antihelten
Golden Member
- Feb 2, 2012
- 1,764
- 274
- 126
Ahem...
It seems they both support a variation of Async Shaders, AMD's is capable of handling more before performance drops, Maxwell 2 can handle up to 31 before performance drops and levels with AMD.
"Maxwell is capable of Async compute (and Async Shaders), and is actually faster when it can stay within its work order limit (1+31 queues). Though, it evens out with GCN parts toward 96-128 simultaneous command lists (3-4 work order loads). Additionally, it exposes how differently Async Shaders can perform on either architecture due to how they're compiled.
These preliminary benchmarks are NOT the end-all-be-all of GPU performance in DX12, and are interesting data points in an emerging DX12 landscape."
https://www.reddit.com/r/nvidia/comments/3j5e9b/analysis_async_compute_is_it_true_nvidia_cant_do/
It's going to be down to developers to code efficiently for all platforms.
The guy in that link (Alarchy) misunderstands what async compute is, he confuses it up with the ability to support multiple compute queues (32 in the case of Maxwell 2), which is a completely separate (albeit complementary) feature.
For a better explanation of the results from the Beyond3D program look at this one (already posted before in this thread I believe):
https://www.reddit.com/r/pcmasterra...e_all_jump_to_conclusions_and_crucify/cumlmwv
