# US 2001/0017626 A1 · Graphics processing unit with transform module capable of handling scalars and vectors

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

- Number: US 2001/0017626 A1
- Document ID: US20010017626A1
- Status: Published application
- Kind: A1
- Application number: 09/775,129
- Filed: 2001-01-31
- Published: 2001-08-30
- Earliest claimed: 1999-12-06
- Assignees: NVIDIA Corporation
- Inventors: John Erik Lindholm; Simon Moy; David B. Kirk; Paolo E. Sabella
- Web page: https://patentgrep.com/patent/US20010017626A1

## CPC

- G06T 11/40: Filling planar surfaces by adding surface attributes, e.g. adding colours or textures
- G06T 15/005: {General purpose rendering architectures}
- G06T 15/04: Texture mapping
- G06T 15/50: Lighting effects
- G06T 15/503: {Blending, e.g. for anti-aliasing}
- G06T 15/506: {Illumination models}
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A method, apparatus and article of manufacture are provided for a transform system for graphics processing as a computer system or on a single integrated circuit. Included is an input buffer adapted for being coupled to a vertex attribute buffer for receiving vertex data therefrom. A multiplication logic unit has a first input coupled to an output of the input buffer. Also provided is an arithmetic logic unit having a first input coupled to an output of the multiplication logic unit. Coupled to an output of the arithmetic logic unit is an input of a register unit. An inverse logic unit is provided including an input coupled to the output of the arithmetic logic unit or the register unit for performing an inverse or an inverse square root operation. Further included is a conversion module coupled between an output of the inverse logic unit and a second input of the multiplication logic unit. In use, the conversion module serves to convert scalar vertex data to vector vertex data. Memory is coupled to the multiplication logic unit and the arithmetic logic unit. The memory has stored therein a plurality of constants and variables for being used in conjunction with the input buffer, the multiplication logic unit, the arithmetic logic unit, the register unit, the inverse logic unit, and the conversion module for processing the vertex data. Finally, an output converter is coupled to the output of the arithmetic logic unit for being coupled to a lighting module to output the processed vertex data thereto.

## Claims

### Claim 1

A transform system for graphics processing, comprising:
(a) an input buffer adapted for being coupled to a vertex attribute buffer for receiving vertex data therefrom;
(b) a multiplication logic unit having a first input coupled to an output of the input buffer;
(c) an arithmetic logic unit having a first input coupled to an output of the multiplication logic unit;
(d) a register unit having an input coupled to an output of the arithmetic logic unit;
(e) an inverse logic unit including an input coupled to the output of the arithmetic logic unit or the register unit for performing an inverse or an inverse square root operation;
(f) a conversion module coupled between an output of the inverse logic unit and a second input of the multiplication logic unit, the conversion module adapted to convert scalar vertex data to vector vertex data;
(g) memory coupled to the multiplication logic unit and the arithmetic logic unit, the memory having stored therein a plurality of constants and variables for being used when processing the vertex data; and
(h) an output converter coupled to the output of the arithmetic logic unit and adapted for being coupled to a lighting module to output the processed vertex data thereto.

### Claim 2 (depends on 1)

The system as recited in claim 1, wherein the memory is coupled to the second input of the multiplication logic unit.

### Claim 3 (depends on 1)

The system as recited in claim 1, wherein the input of the inverse logic unit is coupled to the output of the arithmetic logic unit.

### Claim 4 (depends on 1)

The system as recited in claim 1, wherein the inputs of the multiplication logic unit include multiplexers.

### Claim 5 (depends on 1)

The system as recited in claim 1, wherein at least one of the inputs of the arithmetic logic unit includes a multiplexer.

### Claim 6 (depends on 1)

The system as recited in claim 1, wherein the memory has a write terminal coupled to the output of the arithmetic logic unit.

### Claim 7 (depends on 1)

The system as recited in claim 1, wherein the output of the multiplication logic unit has a feedback loop coupled to the first input thereof.

### Claim 8 (depends on 1)

The system as recited in claim 1, wherein the output of the register unit is coupled to the first input of the multiplication logic unit.

### Claim 9 (depends on 8)

The system as recited in claim 8, wherein the output of the register unit is coupled to the second input of the multiplication logic unit.

