• Journal of IKEEE
• /
• v.25 no.1
• /
• pp.180-192
• /
• 2021

One widely used algorithm for image recognition and pattern detection is the convolution neural network (CNN). To efficiently handle convolution operations, which account for the majority of computations in the CNN, we use hardware accelerators to improve the performance of CNN applications. In using these hardware accelerators, the CNN fetches data from the off-chip DRAM, as the massive computational volume of data makes it difficult to derive performance improvements only from memory inside the hardware accelerator. In other words, data communication between off-chip DRAM and memory inside the accelerator has a significant impact on the performance of CNN applications. In this paper, a simulator for the CNN is developed to analyze the main memory or DRAM access with respect to the size of the on-chip memory or global buffer inside the CNN accelerator. For AlexNet, one of the CNN architectures, when simulated with increasing the size of the global buffer, we found that the global buffer of size larger than 100kB has 0.8x as low a DRAM access count as the global buffer of size smaller than 100kB.

• The KIPS Transactions:PartC
• /
• v.17C no.3
• /
• pp.299-306
• /
• 2010

IPTV and VoIP services are considered as killer applications over Mobile WiMAX network, which provideshigh mobility and data rate. Among those which affect the quality of voice in those services, the jitter buffer or playout buffer can compensate the poor voice quality caused by the packet drop due to frequent route change or differences among routes between service endpoints. In this paper, we analyze the correlation between the sleep interval length and jitter buffer size in order to guarantee a predefined level of voice quality. For this purpose, we present an end-to-end delay model considering additional delay incurred by the WiMAX PSC-II sleep mode and a VoIP service quality requirement based on the delay constraints. Through extensive simulation experiments, we also show that the increase of jitter buffer size may degrade the voice quality since it can introduce additional packet drop in the jitter buffer under WiMAX power saving mode.

• Journal of the Society of Cosmetic Scientists of Korea
• /
• v.37 no.1
• /
• pp.55-60
• /
• 2011

In this study, DOPC liposomes were prepared with distilled water, phosphate buffer and phosphate buffered saline to evaluate the effects of salt on the stability of DOPC liposome. The changes in physical properties (likeparticle size and zeta potential) of liposome were measured after adding the salt. Liposomes were diluted 40 times and 80 times with hydration solvent to confirm the effect of dilution. Consequently, the stability of liposome was maintained up to 40 times dilution with hydration solvent. The liposome that prepared with distilled water was diluted with distilled water, phosphate buffer and phosphate buffered saline, and the liposome that prepared with phosphate buffer was diluted with phosphate buffer and phosphate buffered saline to evaluate the salt-induced changes in particle size and zeta potentia. As results, the particle size increased slightly and zeta potential became closer to 0 when the salt concentration was increased. In conclusion, particle size and zeta potential of liposome could be reasonable factors to evaluate the stability of liposome. In addition, we suggest that salt concentration of hydration solvent has a significant effect on the stability of liposome.

• Journal of Korea Multimedia Society
• /
• v.1 no.1
• /
• pp.67-79
• /
• 1998

In this paper, we present the design and analyze the transmission rate of a control algorithm for the client buffer on the VOD systems. The design is based on the server's data transmission rate and clients' data consumption rate. The proposed algorithm stabilizes the client's buffer by reducing the oscillation phenomena of the buffer. And it uses the BCT (Buffer Check Time) to reduce the scheduling load of the client system. The client's data consumption rate and the buffer size are calculated on the basis of BCT. In case that the predicted buffer size operates in the overrun or starvation the buffer. As the load fluctuation of the network, the transmission rate control policy cannot stabilize the client's buffer alone. The media scaling policy, therefore, also performs its complementary part in stabilizing the client's buffer, and performance evaluation of the proposed algorithm is provided through simulation.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.11
• /
• pp.5343-5364
• /
• 2016

