According to the latest report, in addition to the acquisition of Linutronix, a company known for real-time (RT) kernel patches and other contributions, and the acquisition of Codeplay Software in June, Intel today also acquired another development team – ArrayFire, to further their ambitious software business.
ArrayFire is designed to provide a variety of parallel computing and GPU computing software, including C/C++/Fortran providing GPGPU extensions, which can execute commands across various accelerators from multiple vendors as well as CPU-based.
JOIN XIAOMI ON TELEGRAM
ArrayFire has previously been partially funded by DARPA. It has reportedly been used by various defense industry customers. ArrayFire has been an open source library for over 5 years and they also provide OpenCL and NVIDIA CUDA consulting services.
Intel acquired Codeplay this summer to bolster their work on oneAPI, which supports NVIDIA and AMD, given the company’s work on SYCL computing. Intel followed a similar path with the acquisition of the ArrayFire software team and further strengthened the oneAPI ecosystem.
Intel announced that the ArrayFire team has joined Intel to focus exclusively on oneAPI: “At ArrayFire, the team pioneered ways to efficiently accelerate computing. They created the ArrayFire library, which contains fast, easy-to-use math routines that rely on a clever runtime to optimize low-level performance details through clever JIT compilation of kernel code and intelligent memory management.
Because oneAPI is a foundational effort to ensure that accelerated computing is open, multi-vendor, and multi-architecture; the team knew right away that we were like-minded. With their pioneering track record and demonstrated skills in bringing open solutions to accelerated computing, we are all excited about the opportunity to join forces to help everyone achieve open accelerated computing.
This team at Intel is now building an oneAPI backend that will allow library users to migrate to oneAPI. This exciting work will further our vision for open accelerated technical computing.”
