Abstract: An arrangement and method for decoding coded instructions and playing and replaying decoded instructions to a machine. The arrangement has a source of coded instructions. Connected to this coded instruction source is a decoder for receiving and decoding the coded instructions and for outputting the decoded instructions to a machine. A silo is connected to the output of the decoder and siloes and outputs the decoded instructions to the machine. The outputting of the decoded instructions to the machine are switched between the silo and the decoder, so that the machine receives the siloed decoded instructions. By siloing and then replaying already decoded instructions at the time of a trap occurrence, a speed increase is achieved, since the instructions which are in the trap shadow do not have to be decoded again.