Patent US5734829A successfully refreshed from Google Patents.
US5734829A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Method and program for processing a volume of data on a parallel computer system
Assignee: International Business Machines Corp
•
Filing: 1995-10-20
•
Granted: 1998-03-31
•
Expires: 2015-10-20
Abstract
Abstract A method and program for processing a volume of data such as seismic data on a parallel computer system having a distributed memory system, a host node and a plurality of worker nodes. The output data volume is sliced into horizontal slices and one or more slices are assigned to each worker node. Each worker node performs Kirchoff 3D DMO operations on that portion of the input traces for processing the horizontal slice or slices assigned to that processing node, and passes the output slices to the host node. The host node then assembles the slices into the output data volume. Where there are M worker nodes and the number of slices exceeds M, every M th slice is assigned to a worker node in a round-robin fashion to balance the load among the worker nodes.
Inventors
- 👤 Billy Ray Robinson
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-07 13:34
CPC Classifications
G; G01; G01V; G01V1/00; G01V1/28
Claims Analysis
Total Claims: 20
Claim 1
Independent Claim
127 words
In a multiprocessor parallel computer system having distributed memory, a method for processing data having at least three dimensions, the method comprising: allocating one of said processors of said parallel computer system as a host node for distributing said data; allocating of multiple processors as worker nodes for processing the data in parallel; dividing said data into slices along one of the dimensions of the data representing depth; distributing each slice to memory segments of said distributed memory; and assigning each worker node to memory segments of said distributed memory to which a slice has been assigned such that contiquous slices are assigned to different worker nodes, each worker node processing a slice in parallel with and independent from the slices being processed by other worker nodes.
Claim 2
Dependent Claim
127 words
In a multiprocessor parallel computer system having distributed memory, a method for processing data having at least three dimensions, the method comprising: allocating one of said processors of said parallel computer system as a host node for distributing said data; allocating of multiple processors as worker nodes for processing the data in parallel; dividing said data into slices along one of the dimensions of the data representing depth; distributing each slice to memory segments of said distributed memory; and assigning each worker node to memory segments of said distributed memory to which a slice has been assigned such that contiquous slices are assigned to different worker nodes, each worker node processing a slice in parallel with and independent from the slices being processed by other worker nodes.
Claim 3
Dependent Claim
26 words
The method according to claim 1 further comprising assigning said worker nodes so as to balance the load of processing said data among said worker nodes.
Claim 4
Dependent Claim
141 words
In a multiprocessor parallel computer system having distributed memory, a method for processing data having at least three dimensions, the method comprising: allocating one of said processors of said parallel computer system as a host node for distributing said data; allocating multiple processors into M worker nodes for processing data in parallel; storing said data in said distributed memory of said parallel computer system; dividing said data into M slabs along one dimension of the data representing depth; subdividing each slab of data into M slices along said one dimension; assigning a respective slice of each slab to the worker nodes, respectively, for processing the data in parallel; and as the processing of data in each slice is completed by a worker node, assigning the respective slice of the next slab, in turn, to the worker node that has thus completed.
Claim 5
Dependent Claim
141 words
In a multiprocessor parallel computer system having distributed memory, a method for processing data having at least three dimensions, the method comprising: allocating one of said processors of said parallel computer system as a host node for distributing said data; allocating multiple processors into M worker nodes for processing data in parallel; storing said data in said distributed memory of said parallel computer system; dividing said data into M slabs along one dimension of the data representing depth; subdividing each slab of data into M slices along said one dimension; assigning a respective slice of each slab to the worker nodes, respectively, for processing the data in parallel; and as the processing of data in each slice is completed by a worker node, assigning the respective slice of the next slab, in turn, to the worker node that has thus completed.
Claim 6
Dependent Claim
133 words
In a multiprocessor parallel computer system having distributed memory, a method for processing seismic data comprising: allocating one of said processors of said parallel computer system as a host node for distributing said seismic data; allocating of multiple processors as worker nodes for processing seismic data in parallel; dividing said seismic data into time slices, each time slice representing a horizontal slice of the volume represented by the seismic data; distributing each time slice to memory segments of said distributed memory; and assigning each worker node to memory segments of said distributed memory to which a time slice has been assigned such that contiguous slices are assigned to different worker nodes, each worker node processing a time slice in parallel with and independent from the time slices being processed by other worker nodes.
Claim 7
Dependent Claim
133 words
In a multiprocessor parallel computer system having distributed memory, a method for processing seismic data comprising: allocating one of said processors of said parallel computer system as a host node for distributing said seismic data; allocating of multiple processors as worker nodes for processing seismic data in parallel; dividing said seismic data into time slices, each time slice representing a horizontal slice of the volume represented by the seismic data; distributing each time slice to memory segments of said distributed memory; and assigning each worker node to memory segments of said distributed memory to which a time slice has been assigned such that contiguous slices are assigned to different worker nodes, each worker node processing a time slice in parallel with and independent from the time slices being processed by other worker nodes.
Claim 8
Dependent Claim
27 words
The method according to claim 4 further comprising assigning said worker nodes so as to balance the load of processing said seismic data among said worker nodes.
Claim 9
Dependent Claim
18 words
The method according to claim 5 wherein said data processing of seismic data comprises Kirchhoff 3D DMO calculations.
