Patent US3087444A successfully refreshed from Google Patents.
Patent US5821949A successfully refreshed from Google Patents.
US5821949A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Three-dimensional graphics accelerator with direct data channels for improved performance
Assignee: Oracle America Inc
•
Filing: 1996-07-01
•
Granted: 1998-10-13
•
Expires: 2016-07-01
Abstract
Abstract A 3-D graphics accelerator which includes a command block or preprocessor, a plurality of floating point processors or blocks, and one or more draw processors or blocks. The 3-D graphics accelerator includes a plurality of direct data channels or point-to-point buses which connect the command preprocessor to each of the plurality of floating point processors. The 3-D graphics accelerator also includes a plurality of direct data channels or point-to-point buses which connect the plurality of floating point processors to each of the draw processors. These direct data channels or point-to-point buses provide data transfer throughput similar to prior art designs with improved electrical performance. The plurality of direct data channels or point-to-point buses enables smaller data paths, e.g., 8 bit data paths, while providing similar bandwidth to prior art shared bus designs. The use of these smaller direct data paths also provides better electrical characteristics for the graphical architecture. First, the direct data channel output pins on the command chip are only required to drive a single device, as opposed to driving multiple devices in a shared bus architecture. Also, each of the floating point processors have a reduced number of pins, since each only connects to an 8 bit bus. Further, the direct data paths provide improved connectivity between multiple boards. The improved electrical characteristics also enable the user of higher clock speeds, thus providing increased transfer bandwidth.
Inventors
- 👤 Michael F. Deering
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 17:32
CPC Classifications
G; G06; G06T; G06T1/00; G06T1/20; G06F; G06F9/00; G06F9/06; G06F9/30; G06F9/38; G06F9/3877; G06F9/3879
Claims Analysis
Total Claims: 16
Claim 1
Independent Claim
351 words
A 3-D graphics accelerator for performing three-dimensional graphics acceleration functions, comprising: a frame buffer memory; a command block for receiving geometry data and outputting initial geometric primitive data; a plurality of floating point blocks for receiving said initial geometric primitive data from said command block, and wherein each of said plurality of floating point blocks is configured to perform geometry processing operations on said initial geometric primitive data in order to produce modified geometric primitive data; a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks, wherein said command block couples to each of said plurality of floating point blocks through said first plurality of direct data channels, wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and one of said plurality of floating point blocks, wherein said command block includes round robin arbitration logic coupled to said first plurality of direct data channels, wherein said round robin arbitration logic operates to provide data to each of said plurality of floating point blocks in a substantially round robin fashion; one or more draw blocks coupled to the frame buffer memory, wherein said one or more draw blocks are configured to receive said modified geometric primitive data from said plurality of floating point blocks, and wherein said one or more draw blocks are configured to render pixel data corresponding to said modified geometric primitive data into the frame buffer memory; a second plurality of direct data channels coupled between each of said plurality of floating point blocks and said one or more draw blocks, wherein each of said plurality of floating point blocks is coupled to each of said one or more draw blocks by one of said second plurality of direct data channels, wherein each of said plurality of floating point blocks provide said modified geometric primitive data to said one or more draw blocks; a digital-to-analog converter coupled to said frame buffer memory for receiving said pixel data from said frame buffer memory and providing an analog output to a video monitor.
Claim 2
Dependent Claim
351 words
A 3-D graphics accelerator for performing three-dimensional graphics acceleration functions, comprising: a frame buffer memory; a command block for receiving geometry data and outputting initial geometric primitive data; a plurality of floating point blocks for receiving said initial geometric primitive data from said command block, and wherein each of said plurality of floating point blocks is configured to perform geometry processing operations on said initial geometric primitive data in order to produce modified geometric primitive data; a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks, wherein said command block couples to each of said plurality of floating point blocks through said first plurality of direct data channels, wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and one of said plurality of floating point blocks, wherein said command block includes round robin arbitration logic coupled to said first plurality of direct data channels, wherein said round robin arbitration logic operates to provide data to each of said plurality of floating point blocks in a substantially round robin fashion; one or more draw blocks coupled to the frame buffer memory, wherein said one or more draw blocks are configured to receive said modified geometric primitive data from said plurality of floating point blocks, and wherein said one or more draw blocks are configured to render pixel data corresponding to said modified geometric primitive data into the frame buffer memory; a second plurality of direct data channels coupled between each of said plurality of floating point blocks and said one or more draw blocks, wherein each of said plurality of floating point blocks is coupled to each of said one or more draw blocks by one of said second plurality of direct data channels, wherein each of said plurality of floating point blocks provide said modified geometric primitive data to said one or more draw blocks; a digital-to-analog converter coupled to said frame buffer memory for receiving said pixel data from said frame buffer memory and providing an analog output to a video monitor.