### Claim 10 (depends on 1)

The system as recited in claim 1, wherein the output of the arithmetic logic unit has a feedback loop connected to the second input thereof.

### Claim 11 (depends on 10)

The system as recited in claim 10, wherein the feedback loop has a delay coupled thereto.

### Claim 12 (depends on 1)

The system as recited in claim 1, wherein the multiplication logic unit is capable of performing a rotate operation on vector vertex data.

### Claim 13 (depends on 1)

The system as recited in claim 1, wherein the inverse logic unit is capable of clamping a value of an inverse operation if the value of the inverse operation meets predetermined criteria.

### Claim 14 (depends on 1)

The system as recited in claim 1, wherein included are six input buffers coupled to the first input of the multiplication logic unit.

### Claim 15 (depends on 1)

The system as recited in claim 1, wherein the multiplication logic unit includes four multipliers coupled in parallel.

### Claim 16 (depends on 1)

The system as recited in claim 1, wherein the arithmetic logic unit includes three adders coupled in parallel and series.

### Claim 17 (depends on 1)

The system as recited in claim 1, wherein the register unit includes four sets of registers each having an output coupled to a first input of an associated multiplexer which has a second input coupled to the input of the corresponding set of registers.

### Claim 18 (depends on 1)

The system as recited in claim 1, wherein the register unit is threaded.

### Claim 19 (depends on 1)

The system as recited in claim 1, wherein the output converter is adapted to be coupled to the lighting module via output buffers.

### Claim 20 (depends on 1)

The system as recited in claim 1, wherein a register is coupled between the output of the inverse logic unit and an input of the conversion unit.

### Claim 21 (depends on 20)

The system as recited in claim 20, wherein the register is threaded.

### Claim 22 (depends on 1)

The system as recited in claim 1, wherein the register unit is capable of being masked at a vector component level.

### Claim 23

A system for handling scalar and vector components during graphics processing, comprising:
(a) a vector operation module for receiving vertex data in the form of vectors and performing vector operations on the vector vertex data;
(b) a conversion module coupled to the vector operation module for converting scalar vertex data from the vector operation module into vector vertex data; and
(c) a register coupled to the vector operation module for storing an output of the vector operation module for feeding the output back to the vector operation module.

### Claim 24 (depends on 23)

The system as recited in claim 23, wherein the vector operation module includes at least one of multiplier and an adder.

### Claim 25 (depends on 23)

The system as recited in claim 23, wherein zero latency is achieved by bypassing the register.

### Claim 26 (depends on 25)

The system as recited in claim 25, wherein the register includes a vector component write mask for generating vector vertex data.

### Claim 27 (depends on 23)

The system as recited in claim 23, and further comprising a scalar operation module adapted for executing scalar operations on an output of the vector operation module, thereby rendering vertex data in the form of scalars.

### Claim 28 (depends on 27)

The system as recited in claim 27, wherein the scalar operations include inverse or inverse square root operations.

### Claim 29

A method for handling scalar and vector components during graphics processing, comprising:
(a) receiving vertex data in the form of vectors;
(b) performing vector operations on the vector vertex data;
(c) converting scalar vertex data resulting from the vector operations into vector vertex data;
(d) storing an output of the vector operations; and
(e) performing additional vector operations on the stored output of the vector operations.

### Claim 30 (depends on 29)

The method as recited in claim 29, wherein the vector operations include multiplication or addition operations.

### Claim 31 (depends on 29)

The method as recited in claim 29, wherein the vector operations are performed on the output of the vector operations with zero latency.

### Claim 32 (depends on 31)

The method as recited in claim 31, wherein the output of the vector operations is stored in a register unit, and the zero latency is achieved by bypassing the register unit.

### Claim 34 (depends on 29)

The method as recited in claim 29, and further comprising executing scalar operations on an output of the vector operations, thereby rendering vertex data in the form of scalars.

### Claim 35 (depends on 34)

The method as recited in claim 34, wherein the scalar operations include inverse or inverse square root operations.

### Claim 36 (depends on 34)

The method as recited in claim 34, and further comprising extracting scalar vertex data from the output of the vector operations if the output is in the form of vectors.

