worthless benchmark.
I.E. my hd 7950 gets the same results as my 8 gig RX 580......

 

[pastrychef]

Looking at the results posted here, and my own testing shows all on screen test are capped at 60 fps. In my mind this is a worthless benchmark.
I.E. my hd 7950 gets the same results as my 8 gig RX 580......

I remember the app itself recommends comparing offscreen results.

 

[dfritchie]

So its only good for compute background tasks, if you want to know how well your system will run say, metal based games, you are out of luck.