`
All prior lib_xs3_math v1 header includes were changed.
Renamed functions
All occurrences of double_to_float_s32 were replaced with f64_to_float_s32
All occurrences of float_to_float_s32 were replaced with f32_to_float_s32
The xs3_ prefix was removed from all function calls to the older, lib_xs3_math API
Typedef changes
The type fixed_s32_t has been removed from the library.
fixed_s32_t types were replaced with int32_t where the q-format was uncertain. fixed_s32_t types were replaced with the new typedefs uq0_32, uq1_31, uq2_30, uq8_24 (or appropriate) where the q-format could be determined from the variable name.
Notes to Reviewers
Ensure the python that generates source code is correct
Ensure that the q-format types are correct
The remainder of the changes should be straight forward. Mostly needs another set of eyes to make sure no silly mistakes were made. 120 files have been modified.
Summary of Changes
Includes
The recommended API include is now:
`
include "xmath/xmath.h"
` All prior lib_xs3_math v1 header includes were changed.
Renamed functions
All occurrences of
double_to_float_s32
were replaced withf64_to_float_s32
All occurrences offloat_to_float_s32
were replaced withf32_to_float_s32
Thexs3_
prefix was removed from all function calls to the older, lib_xs3_math APITypedef changes
The type
fixed_s32_t
has been removed from the library.fixed_s32_t
types were replaced withint32_t
where the q-format was uncertain.fixed_s32_t
types were replaced with the new typedefsuq0_32
,uq1_31
,uq2_30
,uq8_24
(or appropriate) where the q-format could be determined from the variable name.Notes to Reviewers
Ensure the python that generates source code is correct
Ensure that the q-format types are correct
The remainder of the changes should be straight forward. Mostly needs another set of eyes to make sure no silly mistakes were made. 120 files have been modified.