### Claim 37 (depends on 36)

The method as recited in claim 36, wherein the extraction is carried out by a multiplexer.

### Claim 38 (depends on 29)

The method as recited in claim 29, wherein the received vertex data is manipulated by a multiplexer.

### Claim 39

A computer program embodied on a computer readable medium for handling scalar and vector components during graphics processing, comprising:
(a) a code segment for receiving vertex data in the form of vectors;
(b) a code segment for performing vector operations on the vector vertex data;
(c) a code segment for converting scalar vertex data resulting from the vector operations into vector vertex data;
(d) storing an output of the vector operations; and
(e) performing additional vector operations on the stored output of the vector operations.

### Claim 40 (depends on 39)

The computer program as recited in claim 39, wherein the vector operations include multiplication or addition operations.

### Claim 41 (depends on 39)

The computer program as recited in claim 39, wherein the vector operations are performed on the output of the vector operations with zero latency.

### Claim 42 (depends on 41)

The computer program as recited in claim 41, wherein the output of the vector operations is stored in a register unit, and the zero latency is achieved by bypassing the register unit.

### Claim 43 (depends on 39)

The computer program as recited in claim 39, and further comprising a code segment for executing scalar operations on an output of the vector operations, thereby rendering vertex data in the form of scalars.

### Claim 44 (depends on 43)

The computer program as recited in claim 43, wherein the scalar operations include inverse or inverse square root operations.

### Claim 45 (depends on 43)

The computer program as recited in claim 43, and further comprising a code segment for extracting scalar vertex data from the output of the vector operations if the output is in the form of vectors.

### Claim 46 (depends on 45)

The method as recited in claim 45, wherein the extraction is carried out by a multiplexer.

### Claim 47 (depends on 39)

The method as recited in claim 39, wherein the received vertex data is manipulated by a multiplexer.

### Claim 48

A method for performing a blending operation during graphics processing in a hardware-implemented graphics pipeline, comprising:
(a) receiving a plurality of matrices, a plurality of weight values each corresponding with one of the matrices, and vertex data in a buffer;
(b) calculating a sum of a plurality of products with each product calculated by the multiplication of the vertex data, one of the matrices, and the weight corresponding to the matrix, wherein the calculation is executed on a single integrated circuit; and
(c) outputting the sum of products for additional processing.

### Claim 49 (depends on 48)

The method as recited in claim 48, wherein the matrices include model view matrices.

### Claim 50 (depends on 49)

The method as recited in claim 49, wherein the additional processing includes multiplying the sum of products by a composite matrix for displaying purposes.

### Claim 51 (depends on 49)

The method as recited in claim 49, wherein the additional processing includes a lighting operation.

### Claim 52 (depends on 48)

The method as recited in claim 48, wherein the matrices include inverse matrices and the vertex data includes a normal vector.

### Claim 53 (depends on 48)

The method as recited in claim 48, wherein the single integrated circuit includes: a multiplication logic unit having a first input coupled to an output of the buffer for receiving the vertex data; an arithmetic logic unit having a first input coupled to an output of the multiplication logic unit; a register unit having an input coupled to an output of the arithmetic logic unit; memory coupled to the multiplication logic unit and the arithmetic logic unit, the memory having stored therein a plurality of constants and variables for being when processing the vertex data.

### Claim 54 (depends on 48)

The system as recited in claim 48, wherein the single integrated circuit includes: a multiplication logic unit having a first input coupled to an output of the buffer; an arithmetic logic unit having a first input coupled to an output of the multiplication logic unit; a register unit having an input coupled to an output of the arithmetic logic unit; an inverse logic unit including an input coupled to the output of the arithmetic logic unit or the register unit for performing an inverse or an inverse square root operation; a conversion module coupled between an output of the inverse logic unit and a second input of the multiplication logic unit, the conversion module adapted to convert scalar vertex data to vector vertex data; memory coupled the multiplication logic unit and the arithmetic logic unit, the memory having stored therein a plurality of constants and variables for being used when processing the vertex data; and an output converter coupled to the output of the arithmetic logic unit for being coupled to a lighting module to output the processed vertex data thereto.

