Journal
Scientific and technical journal «Priborostroenie»
UDK519.688
Issue:10 (55)
Download PDF496 Kbyte
ANALYSIS OF CUDA EFFICIENCY IN SOLVING LINEAR TRIDIAGONAL SYSTEMS FOR THEORETICAL OPTION PRICING
Annotation
Parallel cyclic reduction method for solving linear tridiagonal systems is implemented on GPU. The advisability of matrix formation directly in GPU global memory is shown. The approach provides a more than 20-fold acceleration as compared to single-threaded calculation. With the account for data transfer between RAM and GPU, a 5—8-fold acceleration is attained with the use of mapped memory.
Keywords
Постоянный URL
Articles in current issue
- CONTEXT MANAGEMENT IN INFORMATION SYSTEMS
- APPLICATION OF CUDA TECHNIQUE TO QUICKEN CALCULATION OF EUROPEAN OPTION PRICES BY FINITE-DIFFERENCE METHOD
- STUDY OF TRANSFORM CODING ALGORITHMS IN COMPRESSION OF VIDEO SEQUENCE FRAMES
- NEURAL NETWORK METHODS OF PERSON IDENTIFICATION BY FACE IMAGE
- MODEL OF CONTROL OVER ACCESS TO NEWLY CREATED FILE OBJECTS
- CONVERSION OF ALGORITHM MODELS
- CREATING A CAD SIMULATOR OF A ROUTED COMPUTER NETWORK USING OPEN SOURCE COMPONENTS
- FUNCTIONAL RELIABILITY OF COMPUTING SYSTEMS WITH REDISTRIBUTION OF INQUIRIES
- PROBLEMS OF SYNTHESIS OF SYSTEMS WITH LOSSES
- PRIORITY-BASED MECHANISMS OF SERVICE QUALITY PROVISION IN MULTISERVICE COMPUTER NETWORKS
- COMPOSITE CODE SCALES FOR CONVERTERS OF LINEAR MOVEMENTS
- DESIGN OF INSTRUMENT CONTROLLERS
- ASPECT TECHNOLOGIES IN DEVELOPMENT OF SW-INTENSIVE SYSTEMS
- A COMPUTER-AIDED SYSTEM FOR PSEUDO-REGULAR CODE SCALE DESIGN
- METHOD OF REVERSE RANDOM SAMPLING FOR MEASUREMENT OF TIME DELAY IN ELEMENT OF THE INTEGRATED CIRCUIT WITH PICOSECOND ACCURACY