bfp_complex_s16.h
This header contains functions implementing arithmetic operations on 16-bit complex block floating-point vectors.Functions for initializing complex BFP vectors can be found in
bfp_init.h.
- Note
- This header is included automatically through
bfp_math.h
.
- See also
bfp_complex_s16_t
bfp_complex_s32.h
This header contains functions implementing arithmetic operations on 32-bit complex block floating-point vectors.Functions for initializing complex BFP vectors can be found in
bfp_init.h.
- Note
- This header is included automatically through
bfp_math.h
.
- See also
bfp_complex_s32_t
bfp_fft.h
This header contains functions for performing block floating-point Fast Fourier Transforms (FFTs) and inverse FFTs on BFP vectors of types
bfp_s32_t
,
bfp_complex_s32_t
.
- Note
- This header is included automatically through
bfp_math.h
.
bfp_init.h
This header contains functions for initializing the structs representing block floating-point vectors. Initialization comes in two flavors: static and dynamic.Static initialization requires the user to supply the buffer which is used to store the mantissa vector. Static initiaization is accomplished using the
bfp_*_init()
functions.Dynamic initialization allocates the buffer for the mantissa vectors from the heap using
malloc()
(unless overridden by the user via configuration of
XS3_MALLOC). Dynamic allocation is accomplished using the
bfp_*_alloc()
functions. If dynamically allocated vectors are to be temporary, to avoid memory leaks they much be deallocated using the corresponding
bfp_*_dealloc()
functions.
- Note
- This header is included automatically through
bfp_math.h
.
- See also
bfp_s16_t
, bfp_s32_t
, bfp_complex_s16_t
, bfp_complex_s32_t
bfp_misc.h
bfp_s16.h
This header contains functions implementing arithmetic operations on 16-bit block floating-point vectors.Functions for initializing BFP vectors can be found in
bfp_init.h.
- Note
- This header is included automatically through
bfp_math.h
.
- See also
bfp_s16_t
bfp_s32.h
This header contains functions implementing arithmetic operations on 32-bit block floating-point vectors.Functions for initializing BFP vectors can be found in
bfp_init.h.
- Note
- This header is included automatically through
bfp_math.h
.
- See also
bfp_s32_t
bfp_math.h
This header includes the header files necessary to use lib_xs3_math
's high-level block floating- point API as well as its low-level vector API, including types and functions for 16- and 32-bit BFP vectors.
xs3_scalar_float.h
The functions contained here are part of the scalar (non-IEEE754) floating-point API.
xs3_fft.h
This header contains XS3-optimized functions for performing Fast Fourier Transforms (FFTs) and inverse FFTs on 32-bit vectors, as well as several utility functions required to perform the transforms.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_filters.h
This header contains XS3-optimized functions and types for initializing and executing 16- and 32-bit FIR filters, as well as 32-bit biquad filters.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_mixed.h
This header contains XS3-optimized functions implementing arithmetic operations where the input(s) and output vectors are not of the same bit-depth.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vect_complex_s16.h
This header contains functions implementing arithmetic operations on complex 16-bit vectors, optimized for xCore XS3.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vect_complex_s32.h
This header contains functions implementing arithmetic operations on complex 32-bit vectors, optimized for xCore XS3.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vect_f32.h
This header contains functions implementing operations on single-precision (32-bit) floats, optimized for xCore XS3.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vect_s16.h
This header contains functions implementing arithmetic operations on 16-bit vectors, optimized for xCore XS3.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vect_s32.h
This header contains functions implementing arithmetic operations on 32-bit vectors, optimized for xCore XS3.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vect_s8.h
This header contains functions implementing arithmetic operations on 8-bit vectors, optimized for xCore XS3.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_api.h
This header defines several macros used throughout the API to provide consistent, descriptive and human-readable code.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_math.h
This header includes the header files necessary to use
lib_xs3_math
's low-level vector API, including types and functions for 16- and 32-bit vectors.To access the block floating-point API, include
bfp_math.h instead (which will include this file).
xs3_math_conf.h
This header is used for detecting/setting several compile-time configuration macros.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
- See also
- Compile Time Options
xs3_math_types.h
This header contains definitions for most of the library-defined types which are used throughout this API. Several other types (such as those used for filtering) are defined locally if they're only used within a single header.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_util.h
This header contains various macros and functions.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vpu_info.h
This header contains macros and types which may be useful for interacting with the XS3 VPU at a relatively low level.
- Note
- This header is included automatically through
xs3_math.h
or bfp_math.h
.
xs3_vpu_scalar_ops.h
This header contains scalar versions of many of the XS3 VPU's vector instructions. These are mostly used for unit testing and implementing bit-exact pure C implementations of optimized functions.
- Note
- This file is not included via
xs3_math.
or bfp_math.h
- Warning
- This is not considered to be part of the public API.