Claim 3
Dependent Claim
73 words
The 3-D graphics accelerator of claim 1, wherein said round robin arbitration logic determines which of the plurality of floating point processors is to receive a next packet of said initial geometric primitive data; wherein said round robin arbitration logic provides said next packet of said initial geometric primitive data to a determined one of said plurality of floating point blocks via a corresponding one of said first plurality of direct data channels.
Claim 4
Dependent Claim
53 words
The 3-D graphics accelerator of claim 1, wherein said round robin arbitration logic operates to distribute said initial geometric primitive data substantially evenly between said plurality of floating point blocks; wherein said round robin arbitration logic maintains a substantially even flow of data across each of said first plurality of direct data channels.
Claim 5
Dependent Claim
54 words
The 3-D graphics accelerator of claim 3, wherein said round robin arbitration logic utilizes a next-available round arbitration scheme, wherein said round robin arbitration logic skips a particular one of said first plurality of direct data channels if said particular one of said first plurality of direct data channels is unable to transfer data.
Claim 6
Dependent Claim
367 words
A 3-D graphics accelerator for performing three-dimensional graphics acceleration functions, comprising: a frame buffer memory; a command block for receiving geometry data and outputting initial geometric primitive data; a plurality of floating point blocks for receiving said initial geometric primitive data from said command block, wherein each of said plurality of floating point blocks is configured to perform geometry processing operations on said initial geometric primitive data in order to produce modified geometric primitive data; a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks, wherein said command block couples to each of said plurality of floating point blocks through said first plurality of direct data channels, wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and one of said plurality of floating point blocks; one or more draw blocks coupled to the frame buffer memory, wherein said one or more draw blocks are configured to receive said modified geometric primitive data from said plurality of floating point blocks, and wherein said one or more draw blocks are configured to render pixel data corresponding to said modified geometric primitive data into the frame buffer memory; a second plurality of direct data channels coupled between each of said plurality of floating point blocks and said one or more draw blocks, wherein each of said plurality of floating point blocks is coupled to each of said one or more draw blocks by one of said second plurality of direct data channels, wherein each of said plurality of floating point blocks provide said modified geometric primitive data to said one or more draw blocks; a digital-to-analog converter coupled to said frame buffer memory for receiving said pixel data from said frame buffer memory and providing an analog output to a video monitor; wherein each of said one or more draw blocks includes a plurality of input buffers, and wherein each of said plurality of input buffers for a given draw block is coupled to a respective one of said second plurality of direct data channels which couples said given draw block to a corresponding one of said plurality of floating point blocks.
Claim 7
Dependent Claim
367 words
A 3-D graphics accelerator for performing three-dimensional graphics acceleration functions, comprising: a frame buffer memory; a command block for receiving geometry data and outputting initial geometric primitive data; a plurality of floating point blocks for receiving said initial geometric primitive data from said command block, wherein each of said plurality of floating point blocks is configured to perform geometry processing operations on said initial geometric primitive data in order to produce modified geometric primitive data; a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks, wherein said command block couples to each of said plurality of floating point blocks through said first plurality of direct data channels, wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and one of said plurality of floating point blocks; one or more draw blocks coupled to the frame buffer memory, wherein said one or more draw blocks are configured to receive said modified geometric primitive data from said plurality of floating point blocks, and wherein said one or more draw blocks are configured to render pixel data corresponding to said modified geometric primitive data into the frame buffer memory; a second plurality of direct data channels coupled between each of said plurality of floating point blocks and said one or more draw blocks, wherein each of said plurality of floating point blocks is coupled to each of said one or more draw blocks by one of said second plurality of direct data channels, wherein each of said plurality of floating point blocks provide said modified geometric primitive data to said one or more draw blocks; a digital-to-analog converter coupled to said frame buffer memory for receiving said pixel data from said frame buffer memory and providing an analog output to a video monitor; wherein each of said one or more draw blocks includes a plurality of input buffers, and wherein each of said plurality of input buffers for a given draw block is coupled to a respective one of said second plurality of direct data channels which couples said given draw block to a corresponding one of said plurality of floating point blocks.
Claim 8
Dependent Claim
187 words
A 3-D graphics accelerator for performing three-dimensional graphics acceleration functions, comprising: a frame buffer memory; a command block for receiving geometry data and outputting initial geometric primitive data; a plurality of floating point blocks for receiving said initial geometric primitive data from said command block, wherein each of said plurality of floating point blocks is configured to perform geometry processing operations on said initial geometric primitive data in order to produce modified geometric primitive data, and a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks, wherein said command block couples to each of said plurality of floating point blocks through said first plurality of direct data channels, wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and one of said plurality of floating point blocks, wherein said command block includes round robin arbitration logic coupled to said first plurality of direct data channels, wherein said round robin arbitration logic operates to provide data to each of said plurality of floating point blocks in a substantially round robin fashion.
