p | p | Number of processors or cores | Sushil Prasad | Processors, cores, and nodes are alternative ways of describing the number of processing units employed. The same symbol can be used if the description of a parallel algorithm or program uses “processes” or another abstraction for processing units | |
T1* | T∗1 | Fastest sequential time | | This (usually) corresponds to the best sequential algorithm | |
Tp | Tp | Execution time of a parallel program/algorithm using p processors | | T1 is thus the time taken by a parallel program using one processor and may involve parallel overheads. As compared with this, T1* is the pure sequential time | |
Sp | Sp | (Absolute) speedup of a parallel program using p processors compared with the fastest sequential time | | Thus, the “absolute speedup” is defined as Sp=T∗1/Tp. T1/Tp is the “relative speedup,” comparing a parallel program’s performance relative to its own performance using one processor | |
Cost | | Cost of a parallel program | | Cost = pTp | |
Work | | Total work collectively done by all processors | | Work <= cost | |
Ep | Ep | Efficiency of a parallel program | | Ep=Sp/p=T∗1/pTp=T∗1/cost | |
Cost-optimal | | A parallel algorithm whose cost is of the same order as T∗1 | | | |
G | G | Graph G | | G = (V,E) | |
V | V | Set of nodes in graph G | | | |
E | E | Set of edges in graph G | | | |
n, m | n, m | Variables related to the size of the input | Anshul Gupta | Lower case letters such as n, m, etc., should be used to denote variables that determine the size of the input; for example, a set of n records, or an n × m matrix, or a graph with n vertices and m edges, etc. | |
Design pattern | | A description of an effective solution to a recurring problem in a particular context | Dick Brown | | |
N | N | Input size | Ramachandran Vaidyanathan | Uppercase is useful when N = 2k is a power of 2 | |
l | ℓ | Level of node in a tree | Ramachandran Vaidyanathan | | |
t | ti | Time-related quantity | Ramachandran Vaidyanathan | | |
x’, X’ | x′ or X′ | Complement of a Boolean variable x or X | Ramachandran Vaidyanathan | | |
< a,b > | < a,b > | Doublet or ordered pair < a,b > | Ramachandran Vaidyanathan | | |
oplus | ⊕ | Generic associative operator | Thomas Cormen | When describing reduce and scan operations, we need a generic operator | |
@ | @ | Operator to perform segmented scans via unsegmented scans | Thomas Cormen | When performing segmented scan operations, we use a special operator to use the mechanism for unsegmented scans | |