Draft: [DO NOT MERGE] issue 2653: Possible bug regarding exponent handling in function mix_output_block_fx()

Reason why this change is needed

  • This change aims to first identify test cases where this function is used. Second, if the testcase actually leads to mismatch between FX and FP versions due to a wrong exponent, further changes will correct it.

Description of the change

  • For now, an assert is put where the function is called.

Affected operating points

  • I would assume it should be Bit exact.

Merge request reports

Loading