Narrow down by specifying conditions
There are currently 4292 hits.
Intel: Is it possible to perform half-precision floating point calculations such as int8 and FP16 on FPGA with OpenCL™?
Category: OpenCL™
Tools: Intel® FPGA SDK for OpenCL™
device:-
As a half-precision floating-point support option (pragma) in your kernel code statement:
#pragma OPENCL EXTENSION cl_khr_fp16 : enable
You can compile by giving
Please check the following document for details.
(reference)
The OpenCL Specification Version: 1.0 Document Revision: 48
https://www.khronos.org/registry/OpenCL/specs/opencl-1.0.pdf
(item of 9.10 Half Floating-Point)
Experienced FAE
Free consultation is available.
From specific product specifications to parts selection, the Company FAE will answer your technical concerns free of charge. Please feel free to contact us.