Parallel computing system for modeling and data processing
First Claim
1. A parallel computing system for generating and processing voxel-based objects within a three-dimensional memory storage array represented in 3-D space, comprising:
- a system bus for transferring coordinate data sets and local programs;
a plurality of local computing units, each said local computing unit being connected to said system bus and having local program storage means for storing one or more local programs, and local processing means for locally processing one or more coordinate data sets in accordance with one or more of said local programs and producing data elements, including voxel data elements representative of a portion of a 3-D voxel-based object represented in said 3-D space;
a plurality of local memory storage modules, each said local memory storage module having a set of memory storage elements and each said memory storage element being capable of storing one or more voxel data elements, wherein a plurality of said memory storage elements in said plurality of local memory storage modules are indexed so as to represent said three-dimensional memory storage array in said 3-D space;
a plurality of local buses, each said local bus being operably connected to one said local computing unit and one said local memory storage module, for transferring voxel data elements therebetween; and
a control computing unit, operably connected to said plurality of local computing units by way of said system bus, for controlling and coordinating said plurality of local computing units during the local production and storage of said voxel data in said three-dimensional memory storage array.
9 Assignments
0 Petitions
Accused Products
Abstract
A parallel computing system useful for volume visualization, numeric, symbol, and other computing applications. The parallel computing system includes a memory storage system indexed to provide a three dimensional memory storage array, a plurality of local computing units, and a display unit. The plural computing units have the capability of accessing, in parallel and without conflict, the memory storage system. Local and global data transfer mechanisms are provided to facilitate transfer of data elements among the memory storage elements. This system can be used to generate voxel-based 3-D objects and to generate a 2-D image of any such 3-D virtual object along any viewing direction (such as by simulating parallel or perspective viewing ray processes).
92 Citations
34 Claims
-
1. A parallel computing system for generating and processing voxel-based objects within a three-dimensional memory storage array represented in 3-D space, comprising:
-
a system bus for transferring coordinate data sets and local programs; a plurality of local computing units, each said local computing unit being connected to said system bus and having local program storage means for storing one or more local programs, and local processing means for locally processing one or more coordinate data sets in accordance with one or more of said local programs and producing data elements, including voxel data elements representative of a portion of a 3-D voxel-based object represented in said 3-D space; a plurality of local memory storage modules, each said local memory storage module having a set of memory storage elements and each said memory storage element being capable of storing one or more voxel data elements, wherein a plurality of said memory storage elements in said plurality of local memory storage modules are indexed so as to represent said three-dimensional memory storage array in said 3-D space; a plurality of local buses, each said local bus being operably connected to one said local computing unit and one said local memory storage module, for transferring voxel data elements therebetween; and a control computing unit, operably connected to said plurality of local computing units by way of said system bus, for controlling and coordinating said plurality of local computing units during the local production and storage of said voxel data in said three-dimensional memory storage array. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13)
-
-
14. A method of generating 3-D voxel-based objects using a parallel computing system which includes (1) a system bus for transferring coordinate data sets and local programs, (2) a plurality of local computing units, each being connected to said system bus and having local program storage means for storing a local voxelization program, and local processing means for locally processing received coordinate data sets in accordance with said local voxelization program, (3) a plurality of local memory storage modules, each having a set of memory storage elements and each said memory storage element being capable of storing one or more voxel data elements, wherein a plurality of said memory storage elements in said plurality of local memory storage modules are indexed so as to represent a three-dimensional memory storage array in 3-D space within which the voxel data elements of said 3-D voxel-based object can be stored, (4) a plurality of local buses, each said local bus being operably connected to one said local computing unit and one said local memory storage module, for transferring voxel data elements therebetween, (5) a control computing unit, operably connected to said plurality of local computing units by way of said system bus, for coordinating said plurality of local computing units during each access to a sequence of memory storage elements in said three-dimensional memory storage array, and (6) data element buffering means, operably connected to said system bus and said plurality of local buses and being controlled by said control computing unit, for buffering data elements transferred from said local computing units over said local buses, to said data element buffering means, said method comprising:
-
(a) transferring coordinate data sets over said system bus to said plurality of local computing units; (b) each said local computing unit using said local voxelization program to locally process received coordinate data sets, and produce voxel data elements representative of a portion of a 3-D voxel-based object represented in said 3-D space; and (c) storing the produced voxel data elements of said 3-D voxel-based object in said plurality of local memory storage modules. - View Dependent Claims (15, 16, 17)
-
-
18. A parallel computing system for producing, storing, and processing voxel data elements in parallel within a three-dimensional memory storage array, said parallel computing system comprising:
-
a system bus for transferring parameters and local programs; a plurality of local computing units indexable as k=0,1,2, . . . ,N-1, each said local computing unit being connected to said system bus and having means for storing one or more local programs, including a local voxelization program, capable of locally producing, storing and processing voxel data elements within said three dimensional memory storage array, and means for executing said one or more local programs, including said local voxelization program, using said parameters so as to locally produce, store and process within said three-dimensional memory storage array, voxel data elements representative of at least a portion of 3-D geometric object; a plurality of local memory storage modules indexable as k=0,1,2, . . . ,N-1, each said local memory storage module having a set of memory storage elements and each said memory storage element being capable of storing one or more voxel data elements, wherein a plurality of said memory storage elements in said plurality of local memory storage modules are indexed so as to represent said three-dimensional memory storage array in 3-D space; a plurality of local buses indexable as k=0,1,2, . . .N-1, each said local bus being operably connected to one said local computing unit and one said local memory storage module, for transferring voxel data elements therebetween; and a control computing unit, operably connected to said plurality of local computing units by way of said system bus, for coordinating parallel execution of said one or more local programs by said plurality of local computing units. - View Dependent Claims (19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34)
-
Specification