search

xmos / lib_xcore_math

XMOS optimised arithmetic and vector processing library
Other
6 stars 14 forks source link

Bug in vect_packed_complex_f32_macc() #116

Closed lucianomartin closed 1 year ago

lucianomartin commented 1 year ago

The Input/Output accumulator vector must be incremented before it is assigned in vect_packed_complex_f32_macc():

  a[0].re = b[0].re * c[0].re;
  a[0].im = b[0].im * c[0].im;

must be

  a[0].re += b[0].re * c[0].re;
  a[0].im += b[0].im * c[0].im;

The same as in vect_packed_complex_f32_conj_macc().

astewart-xmos commented 1 year ago

This issue has been addressed in PR https://github.com/xmos/lib_xcore_math/pull/120 and a fix will be available in v2.1.1