Intel Developing "oneAPI" For Optimized Code Across CPUs, GPUs, FPGAs & More

Intel's 2018 Architecture Day was primarily focused on the company's hardware architecture road-map, but one of the software (pre)announcements was their oneAPI software stack.

OneAPI is so Intel leaves "no transistor left behind" and is their software stack for providing optimized applications and frameworks across processors, GPUs, FPGAs, and other platforms. It was only briefly covered but sounds akin to the Heterogeneous System Architecture (HSA) and their response to NVIDIA's CUDA ecosystem.

It also appears oneAPI will be open-source, but no code or substantive details were immediately available. It makes sense though given Intel's growing array of FPGA devices, their future Arctic Sound discrete GPU plans, and the need for achieving optimal performance particularly around machine learning workloads.

Hopefully oneAPI will be supporting existing industry standards like OpenCL and SYCL to avoid further fragmentation in this space. But for now it looks like we'll be waiting until some point in 2019 for learning all about it.