Numba support? #75
Replies: 4 comments
-
Interesting question! |
Beta Was this translation helpful? Give feedback.
-
The main reason I want to use numba with fxp is that it seems like a convenient way to simulate adaptive filters (that work with the gradiant descend methods) that will need to be run on DSPs. While most DSPs nowadays support floating point math, it is sometimes prefered to just use fixed point whenever possible. The idea would be to simlulate and contrast the performance of the algorithms implemented with fxp vs floats. |
Beta Was this translation helpful? Give feedback.
-
The main reason I want to use numba with fxp is that it seems like a convenient way to simulate adaptive filters (that work with the gradiant descend methods) that will need to be run on DSPs. While most DSPs nowadays support floating point math, it is sometimes prefered to just use fixed point whenever possible. The idea would be to simlulate and contrast the performance of the algorithms implemented with fxp vs floats. |
Beta Was this translation helpful? Give feedback.
-
The main reason I want to use numba with fxp is that it seems like a convenient way to simulate adaptive filters (that work with the gradiant descend methods) that will need to be run on DSPs. While most DSPs nowadays support floating point math, it is sometimes prefered to just use fixed point whenever possible. The idea would be to simlulate and contrast the performance of the algorithms implemented with fxp vs floats. |
Beta Was this translation helpful? Give feedback.
-
Is it possible to integrate Fxp into Numba? Maybe using something like the
jitclass
decoratorBeta Was this translation helpful? Give feedback.
All reactions