# Introduction etwork on Chip (NoC) [1] [2] [3] is the latest approach to overcome the limitation of bus based communication network. NoC is a set of routers employed in a network, in which different nodes are inter connected with their cores can communicate with each others. In a network data comes in packets and sent to the destination with IP via routers and links [4] . When a packet reaches its destination address, it means it is switched [5] to the IP attached to the router. On-chip communications among different networks is possible using interconnection network topology [5] [6], switching, routing, queuing .flow control [11] and scheduling. Research can be done for n-dimensional topological structures network on chip design. The idea of NoC is derived from distributed computing and large scale computer networks. There are different routing techniques used in NOC design considerations to meet high throughput and cover time to market. Due to big constraints on hardware and memory resources utilization, the routing methods for NoC should be very simple. According to the need of processors, NoC technology gives chip designer's flexibility in choosing the network topology, according to University of Bologna professor Luca Benini, founder of and scientific advisor for iNoCs, a start-up provider of on-chip interconnection technology. High degree of parallelism and pipelining increases [8] [12] the performance of the system because all the links in the network works simultaneously on different data packets [13]. As the complexity of the system is increasing, NOC is the solution to enhance system performance in comparison to the previous technological architectures such as dedicated, wires, point to point, bridges and shared buses. The scalability of system and throughput [17] [20] will increase because algorithms are designed in such a way that it offers higher degree of parallelism. For example, a mesh NoC topology [18] can function with parallelism and thus is well-suited for multiprocessor SoCs, whose cores must run in parallel. Prototype NoCs by the Electronics and Information Technology Laboratory of the French Atomic Energy Commission's Faust, the Swedish Royal Institute of Technology's Nostrum, and the Technion-Israel Institute of Technology's QNoC work with a mesh topology. a) Tools Utilized Design and implementation of mesh network is carried out using Project Navigator ISE 14.2, Xilinx company. It is a tool used to design the IC and to view their RTL (Register Transfer Logic) schematic. Model Sim EE 10.1b student's edition is a tool of Mentor Graphics Company used for simulation and debugging the functionality. The chip implementation is done using VHDL programming language. The paper is organized as follows: Section I presents the introduction and the tools utilized. Section II describes intercommunication among 2D (8 x 8) mesh networks. Section III describes the FPGA synthesis environment. Section IV describes the Result and Performance Evaluation. Section V presents the Device utilization and timing summary. Conclusion is presented in Section VI. # Intercommunication among 2D Mesh Networks Intercommunication among 2D NOC networks can be understood using arbitration logic selection of networks [12]. First understanding, how a 2D networks behaves, then focusing on the intercommunication among 2D networks. 2D NOC [12] follows the cross link which allows addressing any node at any time [11]. A 2D mesh network is connecting multiple inputs to multiple outputs in a matrix form. The 2D NOC architecture is a m × n mesh of switches [10] and resources are placed on the slots formed by the switches. For an m x n architecture there are m nodes on X axis and n nodes on Y axis respectively. Considering an 8 x 8 structure in which 64 nodes can perform intra communication. Node identification is based on the row address and column address [1]. For example, if row address = 000 and column address = 101, node 6 (N 6 ) is identified. Similarly there is the possibility of identifying any node. Table 1 list the possible node address generation scheme for 2D 8 x 8 structure. Step 2 : reset = 0, same clk is used for synchronization and provide rising edge. Step 3 : Select the address of destination node Node_address [5:0] of 6 bits for 8 x 8 structure. Step 4 : Force the value of row_address and column_address of destination node. For 8 x 8 NOC row_address[2:0] and column_address[2:0] are of 3 bits. Step 5 : Force the value of network address [1:0] of destination network. network_address [1:0] = "00" for network 1, network_address [1:0] = "01" for network 2, network_address [1:0] = "10" for network 2 and network_address [1:0] = "11" for network 3. Step 6 : Give the eight bit value of data_in. Force write_en =1 and read_en =0 and then run. Step 7 : Write_en =0 and read_en =1 and run. Desired output on destination is achieved. When write_en =1 and read_en =0, the data is written in temp variable from the source node, when write_en =0 and read_en =1, the data is read from the temp variable to destination node. Clk is applied at the Table 3 : Design pins and their functional description for (8 x 8) NOC # Pins Functional Description reset used for synchronization of the components by using clk clk Provide rising edge of clock pulse node_address [5:0] Address of the source and destination node of 6 bits row_address [2:0] represents address of the nodes in x direction (3 bits) column_address [2:0] represents address of the nodes in y direction ( 3 bits) read_en control signal to read data (1 bit) write_en control signal to write data (1 bit) network_address [1:0] Selection logic for the network ( 2 bits) data_in [7:0] represents input data in the network (8 bits) data_out [7:0] represents 8 bit output data of the destination node (8 bits) IV. # Synthesis The Spartan-3E starter kit [21] [22] provides easy way to test the various programs in the FPGA itself, by dumping the 'bit' file of the designed program in Xilinx software into the FPGA and then observing the output .The Spartan 3E FPGA board [8] comes built in with many peripherals that help in the proper working of the board and also in interfacing the various signals to the board itself. Some of the peripherals included in the Spartan 3E FPGA board include: 2-line, 16-character LCD screen used for display the output, PS/2 mouse or keyboard port can be connected to the FPGA, VGA display port These switches are locked into FPGA using user constraint file (UCF). Figure 6 shows the flow of inputs given to FPGA device. Four slide switches and four push-button switches are used to give the inputs to the FPGA board. They can also act as the reset switches for the various program Kit also has four-output, SPI-based on board Digital-to-Analog Converter (DAC) on board which is to be interfaced to give the analog output to the digital data values. Two-input, SPI-based [7] [23] Analog-to-Digital Converter (ADC) with programmable gain preamplifier converts the real world analog signals into digital values.' Image processing inputs are given by the switches of kit and functionally tested on the corresponding LED's output.T he output data is flashed on LEDs. These LEDs are also locked in UCF file [16]. The bit file of the program is burn out in The EPROM of FPGA and corresponding result is shown by blinking LEDs. The output can be shown on Digital storage oscilloscope (DSO). As shown in figure 7. The input data is 10101010, when reset switch = 0, no output is display on DSO. When reset switch =1, output data is 10101010. 2013![Global Journals Inc. (US) Global Journal of Computer Science and Technology Volume XIII Issue XII Version I](image-2.png "N © 2013") 1![Figure 1 : Two dimensional (2D) cross point topological (8 8) structure[12] ](image-3.png "Figure 1 :") 2![Figure 2 : Intercommunication Mesh Network](image-4.png "©E×Figure 2 :") 3![Figure3shows the simulated result for the 8 x 8 intercommunication architecture, which shows 8 bit data transfer for network 1 to network 4. The functional simulation depends on the steps and Modelsim output is extracted after completion of these steps.a) Simulation Process SequenceStep 1 : reset = 1, clk is used for synchronization and then run.](image-5.png "Figure 3") ![Yearpositive edge clock pulse and reset is kept at 1 for the initial state. The Register transfer level (RTL) view of chip is shown yin the figure4and the details of each pin is listed in table 3.](image-6.png "") 34![Figure 3 : Modelsim Output of Mesh Intercommunication Network (2D-2D) (8 x 8)](image-7.png "Figure 3 :Figure 4 :") ![[21] used to display various encoded data on screen. The Sparten 3E kit is shown in the figure 5. Switches are the input for clk, reset, row_addressn[2:0], column_address [2:0] , network_address [1:0]. Eight bits data data_in[7:0] is also given using switches.](image-8.png "") 678![Figure 6 : Sparten-3 FPGA view [21]](image-9.png "Figure 6 :Figure 7 :Figure 8 :") 1RowColumn AddressAddress000001010011100101110111000n 1N 2N 3N 4N 5N 6N 7N 8001N 9N 10N 11N 12N 13N 14N 15N 16010N 17N 18N 19N 20N 21N 22N 23N 24011N 25N 26N 27N 28N 29N 30N 31N 32100N 33N 34N 35N 36N 37N 38N 39N 40101N 41N 42N 43N 44N 45N 46N 47N 48110N 49N 50N 51N 52N 53N 54N 55N 56111N 57N 58N 59N 60N 61N 62N 63N 64Network 1, network_address is 01 Network 2, network_address is 10 Network 3, and network_address is 11 Network 4 is identified. It is also illustrated using figure3and table 2. 