Claim 10
Dependent Claim
143 words
In a multiprocessor parallel computer system having distributed memory, a method for processing seismic data comprising: allocating one of said processors of said parallel computer system as a host node for distributing said seismic data; allocating multiple processors into M worker nodes for processing seismic data in parallel; storing said seismic data in said distributed memory of said parallel computer system; dividing said seismic data into M horizontal slabs of data from the volume represented by the seismic data; subdividing each slab of data into M time slices; assigning a respective time slice of each slab to the worker nodes, respectively, for processing seismic data in parallel; and as the processing of seismic data in each time slice is completed by a worker node, assigning the respective time slice of the next slab, in turn, to the worker node that has thus completed.
Claim 11
Dependent Claim
143 words
In a multiprocessor parallel computer system having distributed memory, a method for processing seismic data comprising: allocating one of said processors of said parallel computer system as a host node for distributing said seismic data; allocating multiple processors into M worker nodes for processing seismic data in parallel; storing said seismic data in said distributed memory of said parallel computer system; dividing said seismic data into M horizontal slabs of data from the volume represented by the seismic data; subdividing each slab of data into M time slices; assigning a respective time slice of each slab to the worker nodes, respectively, for processing seismic data in parallel; and as the processing of seismic data in each time slice is completed by a worker node, assigning the respective time slice of the next slab, in turn, to the worker node that has thus completed.
Claim 12
Dependent Claim
18 words
The method according to claim 7 wherein said data processing of seismic data comprises Kirchhoff 3D DMO calculations.
Claim 13
Dependent Claim
124 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing data having at least three dimensions comprising: dividing said data into slices along one of the dimensions of the data representing depth; distributing by said host node, each slice to memory segments of said distributed memory; and assigning by said host node, each worker node to memory segments of said distributed memory to which a slice has been assigned such that contiguous slices are assigned to different worker nodes, each worker node processing a slice in parallel with and independent from the slices being processed by other worker nodes.
Claim 14
Dependent Claim
124 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing data having at least three dimensions comprising: dividing said data into slices along one of the dimensions of the data representing depth; distributing by said host node, each slice to memory segments of said distributed memory; and assigning by said host node, each worker node to memory segments of said distributed memory to which a slice has been assigned such that contiguous slices are assigned to different worker nodes, each worker node processing a slice in parallel with and independent from the slices being processed by other worker nodes.
Claim 15
Dependent Claim
151 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing data having at least three dimensions comprising: storing said data in said distributed memory of said parallel computer system; dividing by said host node, said data into M slabs along one dimension of the data, where M is equal to the number of worker nodes; subdividing by said host node, each slab of data into M slices along said one dimension; assigning by said host node, a respective slice of each slab to the worker nodes, respectively, for processing the data in parallel; and as the processing of data in each slice is completed by a worker node, assigning the respective slice of the next slab, in turn, to the worker node that has thus completed.
Claim 16
Dependent Claim
151 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing data having at least three dimensions comprising: storing said data in said distributed memory of said parallel computer system; dividing by said host node, said data into M slabs along one dimension of the data, where M is equal to the number of worker nodes; subdividing by said host node, each slab of data into M slices along said one dimension; assigning by said host node, a respective slice of each slab to the worker nodes, respectively, for processing the data in parallel; and as the processing of data in each slice is completed by a worker node, assigning the respective slice of the next slab, in turn, to the worker node that has thus completed.
Claim 17
Dependent Claim
123 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing seismic data comprising: dividing said seismic data into time slices, each time slice representing a horizontal slice of the volume represented by the seismic data; distributing each time slice to memory segments of said distributed memory; and assigning each worker node to memory segments of said distributed memory to which a time slice has been assigned such that contiguous slices are assigned to different worker nodes, each worker node processing a time slice in parallel with and independent from the time slices being processed by other worker nodes.
Claim 18
Dependent Claim
123 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing seismic data comprising: dividing said seismic data into time slices, each time slice representing a horizontal slice of the volume represented by the seismic data; distributing each time slice to memory segments of said distributed memory; and assigning each worker node to memory segments of said distributed memory to which a time slice has been assigned such that contiguous slices are assigned to different worker nodes, each worker node processing a time slice in parallel with and independent from the time slices being processed by other worker nodes.
Claim 19
Dependent Claim
133 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing seismic data comprising: storing said seismic data in said distributed memory of said parallel computer system; dividing said seismic data into M horizontal slabs of data from the volume represented by the seismic data; subdividing each slab of data into M time slices; assigning a respective time slice of each slab to the worker nodes, respectively, for processing seismic data in parallel; and as the processing of seismic data in each time slice is completed by a worker node, assigning the respective time slice of the next slab, in turn, to the worker node that has thus completed.
Claim 20
Dependent Claim
133 words
A program recorded on a computer readable medium for use in a multiprocessor parallel computer system having distributed memory, a host node and a plurality of worker nodes, the program performing a method of processing seismic data comprising: storing said seismic data in said distributed memory of said parallel computer system; dividing said seismic data into M horizontal slabs of data from the volume represented by the seismic data; subdividing each slab of data into M time slices; assigning a respective time slice of each slab to the worker nodes, respectively, for processing seismic data in parallel; and as the processing of seismic data in each time slice is completed by a worker node, assigning the respective time slice of the next slab, in turn, to the worker node that has thus completed.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2001-03-30 | 2001-09-29 | 2002-03-31 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2005-03-30 | 2005-09-29 | 2006-03-31 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2009-03-30 | 2009-09-29 | 2010-03-31 | $7700.00 | Window Closed / Lapsed |