It might be that there is some startup time to switch the display to the dGPU. By running the game you may perform this switch ahead of time, thus reducing the time to run the benchmark.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
15-inch MBP Geekbench GPU results much worse on 10.12.3
- Thread starter Eldar Gezalov
- Start date
-
- Tags
- geekbench macbook pro
- Sort by reaction score
It might be that there is some startup time to switch the display to the dGPU. By running the game you may perform this switch ahead of time, thus reducing the time to run the benchmark.
I wonder if, perhaps, macOS 10.12.3 is more aggressive about clocking the 460 down unless it's really needed. That way application performance in something like a game that really pushes the GPU doesn't decrease, but maybe Geekbench doesn't force it into "full power" mode?
EDIT: Just tested this myself. Ran GB4 OpenCL test with nothing else running, got 46522. Opened Portal 2 for a minute, went in-game and messed around to try to work the GPU, then cmd-tabbed back to Geekbench and ran it again, got 52204. I think this may be on to something.
Last edited:
Good point, however the OP mentioned they manually set to dGPU (I imagine through GFXStatus or something similar) and the Geekbench results were the same. Perhaps it's dialled down more in 10.12.3 but it only cranks up under bespoke circumstances when it detects it needs more beef from the GPU – though you'd imagine benchmarking would certainly count as that.
The main thing is that @Creep89 isn't seeing any loss of framerate in his game, so whatever the Geekbench results, early testing seems to indicate that real-world performance isn't affected (and that's the main thing!).
