Area efficient transcendental estimate algorithm

A circuit for generating a transcendental value, comprising:

• logic for determining whether the opcode corresponds to a trigonometric operation or a power-of-two operation;
• logic for calculating at least a fractional value from an input value provided in an instruction;
• logic for determining at least one of a shifted fractional value and a constant Value, wherein the constant value is determined based on at least a first portion of the fractional value, wherein the shifted fractional value is produced by shifting a second portion of the fractional value, wherein the second portion of the fractional value is different from the first portion of the fractional value, and wherein the first and second portions of the fractional value each include one or more, but not all, bits of the fractional value;
• and
• logic for generating the transcendental value based on at least the fractional value by summing at least the second portion of the fractional value with at least one of the shifted fractional value produced by shifting the portion of the fractional value and the constant value,
• wherein, if the opcode is determined to correspond to the power-of-two operation, the transcendental value is a power-of-two value, and the logic for generating the transcendental value is configured to:
• calculate an integer value from the input value;
• and
• determine based on the input value whether to perform complementing of the integer value and the fractional value, and
• wherein, if the opcode is determined to correspond to the trigonometric operation, the transcendental value is a trigonometric value and the logic for generating the transcendental value is configured to:
• determine, based on the fractional value, whether to perform at least one of inverting the fractional value and negating the transcendental value.

A processor, comprising:

• at least one execution unit having:
• logic for determining whether an opcode provided in the instruction corresponds to a trigonometric operation or a power-of-two operation and controlling generating of the transcendental value based on the determination;
• logic for calculating at least a fractional value from an input value provided in an instruction;
• logic for determining at least one of a shifted fractional value and a constant Value, wherein the constant value is determined based on at least a first portion of the fractional value, wherein the shifted fractional value is produced by shifting a second portion of the fractional value, wherein the second portion of the fractional value is different from the first portion of the fractional value, and wherein the first and second portions of the fractional value each include one or more, but not all, bits of the fractional value;
• and
• logic for generating a transcendental value based on at least the fractional value by summing at least the second portion of the fractional value with at least one of the shifted fractional value and the constant value,
• wherein, if the opcode is a power-of-two operation, the transcendental value is a power-of-two value, and the logic for generating the transcendental value is configured to:
• calculate an integer value from the input value;
• and
• determine based on the input value whether to perform complementing of the integer value and the fractional value, and
• wherein, if the opcode is determined to correspond and to the trigonometric operation, the transcendental value is a trigonometric value, and the logic for generating the transcendental value is configured to:
• determine, based on the fractional value, whether to perform at least one of inverting the fractional value and negating the transcendental value.