mindscience.sciops.fft.set_fft_cache_size

mindscience.sciops.fft.set_fft_cache_size(cache_size)

Set cache number of ASD FFT operators to optimize the function call performance. Without cache, every ASD FFT function call will dlopen the function symbols from .so dynamically, which will introduce some overhead. With cache, the function symbols will be loaded into cache when the first ASD FFT function call is made, and will not be loaded again in subsequent calls.

Parameters

cache_size (int) – Cache number of ASD FFT operators.

Examples

>>> from mindscience.sciops.fft import set_fft_cache_size
>>> set_fft_cache_size(1024)  # Set 1024 fft operators cache number