TIS Advanced

TIS Advanced features include two new modules, and several new instructions for the Execution Module.

New Modules

New Instructions for Execution Module

Floating Point Arithmetic Instructions

ADDF, SUBF, MULF, and DIVF behave like their integer counterparts, however, they operate on ACC as though the data contained within represents an IEEE-754 compliant half-precision floating point number.

Note that it is possible to store an integer value in ACC and attempt floating point calculations on it, and vice versa. This is likely to produce incorrect values. It is the user's responsibility to ensure that operations are only performed on the correct data representation.

Floating Point Flow Control Instructions

The JEZF, JNZF, JGZF, and JLZF instructions behave like their integer counterparts JEZ, JNZ, JGZ, and JLZ, however, like the floating point arithmetic operations above, they operate on ACC as though the value contained within is an IEEE-754 half-precision float.

Floating Point Conversion Instructions

The FLT instruction will convert an integer value stored in ACC to a IEEE-754 half-precision floating point representation. The INT instruction will convert a floating point value stored in ACC to a 16-bit signed two's complement integer representation; the same representation used by standard TIS-3D arithmetic operations. Converting to integer representation will round the floating point value to the nearest whole number, with ties rounding upwards.

`MOV` Instruction Changes

The MOV instruction has been extended to support handling floating point literals. For example, the instruction MOV 0.25 ACC will no longer be a compiler error, and will now load the floating point value 0.25 into the ACC register as expected.