Accelerating MATLAB with GPU Computing

Index

Note: Page numbers followed by “f” and “b” refers to figures and boxes respectively.

A

Accelerating MATLAB without GPU, 1

code readability and optimization, 10

file read/write, minimizing, 10

for-loop, 7, 109–111

memory preallocation, 6–7

sparse matrix form, 7–9

variable class, changing, 10

vectorization, 1–6

elementwise operation, 2–3

useful tricks, 4–6

vector/matrix operation, 3–4

Atlas-based segmentation, 193–194

MATLAB codes for, 194–203

atlasSeg_Cuda.m, 228, 228f, 229f

B

Binary segmentation mask, 193, 194f

BLAS (basic linear algebra subroutines), 127

C

C/C++ compilers, 19, 20, 74, 121

cal_regularization, 206, 208f, 209f, 210, 215

Cluster Profile Manager window, 110f

c-mex, 19, 175–179, 180–189, 227

code profiling for CUDA, 52–64