4 ms·
Not the way to do this. Accelerators already have a common middle layer. https://discourse.llvm.org/t/rfc-introducing-llvm-project-offload/74302/23 https://di
by KingLancelot 3y ago
Not the way to do this.
Accelerators already have a common middle layer.
https://discourse.llvm.org/t/rfc-introducing-llvm-project-offload/74302/23 https://discourse.llvm.org/t/rfc-introducing-llvm-project-of...
- superb_dev 3y agoFrom the thread it looks like AMD and Nvidia have something similar, maybe Intel is wants feature parity
- ColonelPhantom 3y agoIt looks like offload is mostly about generic abstractions for accelerator devices. XeGPU on the other hand seems to be a way of encoding Xe-specific hardware features/instructions, so pretty different? I don't see how these are anything but orthogonal.