From 4acfef7b3cad53febebda74fd2f78f3c6674e56c Mon Sep 17 00:00:00 2001 From: Christophe Favergeon Date: Wed, 10 Jan 2024 09:05:18 +0100 Subject: [PATCH] Updated documentation for issue #145 --- Source/SupportFunctions/arm_float_to_q31.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/Source/SupportFunctions/arm_float_to_q31.c b/Source/SupportFunctions/arm_float_to_q31.c index 753de3bc..73b6fb7e 100644 --- a/Source/SupportFunctions/arm_float_to_q31.c +++ b/Source/SupportFunctions/arm_float_to_q31.c @@ -60,6 +60,12 @@ @note In order to apply rounding, the library should be rebuilt with the ROUNDING macro defined in the preprocessor section of project options. + + @note If the input float values are very big (2**32) then the function + won't be able to saturate to the right values. + If you expect very big float values in the input array then you + should force those values to +1 or -1 before calling this function. + For reasonable float values (< 2**32), the function will saturate correctly. */ #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)