Method and apparatus for generating an initial estimate for a floating point reciprocal of a square root
AssigneeSamsung Electronics Co., Ltd.
InventorsRoney S. Wong, Hei T. Fung
Life of this patent
Estimating the term…
Abstract
An initial estimate of a reciprocal of a square root of a floating point number is generated by subtracting the input floating point number from a constant and shifting the results to the right by one bit. Additionally, the initial estimate of a reciprocal of a square root of a floating point number can be determined by decrementing the exponent by one, shifting the exponent and fraction to the right by one bit, and subtracting the result from predetermined constant. The estimate for the reciprocal square root can also be determined by shifting the floating point number to the right by one bit and subtracting the shift result from a predetermined number to generate the initial estimate.
Claims38
A method of determining an initial estimate for a reciprocal of a square root of a floating point number having a fraction, f, and an exponent, E, comprising the steps of:
- storing a predetermined number;
- subtracting, using a subtraction circuit, the floating point number from the predetermined number to provide a first result;
- and
- shifting, in a shift circuit, the first result to the right by one bit position to obtain the initial estimate for the reciprocal.
A method of determining an initial estimate for a reciprocal of a square root of a floating point number having a fraction, f, and an exponent, E, comprising the steps of:
- storing a predetermined number, m';
- shifting, in a shift circuit, the floating point number to the right by one bit position to obtain a first result;
- and
- subtracting, using a subtraction circuit, the first result from the predetermined number m' to provide the initial estimate.
An apparatus for determining an initial estimate for a reciprocal of a square root of floating point number having a fraction, f, and an exponent, E, comprising:
- storage locations for the floating point number and a predetermined number;
- a subtraction circuit, coupled to the floating point number and the predetermined number, the subtraction circuit providing as a first result, the floating point number subtracted from the predetermined number;
- and
- a shift circuit, receiving the first result, and providing the first result shifted to the right by one bit position to generate the initial estimate.
An apparatus for determining an initial estimate for a reciprocal of a square root of floating point number having a fraction, f, and an exponent, E, comprising:
- first and second storage locations for the floating point number and a predetermined number;
- a shift circuit coupled to the first storage location, receiving the floating point number and outputting the floating point number shifted by one bit as a first result;
- and
- a subtraction circuit, coupled to the shift circuit and the second storage location, the subtraction circuit subtracting, the first result from the predetermined number to provide the initial estimate.
A method of determining an initial estimate for a reciprocal of a square root of a floating point number having a fraction, f and an exponent, E, comprising the steps of:
- storing a first and second predetermined number;
- selecting one of the first and second predetermined numbers according to a predetermined bit of the exponent;
- subtracting, using a subtraction circuit, the floating point number from the selected one of the predetermined numbers to provide a first result;
- and
- shifting, in a shift circuit, the first result to the right by one bit position to obtain the initial estimate for the reciprocal.
An apparatus for determining an initial estimate for a reciprocal of a square root of floating point number having a fraction and an exponent, comprising:
- first and second storage locations for a first and second predetermined number;
- a shift circuit receiving the floating point number and outputting the floating point number right shifted by one bit as a first result;
- a multiplexer circuit, coupled to the first and second storage locations and outputting one of the first and second predetermined numbers according to the value of the least significant bit of the exponent;
- and
- a subtraction circuit, coupled to the shift circuit and the multiplexer circuit, the subtraction circuit subtracting, the first result from the selected predetermined number to provide the initial estimate.
A method of determining an initial estimate for a reciprocal of a square root of a floating point number having a fraction, f, and an exponent, E, comprising the steps of:
- providing a subtraction circuit;
- subtracting, using the subtraction circuit, one from the least significant bit of the exponent to generate a first result;
- shifting, in a shift circuit, the first result and the fraction to the right by one bit position to obtain a second result;
- and
- subtracting the second result from a predetermined number to obtain the initial estimate for the reciprocal.
An apparatus for determining an initial estimate for a reciprocal of a square root of floating point number having a fraction, f, and an exponent, E, comprising:
- a first and second multiplexer circuit;
- storage locations for the floating point number and for a first and second predetermined number, the floating point number and the second predetermined number being connected to the first multiplexer circuit;
- a subtraction circuit, coupled to the first and second multiplexer circuits, receiving the floating point number and the first predetermined number through the first and second multiplexer circuits, and providing as a first result, the first predetermined number subtracted from the floating point number;
- a shift circuit, coupled to the subtraction circuit and to the second multiplexer circuit, and providing the first result shifted to the right by one bit position to the second multiplexer circuit as a second result;
- and
- the first and second multiplexer circuits providing the subtraction circuit with the second predetermined number and the second result, the subtraction circuit subtracting the second result from the second predetermined number to generate the initial estimate.
An apparatus for determining an initial estimate for a reciprocal of a square root of floating point number having a fraction, f, and an exponent, E, comprising:
- means for storing the floating point number and a predetermined number;
- means, coupled to the floating point number and the predetermined number, for subtracting the floating point number from the predetermined number to provide a first result;
- and
- means, receiving the first result, for shifting the first result by one bit position to the right to generate the initial estimate.
An apparatus for determining an initial estimate for a reciprocal of a square root of floating point number having a fraction, f, and an exponent, E, comprising:
- means for storing the floating point number and a predetermined number;
- means, coupled to the floating point number, for shifting the floating point number by one bit position to the right to generate a first result;
- and
- means, coupled to the shifter means and the predetermined number, for subtracting the first result from the predetermined number to provide the initial estimate.
Dependencies are inferred from the claim text.