2Network_AddressSelection Logic00Network 1 is selected01Network 2 is selected10Network 3 is selected11Network 4 is selectedIII. Result & Performance Evaluation 5and 6 5Logic UtilizationUsed Available UtilizationNumber of Slices335244813 %Number of Slice Flip Flops8048961 %Number of 4 input LUTs653489613 %Number of bonded IOBs3115819 %Number of GCLKs2248 %a) Timing Summary for 8 x 8 NOCTiming details provides the information of delay,minimum period, minimum input arrival time beforeclock and maximum output required time after clock [1].Speed Grade: -5Minimum Period : 4.937 ns (Maximum Frequency:202.536 MHz).VI.ConclusionThehardwarechipfor2D-2Dintercommunication network is © 2013 Global Journals Inc. (US) © 2013 Global Journals Inc. (US) Global Journal of Computer Science and Technology * Modeling and Simulation of 2D Mesh Topological Network on Chip (NOC) PrachiAgarwal AnilKumarSharma AdeshKumar International Journal of Computer Applications 72 21 June 2013 * Express Cube Topologies for On-Chip Interconnects BGrot JHestness SWKeckler OMutlu 15th International Symposium on Computer Architecture (HPCA) 2009 * Network-on-Chip design and synthesis outlook DavidAtienzaa FedericoAngiolini SrinivasanMurali AntonioPullinid LucaBenini GiovanniDeMicheli INTEGRATION, the VLSI journal Elsevier 41 2008 * A Network-on-Chip Architecture for Gigascale Systems-on-Chip, IEEE Circuits and systems magazine, second quarter DavideBertozzi LucaBenini Xpipes 2004 * An interconnect architecture for networking systems on chips F KarimANguyen SDey IEEE Journal on Micro High Performance Interconnect 22 5 Sept 2002 * JasonCong YuhuiHuang BoYuan Los Angeles Los Angeles, USA Computer Science Department University of California * 26.00 ©2011 IEEE, A Tree-Based Topology Synthesis for On-Chip Network * Research challenges for on-chip inter connection networks JDOwens WJDally IEEE MICRO 27 5 Oct. 2007 * On-chip communication architecture exploration: A quantitative evaluation of point-topoint, bus, and network-on-chip approaches HGLee NChang UYOgras RMarculescu ACM Trans. Des. Autom. Electron. Syst 12 3 Aug. 2007 * Mohammad Ayoub Khan Abdul Quaiyum Ansari, A Quadrant-XYZ Routing Algorithm for 3 * The Research Bulletin of Jordan ACM 2078-7952 II Asymmetric Torus Network-on-Chip II) pp * White paper on OCCN: A Network-On-Chip Modeling and Simulation Framework, ISD Integrated system developments, page 8. 11. Naveen Chaudhary, Bursty Communication Performance Analysis of Network-on-Chip with Diverse Traffic Permutations MCoppola SCuraba MGrammatikakis RLocatelli GMaruccia FPapariello LPieralisi International Journal of Soft Computing and Engineering (IJSCE) 2231-2307 1 January 2012 * Network on Chip for 3D Mesh Structure with Enhanced Security Algorithm in HDL Environment AdeshKumar SonalSinghal PiyushKuchhal International Journal of Computer Applications 59 17 IJCA * Performance evaluation and design trade-offs for network-on-chip interconnect architectures PPratim Pande CGrecu MJones AIvanov RSaleh IEEE Transactions on Computers 54 8 2005 * A Network on Chip Simulator Sweden Master of Science Thesis in Electronic System Design Aug 2002 Royal Institute of Technology Rikard Thid Thesis on * design challenges of technology scaling SBorkar IEEE Micro 4 2329 July-August 1999 * Semiconductor Complex Limited, Internet PDF: Data sheets of XC 95 series CPLD * 3D Network-on Chip Architectures Using Homogeneous Meshes and Heterogeneous Floorplans VitordePaulo CristinelAbabei International Journal of Reconfigurable Computing 2010 Hindawi Publishing Corporation Article ID603059 * The future of multiprocessor systems-onchips WWolf Proceedings of the 41 st Design Automation Conference (DAC'04) the 41 st Design Automation Conference (DAC'04) June 2004 * The New Torus Network Design Based On 3-Dimensional Hypercube Hyeong-OkWoo-Seo Ki1 Jae-CheolLee Oh Feb.15-18 2009 ICACT * Comparison of a Ring On-Chip Network and a Code-Division Multiple-Access On-Chip Network XinWang JariNurmi Hindawi Publishing Corporation VLSI Design Volume2007 18372 14 page