# US 2018/0321909 A1 · Variable precision floating-point multiplier

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2018/0321909 A1
- Document ID: US20180321909A1
- Status: Published application
- Kind: A1
- Application number: 16/039,029
- Filed: 2018-07-18
- Published: 2018-11-08
- Earliest claimed: 2016-08-22
- Assignees: Altera Corporation
- Applicants: Altera Corporation
- Inventors: Martin Langhammer
- Web page: https://patentgrep.com/patent/US20180321909A1

## CPC

- G06F 7/4876: {Multiplying}
- G06F 7/485: Adding; Subtracting {(G06F7/4833, G06F7/4836 take precedence)}
- G06F 2207/3816: Accepting numbers of variable word length
- G06F 2207/382: Reconfigurable for different fixed word lengths
- G06F 2207/483: Indexing scheme relating to group G06F7/483
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Integrated circuits with specialized processing blocks are provided. The specialized processing blocks may include floating-point multiplier circuits that can be configured to support variable precision. A multiplier circuit may include a first carry-propagate adder (CPA), a second carry-propagate adder (CPA), and an associated rounding circuit. The first CPA may be wide enough to handle the required precision of the mantissa. In a bridged mode, the first CPA may borrow an additional bit from the second CPA while the rounding circuit will monitor the appropriate bits to select the proper multiplier output. A parallel prefix tree operable in a non-bridged mode or the bridged mode may be used to compute multiple multiplier outputs. The multiplier circuit may also include exponent and exception handling circuitry using various masks corresponding to the desired precision width.

## Claims

### Claim 1

Multiplier circuitry on an integrated circuit die, comprising:
a partial product generator configured to receive a first input signal and a second input signal;
a compressor tree circuit configured to receive signals from the partial product generator and to generate a first set of sum and carry vectors and a second set of sum and carry vectors; and
carry propagate adder circuitry configured to receive the first set of sum and carry vectors and the second set of sum and carry vectors from the compressor tree circuit, wherein the carry propagate adder circuitry is operable to generate a first floating-point multiplier output having a first mantissa width in a first mode and to generate a second floating-point multiplier output having a second mantissa width that is different than the first mantissa width in a second mode.

### Claim 2 (depends on 1)

The multiplier circuitry of claim 1, wherein the carry propagate adder circuitry is further operable to generate a third floating-point multiplier output having a third mantissa width that is different than the first and second mantissa widths in a third mode.

### Claim 3 (depends on 2)

The multiplier circuitry of claim 2, wherein the carry propagate adder circuitry is further operable to generate a fourth floating-point multiplier output having a fourth mantissa width that is different than the first, second, and third mantissa widths in a fourth mode.

### Claim 4 (depends on 1)

The multiplier circuitry of claim 1, wherein the first mantissa width is equal to or less than 14 bits.

### Claim 5 (depends on 4)

The multiplier circuitry of claim 4, wherein the second mantissa width is different than 10 bits.

### Claim 6 (depends on 4)

The multiplier circuitry of claim 4, wherein the first mantissa width is adjusted based on an exponent of the first input signal.

### Claim 7 (depends on 1)

The multiplier circuitry of claim 1, wherein the multiplier circuitry is configured to support a first exponent width in the first mode and to support a second exponent width that is different than the first exponent width in the second mode.

### Claim 8 (depends on 7)

The multiplier circuitry of claim 7, wherein the first exponent width is 8 bits.

### Claim 9 (depends on 7)

The multiplier circuitry of claim 7, wherein the second exponent width is 5 bits.

### Claim 10 (depends on 1)

The multiplier circuitry of claim 1, wherein the multiplier circuitry is configured to support at least first and second exponent widths in the first mode.

### Claim 11 (depends on 10)

The multiplier circuitry of claim 10, wherein the first exponent width is 8 bits.

### Claim 12 (depends on 10)

The multiplier circuitry of claim 10, wherein the second exponent width is 5 bits.

### Claim 13

Multiplier circuitry on an integrated circuit die, comprising:
a first input configured to receive a first floating-point number having a first exponent;
a second input configured to receive a second floating-point number having a second exponent; and
exponent handling circuitry configured to combine the first exponent with the second exponent based on an adjustable bias value that varies as a function of the precision of the first and second exponents.

### Claim 14 (depends on 13)

The multiplier of claim 13, wherein the exponent handling circuitry is further configured to combine the first exponent with the second exponent based on an adjustable normalization value.

### Claim 15 (depends on 13)

The multiplier of claim 13, wherein first exponent has 8 bits.

### Claim 16 (depends on 13)

The multiplier of claim 13, wherein first exponent has 5 bits.

### Claim 17 (depends on 13)

The multiplier of claim 13, wherein the first floating-point number has a number of mantissa bits that is adjusted based on the first exponent.

### Claim 18

An integrated circuit die, comprising:
a first input configured to receive a first floating-point number;
a second input configured to receive a second floating-point number; and
multiplier circuitry configured to multiply the first floating-point number with the second floating-point number to generate a corresponding product, wherein the product has an exponent that is more than 5 bits, and wherein the product has a mantissa size that is adjusted based on the exponent.

### Claim 19

The integrated circuit die, wherein the multiplier circuitry has a rounding position that shifts as the mantissa is adjusted.

### Claim 20

The integrated circuit die, wherein the mantissa size of the product is different than 10 bits.