### Claim 55

A system for performing a blending operation during graphics processing in a graphics pipeline, comprising:
(a) a buffer for receiving a plurality of matrices, a plurality of weight values each corresponding with one of the matrices, and vertex data;
(b) a single integrated circuit coupled to the buffer for calculating a sum of a plurality of products with each product calculated by the multiplication of the vertex data, one of the matrices, and the weight corresponding to the matrix; and
(c) wherein the sum of products is outputted from the single integrated circuit for additional processing.

### Claim 56 (depends on 55)

The system as recited in claim 55, wherein the matrices include model view matrices.

### Claim 57 (depends on 56)

The system as recited in claim 56, wherein the additional processing includes multiplying the sum of products by a composite matrix for displaying purposes.

### Claim 58 (depends on 56)

The system as recited in claim 56, wherein the additional processing includes a lighting operation.

### Claim 59 (depends on 55)

The system as recited in claim 55, wherein the matrices include inverse matrices and the vertex data includes a normal vector.

### Claim 60 (depends on 55)

The system as recited in claim 55, wherein the single integrated circuit includes: a multiplication logic unit having a first input coupled to an output of the buffer for receiving the vertex data; an arithmetic logic unit having a first input coupled to an output of the multiplication logic unit; a register unit having an input coupled to an output of the arithmetic logic unit; memory coupled to the multiplication logic unit and the arithmetic logic unit, the memory having stored therein a plurality of constants and variables for being used when processing the vertex data.

### Claim 61 (depends on 55)

The system as recited in claim 55, wherein the single integrated circuit includes: a multiplication logic unit having a first input coupled to an output of the buffer; an arithmetic logic unit having a first input coupled to an output of the multiplication logic unit; a register unit having an input coupled to an output of the arithmetic logic unit; an inverse logic unit including an input coupled to the output of the arithmetic logic unit or the register unit for performing an inverse or an inverse square root operation; a conversion module coupled between an output of the inverse logic unit and a second input of the multiplication logic unit, the conversion module adapted to convert scalar vertex data to vector vertex data; memory coupled to the multiplication logic unit and the arithmetic logic unit, the memory having stored therein a plurality of constants and variables for being used when processing the vertex data; and an output converter coupled to the output of the arithmetic logic unit for being coupled to a lighting module to output the processed vertex data thereto.

### Claim 62

A method for handling output values in a graphics processing module representative of an inverse operation involving a W-attribute of vertex data, comprising:
(a) processing vertex data, wherein the processing of the vertex data includes an inverse operation involving a W-attribute of the vertex data;
(b) outputting the processed vertex data;
(c) identifying a value of the inverse operation involving the W-attribute of the vertex data; and
(d) clamping the value of the inverse operation if the value of the inverse operation meets predetermined criteria.

### Claim 63 (depends on 62)

The method as recited in claim 62, wherein the criteria includes the value of the inverse operation being greater than a predetermined amount.

### Claim 64 (depends on 62)

The method as recited in claim 62, wherein the value is clamped by an inverse logic unit in a transform module.

### Claim 65 (depends on 62)

The method as recited in claim 62, wherein the value is clamped to a minimum and a maximum exponent.

### Claim 66

A computer program embodied on a computer readable medium for handling output values in a graphics processing module representative of an inverse operation involving a W-attribute of vertex data, comprising:
(a) a code segment for processing vertex data, wherein the processing of the vertex data includes an inverse operation involving a W-attribute of the vertex data;
(b) a code segment for outputting the processed vertex data;
(c) a code segment for identifying a value of the inverse operation involving the W-attribute of the vertex data; and
(d) a code segment for clamping the value of the inverse operation if the value of the inverse operation meets predetermined criteria.

### Claim 67 (depends on 66)

The computer program as recited in claim 66, wherein the criteria includes the value of the inverse operation being greater than a predetermined amount.

### Claim 68 (depends on 66)

The computer program as recited in claim 66, wherein the value is clamped by an inverse logic unit in a transform module.

### Claim 69 (depends on 66)

The computer program as recited in claim 66, wherein the value is clamped to a minimum and a maximum exponent.
