PLASMA
Parallel Linear Algebra Software for Multicore Architectures
Level 1: vectors operations, O(n) work

Vector operations that perform \( O(n) \) work on \( O(n) \) data. These are memory bound, since every operation requires a memory read or write. More...

Modules

 asum: Sum vector
 \( \sum_i |x_i| \)
 
 axpy: Add vectors
 \( y = \alpha x + y \)
 
 copy: Copy vector
 \( y = x \)
 
 dot: Dot (inner) product
 \( x^T y \) or \( x^H y \)
 
 iamax: Find max element
 \( \text{argmax}_i\; |x_i| \)
 
 iamin: Find min element
 \( \text{argmin}_i\; |x_i| \)
 
 nrm2: Vector 2 norm
 \( ||x||_2 \)
 
 rot: Apply Given's rotation
 
 rotg: Generate Given's rotation
 
 scal: Scale vector
 \( x = \alpha x \)
 
 swap: Swap vectors
 \( x <=> y \)
 

Detailed Description

Vector operations that perform \( O(n) \) work on \( O(n) \) data. These are memory bound, since every operation requires a memory read or write.