• The KIPS Transactions:PartA
• /
• v.11A no.1
• /
• pp.49-56
• /
• 2004

The performance of the multiprocessor systems is limited by the several factors. The system performance is affected by the processor speed, memory delay, and interconnection network bandwidth/latency. By the evolution of semiconductor technology, off the shelf microprocessor speed breaks beyond GHz, and the processors can be scalable up to multiprocessor system by connecting through the interconnection networks. In this situation, the system performances are bound by the latencies and the bandwidth of the interconnection networks. SCI, Myrinet, and Gigabit Ethernet are widely adopted as a high-speed interconnection network links for the high performance cluster systems. Performance improvement of the interconnection network can be achieved by the bandwidth extension and the latency minimization. Speed up of the operation clock speed is a simple way to accomplish the bandwidth and latency betterment, while its physical distance makes the difficulties to attain the high frequency clock. Hence the system performance and scalability suffered from the interconnection network limitation. Duplicating the link of the interconnection network is one of the solutions to resolve the bottleneck of the scalable systems. Dual-ring SCI link structure is an example of the interconnection network improvement. In this paper, I propose a network topology and a transaction path algorism, which optimize the latency and the efficiency under the duplicated links. By the simulation results, the proposed structure shows 1.05 to 1.11 times better latency, and exhibits 1.42 to 2.1 times faster execution compared to the dual ring systems.

• Journal of KIISE:Computer Systems and Theory
• /
• v.32 no.8
• /
• pp.393-398
• /
• 2005

In interconnection networks, a Hamiltonian path has been utilized in many applications such as the implementation of linear array and multicasting. In this paper, we consider the Hamiltonian properties of mesh networks which are used as the topology of parallel machines. If a network is strongly Hamiltonian laceable, the network has the longest path joining arbitrary two nodes. We show that a two-dimensional mesh M(m, n) is strongly Hamiltonian laceabie, if $m{\geq}4,\;n{\geq}4(m{\geq}3,\;n{\geq}3\;respectively)$, and the number of nodes is even(odd respectively). A mesh is a spanning subgraph of many interconnection networks such as tori, hypercubes, k-ary n-cubes, and recursive circulants. Thus, our result can be applied to discover the fault-hamiltonicity of such networks.

• The Journal of Korean Association of Computer Education
• /
• v.17 no.2
• /
• pp.115-124
• /
• 2014

Hypercube and star graph are widely known as interconnection network. The embedding of an interconnection network is a mapping of a network G into other network H. The possibility of embedding interconnection network G into H with a low cost, has an advantage of efficient algorithms usage in network H, which was developed in network G. In this paper, we provide an embedding algorithm between HCN and HON. HCN(n,n) can be embedded into HON($C_{n+1},C_{n+1}$) with dilation 3 and HON($C_d,C_d$) can be embedded into HCN(2d-1,2d-1) with dilation O(d). Also, star graph can be embedded to half pancake's value of dilation 11, expansion 1, and average dilation 8. Thus, the result means that various algorithms designed for HCN and Star graph can be efficiently executed on HON and half pancake, respectively.

• The KIPS Transactions:PartA
• /
• v.9A no.3
• /
• pp.333-340
• /
• 2002

Embedding is a mapping an interconnection network G to another interconnection network H. If a network G can be embedded to another network H, algorithms developed on G can be simulated on H. In this paper, we first propose a method to embed between Hierarchical Cubic Network HCN(n, n) and Hierarchical Folded-hypercube Network HFN(n, n). HCN(n, n) and HFN(n, n) are graph topologies having desirable properties of hypercube while improving the network cost, defined as degree${\times}$diameter, of Hypercube. We prove that HCN(n, n) can be embedded into HFN(n, n) with dilation 3 and congestion 2, and the average dilation is less than 2. HFN(n, n) can be embedded into HCN(n, n) with dilation 0 (n), but the average dilation is less than 2. Finally, we analyze the fault tolerance of HCN(n, n) and prove that HCN(n, n) is maximally fault tolerant.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.4
• /
• pp.1045-1057
• /
• 1999