Claim 9
Dependent Claim
187 words
A 3-D graphics accelerator for performing three-dimensional graphics acceleration functions, comprising: a frame buffer memory; a command block for receiving geometry data and outputting initial geometric primitive data; a plurality of floating point blocks for receiving said initial geometric primitive data from said command block, wherein each of said plurality of floating point blocks is configured to perform geometry processing operations on said initial geometric primitive data in order to produce modified geometric primitive data, and a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks, wherein said command block couples to each of said plurality of floating point blocks through said first plurality of direct data channels, wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and one of said plurality of floating point blocks, wherein said command block includes round robin arbitration logic coupled to said first plurality of direct data channels, wherein said round robin arbitration logic operates to provide data to each of said plurality of floating point blocks in a substantially round robin fashion.
Claim 10
Dependent Claim
73 words
The 3-D graphics accelerator of claim 6, wherein said round robin arbitration logic determines which of the plurality of floating point processors is to receive a next packet of said initial geometric primitive data; wherein said round robin arbitration logic provides said next packet of said initial geometric primitive data to a determined one of said plurality of floating point blocks via a corresponding one of said first plurality of direct data channels.
Claim 11
Dependent Claim
53 words
The 3-D graphics accelerator of claim 6, wherein said round robin arbitration logic operates to distribute said initial geometric primitive data substantially evenly between said plurality of floating point blocks; wherein said round robin arbitration logic maintains a substantially even flow of data across each of said first plurality of direct data channels.
Claim 12
Dependent Claim
55 words
The 3-D graphics accelerator of claim 8, wherein said round robin arbitration logic utilizes a next-available round robin arbitration scheme, wherein said round robin arbitration logic skips a particular one of said first plurality of direct data channels if said particular one of said first plurality of direct data channels is unable to transfer data.
Claim 13
Independent Claim
206 words
A method of transferring geometry data in a 3-D graphics accelerator, comprising: transferring geometry data from a memory to a command block; assembling said geometry data into initial geometric primitive data within said command block; conveying packets of said initial geometric primitive data from said command block to individual ones of a plurality of floating point blocks, wherein said conveying said packets of said initial geometric primitive data comprises transferring said initial geometric primitive data over individual ones of a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and a respective one of said plurality of floating point blocks; wherein said conveying said packets of said initial geometric primitive data from said command block comprises: determining that a particular one of the plurality of floating point blocks is to receive a next packet of said initial geometric primitive data; and transferring said initial geometric primitive data to said particular one of said plurality of floating point blocks; wherein data is transferred from said command block to said plurality of floating point blocks in a substantially round robin fashion.
Claim 14
Independent Claim
206 words
A method of transferring geometry data in a 3-D graphics accelerator, comprising: transferring geometry data from a memory to a command block; assembling said geometry data into initial geometric primitive data within said command block; conveying packets of said initial geometric primitive data from said command block to individual ones of a plurality of floating point blocks, wherein said conveying said packets of said initial geometric primitive data comprises transferring said initial geometric primitive data over individual ones of a first plurality of direct data channels coupled between said command block and said plurality of floating point blocks wherein each of said first plurality of direct data channels comprises a point-to-point connection between said command block and a respective one of said plurality of floating point blocks; wherein said conveying said packets of said initial geometric primitive data from said command block comprises: determining that a particular one of the plurality of floating point blocks is to receive a next packet of said initial geometric primitive data; and transferring said initial geometric primitive data to said particular one of said plurality of floating point blocks; wherein data is transferred from said command block to said plurality of floating point blocks in a substantially round robin fashion.
Claim 15
Dependent Claim
87 words
The method of claim 10, wherein said conveying said packets of said initial geometric primitive data from said command block comprises transferring said initial geometric primitive data to each of said plurality of floating point blocks in a substantially round robin fashion to distribute primitive data substantially evenly between said plurality of floating point blocks; wherein said conveying said packets of said initial geometric primitive data from said command block maintains a substantially even flow of data across each of said first plurality of direct data channels.
Claim 16
Dependent Claim
71 words
The method of claim 11, wherein said conveying said packets of said initial geometric primitive data from said command block utilizes a next-available round robin arbitration scheme, wherein said conveying said packets of said initial geometric primitive data from said command block skips a particular one of said first plurality of direct data channels if said particular one of said first plurality of direct data channels is unable to transfer data.
References Cited (Backward Citations)
US5345554A
Cited
US5434967A
Cited
US5392393A
Cited
US5408605A
Cited
US5440682A
Cited
US5517611A
Cited
US5463732A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2001-10-12 | 2002-04-13 | 2002-10-13 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2005-10-12 | 2006-04-13 | 2006-10-13 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2009-10-12 | 2010-04-13 | 2010-10-13 | $7700.00 | Window Closed / Lapsed |