1-6hit |
David ALEDO Benjamin CARRION SCHAFER Félix MORENO
This paper describes the advantages and disadvantages observed when describing complex parameterizable Artificial Neural Networks (ANNs) at the behavioral level using SystemC and at the Register Transfer Level (RTL) using VHDL. ANNs are complex to parameterize because they have a configurable number of layers, and each one of them has a unique configuration. This kind of structure makes ANNs, a priori, challenging to parameterize using Hardware Description Languages (HDL). Thus, it seems intuitively that ANNs would benefit from the raise in level of abstraction from RTL to behavioral level. This paper presents the results of implementing an ANN using both levels of abstractions. Results surprisingly show that VHDL leads to better results and allows a much higher degree of parameterization than SystemC. The implementation of these parameterizable ANNs are made open source and are freely available online. Finally, at the end of the paper we make some recommendation for future HLS tools to improve their parameterization capabilities.
Kun-Lin TSAI I-Jui TUNG Feipei LAI
Content addressable memory is widely used for fast lookup table data searching, but it often consumes considerable power. Moreover, designing the suitable content addressable memory architecture for a specific application also consumes lots of time, since the behavioral simulation is often done in the transistor level. SystemC is a system-level modeling language and simulation platform, providing high simulation efficiency for hardware software co-design. Unfortunately, SystemC does not provide the function for estimating power dissipation of a structure design. In this paper, a SystemC-based fast content addressable memory power estimation method is presented for estimating the power dissipation of the match-line circuit, the search-line circuit, and the storage cell array of content addressable memory in the early design stage. The mathematical equations and behavioral patterns are used as the inputs of power estimation model. The simulation results based on 10 Mibench benchmarks show that the simulation time of the proposed method is in average 1233 times faster than that of HSPICE simulator with only 3.51% error rate.
Sung-Rok YOON Min Li HUANG Sangho SEO Hiroshi OCHI Sin-Chong PARK
This paper presents a fast and systematic architecture exploration method that realizes an efficient IEEE 802.11e based hardware/software co-design Medium Access Control (MAC) system architecture, which can achieve near theoretical MAC throughput for burst data transmission while complying with strict channel access time requirements. Our design approach uses SystemC based Transaction Level Modeling (TLM) framework to integrate reconfigurable general purpose computing and communication resources into the application model for rapid evaluation of core parameters, system performance, and application specific optimizations. As a result, a MAC system architecture that achieves a simulated MAC throughput of more than 100 Mbps when transmitted at 260 Mbps of Physical Layer (PHY) data rate is obtained. This result is verified with X-X-IMPLEMENTATION on a Xilinx Field-Programmable Gate Array (FPGA) board.
In this paper, we describe a verification environment which is based on a constrained random layered testbench using SystemVerilog OOP. As SystemVerilog OOP technique does not allow multiple inheritance, we adopt SystemC to design components of a verification environment which employ multiple inheritance. Then SystemC design unit is linked to a SystemVerilog-based verification environment using SystemVerilog DPI and ModelSim macro. Employing multiple inheritance of SystemC makes the design phase of verification environment simple and easy through source code reusability without corruption due to multi-level single inheritance.
Liang-Bi CHEN Chi-Tsai YEH Hung-Yu CHEN Ing-Jer HUANG
3D graphics application is widely used in consumer electronics which is an inevitable tendency in the future. In general, the higher abstraction level is used to model a complex system like 3D graphics SoC. However, the concerned issue is that how to use efficient methods to traverse design space hierarchically, reduce simulation time, and refine the performance fast. This paper demonstrates a system-level design space exploration model for a tile-based 3D graphics SoC refinement. This model uses UML tools which can assist designers to traverse the whole system and reduces simulation time dramatically by adopting SystemC. As a result, the system performance is improved 198% at geometry function and 69% at rendering function, respectively.
Taekyu KIM Jin LEE Seungbeom LEE Sin-Chong PARK
Tracking a large quantity of moving target tags simultaneously is essential for the localization and guidance of people in welfare facilities like hospitals and sanatoriums for the aged. The locating system using active RFID technology consists of a number of fixed RFID readers and tags carried by the target objects, or senior people. We compare the performances of several determination algorithms which use the power measurement of received signals emitted by the moving active RFID tags. This letter presents a study on the effect of collision in tracking large quantities of objects based on active RFID real time location system (RTLS). Traditional trilateration, fingerprinting, and well-known LANDMARC algorithm are evaluated and compared with varying number of moving tags through the SystemC-based computer simulation. From the simulation, we show the tradeoff relationship between the number of moving tags and estimation accuracy.