System, method, and recording medium for mirroring matrices for batched cholesky decomposition on a graphic processing unit
AssigneeInternational Business Machines Corporation
InventorsMinsik Cho, David Shing-ki Kung, Ruchir Puri
Life of this patent
Estimating the term…
Abstract
A batched Cholesky decomposition method, system, and non-transitory computer readable medium for a Graphics Processing Unit (GPU) including at least a first problem and a second problem, include mirroring a second problem matrix of the second problem to a first problem matrix of the first problem, combining the first problem matrix and the mirrored second problem matrix into a single problem matrix, and allocating data read to a thread and to the first problem and the second problem, respectively.
Claims20
A batched Cholesky decomposition method for a Graphics Processing Unit (GPU) including at least a first problem and a second problem, the method comprising:
- mirroring a second problem matrix of the second problem to a first problem matrix of the first problem;
- combining the first problem matrix and the mirrored second problem matrix into a single problem matrix;
- and
- allocating data read to a thread and to the first problem and the second problem, respectively.
A non-transitory computer-readable recording medium recording a batched Cholesky decomposition program for a Graphics Processing Unit (GPU) including at least a first problem and a second problem, the program causing a computer to perform:
- mirroring a second problem matrix of the second problem to a first problem matrix of the first problem;
- combining the first problem matrix and the mirrored second problem matrix into a single problem matrix;
- and
- allocating data read to a thread and to the first problem and the second problem, respectively.
A batched Cholesky decomposition system for at least a first problem and a second problem on a Graphics Processing Unit (GPU), said system comprising:
- a processor;
- and
- a memory, the memory storing instructions to cause the processor to:
- mirror a second problem matrix of the second problem to a first problem matrix of the first problem;
- combine the first problem matrix and the mirrored second problem matrix into a single problem matrix;
- and
- allocate data read to a thread and to the first problem and the second problem, respectively.
Dependencies are inferred from the claim text.