Packet-switched multistage interconnection networks(MINs) have been widely used for digital switching systems and super computers. In this paper we show that multiple packets in a switching element can move to the succeeding switching element in one network cycle by fully utilizing the cycle bandwidth. Only one packet movement was usually assumed in typical MINs. we present an analytical model for the MNs with the multiple packet movement scheme, and validate it by computer simulation. Comparisons with the traditional MINs of single packet movement reveal that the throughput is increased up to about 30% for practical size MINs. Similar result was also obtained for delays. The performance increase is more significant when the network traffic is nonuniform.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10a
• /
• pp.418-420
• /
• 2003

최근에 하이퍼큐브의 망비용을 개선한 상호연결망 하이퍼-스타가 제안되었다. 본 논문에서는 상호연결 망 하이퍼-스타 HS(2n,n)의 평균거리가 (equation omitted)임을 보이고, 스패닝 트리를 이용한 일-대-다 방송 기법을 제안하고, 방송 수행 시간이 2n-1임을 보인다.

• The KIPS Transactions:PartA
• /
• v.18A no.5
• /
• pp.205-214
• /
• 2011

We propose a new interconnection network, called Twisted cube torus(TT) network based on well-known 3-dimensional twisted cube. Twisted cube torus network has smaller diameter and improved network cost than honeycomb torus with the same number of nodes. In this paper, we propose routing algorithm of Twisted cube torus network and analyze its diameter, network cost, bisection width and hamiltonian cycle.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.116-119
• /
• 2011

다중코어(Many-Core) 시스템은 많은 코어들이 상호연결망을 통해서 연결되어있는 시스템으로, 단일코어나 멀티코어 시스템에 비해 보다 많은 병렬 컴퓨팅 자원을 지원한다. Amdahl 의 법칙에 의하면 병렬화되어 처리하는 부분은 이론적으로 프로세서의 개수에 비례하게 가속화 될 수 있지만, 상호연결망에서의 전송 지연을 비롯한 많은 요인에 의해서 성능의 가속화가 저해된다. 특히 캐시 일관성 규약(Cache Coherence Protocol)을 지원하는 대부분의 다중코어 시스템에서는 병렬화를 함에 있어서 캐시 미스로 인해 발생하는 데이터의 전송 지연이 성능에 많은 영향을 미칠 수 있다. 따라서 효과적인 병렬 프로그램을 위해서는 캐시 구조에 대한 이해를 바탕으로 상호연결망에 대한 연구가 필요하다. 본 논문에서는 메쉬(Mesh) 구조의 64 코어 다중코어 시스템인 TilePro64 를 이용하여 상호연결망의 데이터 전송 지연에 따른 프로그램 성능의 민감도를 측정하였다. 결과적으로 코어간 거리(Hop)가 늘어날수록 작업의 수행시간이 평균적으로 4.27%씩 선형적으로 증가하는 관계가 있는 것으로 나타났다.

• The KIPS Transactions:PartA
• /
• v.16A no.6
• /
• pp.501-508
• /
• 2009

In this paper, we prove that folded hyper-star network FHS(2n,n) is node-symmetric and a bipartite network. We show that FHS(2n,n) can be embedded into odd network On+1 with dilation 2, congestion 1 and Od can be embedded into FHS(2n,n) with dilation 2 and congestion 1. Also, we show that $2n{\time}n$ torus can be embedded into FHS(2n,n) with dilation 2 and congestion 2.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10a
• /
• pp.61-63
• /
• 2004

최근에 하이퍼큐브의 망비용을 개선한 상호연결망 하이퍼-스타가 제안되었다. 본 논문에서는 삼진트리가 하이퍼-스타 연결망 HS(2n, n)에 연장율 1로 임베딩 가능함을 보인다.