Technical Note
CUDASW++: optimizing Smith-Waterman sequence database searches for CUDA-enabled graphics processing units
-
* Corresponding author: Yongchao Liu liuy0039@ntu.edu.sg
School of Computer Engineering, Nanyang Technological University, Singapore
BMC Research Notes 2009, 2:73 doi:10.1186/1756-0500-2-73
Published: 6 May 2009Additional files
Additional file 1:
Data dependencies in the alignment matrix for SW algorithm. This figure demonstrates the data dependencies in the alignment matrix for the Smith-Waterman algorithm.
Format: PNG Size: 79KB Download file
Additional file 2:
Execution model of CUDA-enabled GPUs. This figure demonstrates the execution model of CUDA-enabled GPUs, where serial code executes on the host while parallel code executes on the device.
Format: PNG Size: 58KB Download file
Additional file 3:
Hardware model (Tesla) of CUDA-enabled GPUs. This figure demonstrates the hardware model (Tesla) of CUDA-enabled GPUs consisting of a set of SIMT multiprocessors with on-chip shared memory, constant cache, texture cache and device memory.
Format: PNG Size: 111KB Download file
Additional file 4:
The arrangement of the subject sequences in the database. This figure demonstrates the arrangement of the subject sequences in the database for the inter-task and intra-task parallelization: (a) subject sequences arrangement for the inter-task parallelization and (b) subject sequences arrangement for the intra-task parallelization.
Format: PNG Size: 49KB Download file
Additional file 5:
Two global memory allocation patterns for processing entities. This figure demonstrates two global memory allocation patterns of a basic type variable of size N for M processing entities (threads or thread blocks).
Format: PNG Size: 146KB Download file
