renderstate
Senior member
- Apr 23, 2016
- 237
- 0
- 0
It could very well be the case. I don't think NVIDIA discussed at what granularity they can switch between graphics and compute tasks and I agree that a finer granularity is likely to bring more performance to the table.To my understanding Pascal can now switch between Compute and Graphics at the GPC boundary, which definitely is a form of concurrent compute + graphics. I'm just not entirely convinced of the usefulness when you can only switch between the two in 25% granularity (GP106 might be limited to 33% or 50% granularity, GP102 could increase this to 17%).
In comparison, AMD can switch between the two workloads in >3% granularity on Hawaii to stuff holes in the graphics pipeline. This seems a lot more useful to be honest to fill both expected and unexpected gaps in the graphics workload.
I am just trying to stop this non-sense of Pascal not supporting async compute. Perhaps AMD has a better implementation, which is hardly surprising giving they refined this capability over a number of years.