arrayfire
ArrayFire is a general-purpose tensor library that simplifies the software development process for parallel architectures on CPU, GPU and other hardware accelerators. It presents developers with a high-level abstraction of the data on the accelerator through the ⁇ af::array ⁇ object. CUDA provides cross-platform compatibility for oneAPI, OpenCL and local CPU backends.
arrayfire · Scientific Research & Computing
- Using hundreds of accelerated tensor IT functions, you perform linear algebra and machine learning operations.
- With built-in visualization functions and windows, you display data and simulations on screen.
- On Windows, Mac and Linux, you provide hardware acceleration with CUDA, oneAPI, OpenCL and CPU backends.
- ↑ PositiveHacker News
ArrayFire, a general-purpose GPU library, goes open source
“For some context, ArrayFire is a product of AccelerEyes, which began life selling a GPU booster for Matlab (a product called Jacket). This and today's .NET announcement shows how hard it is to sell proprietary developer tools. I had considered using ArrayFire for some of my own…”
- • NeutralHacker News
ArrayFire, a general-purpose GPU library, goes open source
“Hi! A few questions: - How do you deal with software that has been previously run with coarse grained parallelism, optimized for Multicore/Multinode x86? In my experience, GPGPU porting often leads to a tedious, mostly mechanical conversion from coarse grained to fine grained,…”
My note
More in: Scientific Research & Computing
- scientific-agent-skills★ 55,425 · 5 days ago
- deep-research★ 19,863 · 6 mo ago
- burn★ 16,206 · today
- scipy★ 15,168 · today
- RD-Agent★ 15,152 · 9 days ago
- gonum★ 8,454 · 6 days ago