Embedding hex/cells into tree/hypercube networks
Qatawinah, Awad Ibrahim
Abstract
Qatawinah, Awad Ibrahim
Abstract
Graph embedding or graph mapping is an important aspect for interconnection networks used for communication between processors in parallel systems. Some parallel algorithms use communication structures which can be represented by hex-cells. In order to run these algorithms on a tree-hypercube multiprocessor system, without changing the current topology and the running application, their communication graphs need to be embedded into tree- hypercube. In this research, we have designed an algorithm for embedding hex-cells of n nodes into tree-hypercube TH(2,d) where d >= 2. The embedding has dilation one, congestion one, and expansion 1.1. In the algorithm an embedding of irregular shape of hex-cells into tree-hypercube TH(2,d) where d >= 2 is performed.
OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Graph embedding or graph mapping is an important aspect for interconnection networks used for communication between processors in parallel systems. Some parallel algorithms use communication structures which can be represented by hex-cells. In order to run these algorithms on a tree-hypercube multiprocessor system, without changing the current topology and the running application, their communication graphs need to be embedded into tree- hypercube. In this research, we have designed an algorithm for embedding hex-cells of n nodes into tree-hypercube TH(2,d) where d >= 2. The embedding has dilation one, congestion one, and expansion 1.1. In the algorithm an embedding of irregular shape of hex-cells into tree-hypercube TH(2,d) where d >= 2 is performed.
Key concepts: Hypercube, Embedding, Multiprocessing, Computer science, Dilation (metric space), Tree (set theory), Interconnection, Binary tree