• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.459-461
• /
• 2022

One of the big differences between Network-on-Chip (NoC) and the existing parallel processing system based on an off-chip network is that data packet routing is performed using a centralized control scheme. In such an environment, the best-effort packet routing problem becomes a real-time assignment problem in which data packet arriving time and processing time is the cost. In this paper, the Hungarian algorithm, a representative computational complexity reduction algorithm for the linear algebraic equation of the allocation problem, is implemented in the form of a hardware accelerator. As a result of logic synthesis using the TSMC 0.18um standard cell library, the area of the circuit designed through case analysis for the cost distribution is reduced by about 16% and the propagation delay of it is reduced by about 52%, compared to the circuit implementing the original operation sequence of the Hungarian algorithm.

• Journal of the Korean earth science society
• /
• v.38 no.7
• /
• pp.496-510
• /
• 2017

This study presents a new nationwide quantitative precipitation estimation (QPE) based on the hybrid surface rainfall (HSR) technique using the weather radar network of Korea Meteorological Administration (KMA). This new nationwide HSR is characterized by the synthesis of reflectivity at the hybrid surface that is not affected by ground clutter, beam blockage, non-meteorological echoes, and bright band. The nationwide HSR is classified into static (STATIC) and dynamic HSR (DYNAMIC) mosaic depending on employing a quality control process, which is based on the fuzzy logic approach for single-polarization radar and the spatial texture technique for dual-polarization radar. The STATIC and DYNAMIC were evaluated by comparing with official and operational radar rainfall mosaic (MOSAIC) of KMA for 10 rainfall events from May to October 2014. The correlation coefficients within the block region of STATIC, DYNAMIC and MOSAIC are 0.52, 0.78, and 0.69, respectively, and their mean relative errors are 34.08, 30.08, and 40.71%.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.37 no.4
• /
• pp.80-89
• /
• 2000

In this paper, we propose a new CPLD technology mapping algorithm for sequential circuit under time constraints. The algorithm detects feedbacks of sequential circuit, separate each feedback variables into immediate input variable, and represent combinational part into DAG. Also, among the nodes of the DAG, the nodes that the number of outdegree is more than or equal to 2 is not separated, but replicated from the DAG, and reconstructed to fanout-free-tree. To use this construction method is for reason that area is less consumed than the TEMPLA algorithm to implement circuits, and process time is improved rather than TMCPLD within given time constraint. Using time constraint and delay of device the number of partitionable multi-level is defined, the number of OR terms that the initial costs of each nodes is set to and total costs that the$^1$costs is set to after merging nodes is calculated, and the nodes that the number of OR terms of CLBs that construct CPLD is excessed is partitioned and is reconstructed as subgraphs. The nodes in the partitioned subgraphs is merged through collapsing, and the collapsed equations is performed by bin packing so that it fit to the number of OR terms in the CLBs of a given device. In the results of experiments to MCNC circuits for logic synthesis benchmark, we can shows that proposed technology mapping algorithm reduces the number of CLBs by 15.58% rather than the TEMPLA, and reduces process time rather than the TMCPLD.

• Journal of IKEEE
• /
• v.24 no.3
• /
• pp.901-908
• /
• 2020

This paper describes a method for scalable hardware implementation of modular inversion. The proposed scalable architecture has a one-dimensional array of processing elements (PEs) that perform arithmetic operations in 32-bit word, and its performance and hardware size can be adjusted depending on the number of PEs used. The hardware operation of the scalable processor for modular inversion was verified by implementing it on Spartan-6 FPGA device. As a result of logic synthesis with a 180-nm CMOS standard cells, the operating frequency was estimated to be in the range of 167 to 131 MHz and the gate counts were in the range of 60,000 to 91,000 gate equivalents when the number of PEs was in the range of 1 to 10. When calculating 256-bit modular inverse, the average performance was 18.7 to 118.2 Mbps, depending on the number of PEs in the range of 1 to 10. Since our scalable architecture for computing modular inversion in GF(p) has the trade-off relationship between performance and hardware complexity depending on the number of PEs used, it can be used to efficiently implement modular inversion processor optimized for performance and hardware complexity required by applications.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.1
• /
• pp.224-234
• /
• 2000

In this paper, we propose a new CPLD technology mapping algorithm for sequential circuit under time constraints. The algorithm detects feedbacks of sequential circuit, separate each feedback variables into immediate input variable, and represent combinational part into DAG. Also, among the nodes of the DAG, the nodes that the number of outdegree is more than or equal to 2 is not separated, but replicated from the DAG, and reconstructed to fanout-free-tree. To use this construction method is for reason that area is less consumed than the TEMPLA algorithm to implement circuits, and process time is improved rather than TMCPLD within given time constraint. Using time constraint and delay of device the number of partitionable multi-level is defined, the number of OR terms that the initial costs of each nodes is set to and total costs that the costs is set to after merging nodes is calculated, and the nodes that the number of OR terms of CLBs that construct CPLD is excessed is partitioned and is reconstructed as subgraphs. The nodes in the partitioned subgraphs is merged through collapsing, and the collapsed equations is performed by bin packing so that if fit to the number of OR terms in the CLBs of a given device. In the results of experiments to MCNC circuits for logic synthesis benchmark, we can shows that proposed technology mapping algorithm reduces the number of CLBs bu 15.58% rather than the TEMPLA, and reduces process time rather than the TMCPLD.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.38 no.1
• /
• pp.79-85
• /
• 2001

