TIS Advanced

Two new modules come with the mod, the ASIC Module and the Tape Storage Module, joined by an Execution Module. Several new instructions are added to the TIS-3D set, and calculation support is built on top of TIS-3D, which must be installed first. Required dependency: TIS-3D.

Floating point work uses IEEE-754 half-precision values held in the accumulator. ADDF, SUBF, MULF and DIVF follow the integer counterparts, while JEZF, JNZF, JGZF and JLZF mirror JEZ, JNZ, JGZ and JLZ. Going the other direction, INT takes a half-precision value held in ACC and produces a 16-bit signed two's complement integer; a value sitting exactly between two whole numbers is pushed up. The MOV instruction now accepts floating point literals, so MOV 0.25 ACC loads a value directly instead of raising a compiler error.

Because the accumulator holds the same bits for both kinds of value, a program that stores an integer and then runs a floating point calculation on it can produce incorrect results. Keeping each operation on data of the matching representation is left to the programmer.