In this paper, we propose a new technology mapping algorithm for CPLD consider area under time constraint(TMFCPLD). This technology mapping algorithm detect feedbacks from boolean networks, then variables that have feedback are replaced to temporary variables. Creating the temporary variables transform sequential circuit to combinational circuit. The transformed circuits are represented to DAG. After traversing all nodes in DAG, the nodes that have output edges more than two are replicated and reconstructed to fanout free tree. This method is for reason to reduce area and improve total run time of circuits by TEMPLA proposed previously. Using time constraints and delay time of device, the number of graph partitionable multi-level is decided. Initial cost of each node are the number of OR-terms that it have. Among mappable clusters, clusters of which the number of multi-level is least is selected, and the graph is partitioned. Several nodes in partitioned clusters are merged by collapsing, and are fitted to the number of OR-terms in a given CLB by bin packing. Proposed algorithm have been applied to MCNC logic synthesis benchmark circuits, and have reduced the number of CLBs by 62.2% than those of DDMAP. And reduced the number of CLBs by 17.6% than those of TEMPLA, and reduced the number of CLBs by 4.7% than those of TMCPLD. This results will give much efficiency to technology mapping for CPLDs.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.1
• /
• pp.178-184
• /
• 2015

This paper proposes a hardware architecture for high performance Deblocking filter(DBF) in High Efficiency Video Coding for UHD(Ultra High Definition) videos. This proposed hardware architecture which has less processing time has a 4-stage pipelined architecture with two filters and parallel boundary strength module. Also, the proposed filter can be used in low-voltage design by using clock gating architecture in 4-stage pipeline. The segmented memory architecture solves the hazard issue that arises when single port SRAM is accessed. The proposed order of filtering shortens the delay time that arises when storing data into the single port SRAM at the pre-processing stage. The DBF hardware proposed in this paper was designed with Verilog HDL, and was implemented with 22k logic gates as a result of synthesis using TSMC 0.18um CMOS standard cell library. Furthermore, the dynamic frequency can process UHD 8k($7680{\times}4320$) samples@60fps using a frequency of 150MHz with an 8K resolution and maximum dynamic frequency is 285MHz. Result from analysis shows that the proposed DBF hardware architecture operation cycle for one process coding unit has improved by 32% over the previous one.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2013.08a
• /
• pp.289-289
• /
• 2013

While there are plenty of studies on synthesizing semiconducting germanium nanowires (Ge NWs) by vapor-liquid-solid (VLS) process, it is difficult to inject dopants into them with uniform dopants distribution due to vapor-solid (VS) deposition. In particular, as precursors and dopants such as germane ($GeH_4$), phosphine ($PH_3$) or diborane ($B_2H_6$) incorporate through sidewall of nanowire, it is hard to obtain the structural and electrical uniformity of Ge NWs. Moreover, the drastic tapered structure of Ge NWs is observed when it is synthesized at high temperature over $400^{\circ}C$ because of excessive VS deposition. In 2006, Emanuel Tutuc et al. demonstrated Ge NW pn junction using p-type shell as depleted layer. However, it could not be prevented from undesirable VS deposition and it still kept the tapered structures of Ge NWs as a result. Herein, we adopt $C_2H_2$ gas in order to passivate Ge NWs with carbon sheath, which makes the entire Ge NWs uniform at even higher temperature over $450^{\circ}C$. We can also synthesize non-tapered and uniformly doped Ge NWs, restricting incorporation of excess germanium on the surface. The Ge NWs with carbon sheath are grown via VLS process on a $Si/SiO_2$ substrate coated 2 nm Au film. Thin Au film is thermally evaporated on a $Si/SiO_2$ substrate. The NW is grown flowing $GeH_4$, HCl, $C_2H_2$ and PH3 for n-type, $B_2H_6$ for p-type at a total pressure of 15 Torr and temperatures of $480{\sim}500^{\circ}C$. Scanning electron microscopy (SEM) reveals clear surface of the Ge NWs synthesized at $500^{\circ}C$. Raman spectroscopy peaked at about ~300 $cm^{-1}$ indicates it is comprised of single crystalline germanium in the core of Ge NWs and it is proved to be covered by thin amorphous carbon by two peaks of 1330 $cm^{-1}$ (D-band) and 1590 $cm^{-1}$ (G-band). Furthermore, the electrical performances of Ge NWs doped with boron and phosphorus are measured by field effect transistor (FET) and they shows typical curves of p-type and n-type FET. It is expected to have general potentials for development of logic devices and solar cells using p-type and n-type Ge NWs with carbon sheath.