Considering the tradeoff between energy consumption and outage behavior in buffer-aided relay selection, a novel energy-efficient buffer-aided optimal relay selection scheme with power adaptation and Inter-Relay Interference (IRI) cancellation is proposed. In the proposed scheme, energy consumption minimization is the objective with the consideration of relay buffer state, outage probability and relay power control, in order to eliminate IRI. The proposed scheme selects a pair of optimal relays from multiple candidate relays, denoted as optimal receive relay and optimal transmit relay respectively. Source-relay and relay-destination communications can be performed within a time-slot, which performs as Full-Duplex (FD) relaying. Markov chain model is applied to analyze the evolution of relay buffer states. System steady state outage probability and achievable diversity order are derived respectively. In addition, packet transmission delay and power reduction performance are investigated with a specific analysis. Numerical results show that the proposed scheme outperforms other relay selection schemes in terms of outage behavior with power adaptation and IRI cancellation in the same relay number and buffer size scenario. Compared with Buffer State relay selection method, the proposed scheme reduces transmission delay significantly with the same amount of relays. Average transmit power reduction can be implemented to relays with the increasing of relay number and buffer size, which realizes the tradeoff between energy-efficiency, outage behavior and delay performance in green cooperative communications.

• Journal of Navigation and Port Research
• /
• v.33 no.5
• /
• pp.331-337
• /
• 2009

Container crane is main equipment in container terminals and it determines the productivity and the efficiency of container terminals. The typical type of container cranes has the single trolley and one among advanced types of them has the dual trolley. The objective of this paper is to analyze the buffer size of a container crane with the dual trolley in container terminals. We present a simulation model for analysing the buffer space of a container crane with the dual trolley. The buffer space is located between main trolley in sea-side and second trolley in yard-side. We performs various simulation experiments and analyze the buffer size to estimate the required productivity.

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.99-102
• /
• 1998

As multimedia communication services have been widely spreading, the amount of video traffic is rapidly increasing in B-ISDN environment based on the ATM technology. The image quality of MPEG services is very sensitive to the cell losses in ATM network, since each cell contains information needed at decoding process. Since the number of cells in each frame of MPEG is variable, this video smoothing technology need to prepare a buffer for no overflow or underflow at the transmission, requires that some number of cells be taken to the buffer in client before the playback of video. To ensure the high quality image of video, the video smoothing is scheduled by a Group of Picture unit. In this paper, we then apply the theory to reds nightmare encoded in MPEG, and find minimum smoothing buffer size, initial buffer size. It can be used to study the smoothing of stored video.

• Proceedings of the KIEE Conference
• /
• 2004.11c
• /
• pp.629-631
• /
• 2004

The proposed output buffer is presented for driving large-size LCD panels. This output buffer is designed by adding some simple circuitry to the conventional two-stage operational amplifier. The proposed circuit is simulated in a high-voltage 0.35um CMOS process with HSPICE. The simulated result is more improved settling time than that of conventional one.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.8
• /
• pp.2106-2119
• /
• 1997

NDAS(Network Data Access System) is a system that interconnects high speed network(B-ISDN) and Local Area Network(Ethernet, FDDI). Nowadays, the request of multimedia data increases rapidly, so NDAS system must accept and process that request efficiently. Comparing with normal data traffic characteristics, multimedia data traffic characteristics are much complex, so it is necessary to implement the system that can cover such factor. In this paper, we suggest a NDAS hardware system model and also we propose a proper buffer size of NDAS system through the simulation. For measuring the proper buffer size of NDAS, we assumed two different simulation models, one is the case of using only one buffer and the other is the case of using two different buffers for Ethernet and FDDI, respectively. After comparing and analyzing simulation results, we suggest the proper NDAS system buffer size.

• KSBB Journal
• /
• v.17 no.1
• /
• pp.106-109
• /
• 2002

Gel electrophoresis of DNA is a well known technique in molecular biology. This technique is simple, rapid to perform, and capable of adequately separating fragments of DNA. A number of mixtures of DNA fragments ("DNA size markers") are frequently employed in a purpose of extrapolating the sizes or the amount of DNA molecules during gel electrophoresis. DNA size markers are constructed by digesting plasmid DNA, bacteriophage DNA, or recombinant DNA molecules with one or more restriction enzymes. However, liquid suspension containing DNA size marker needs to be kept at a low temperature during storage and shipping. In an attempt to maintain the DNA samples at room temperature for extended period of time, lyophilization of DNA with addition of nuclease inhibitor was studied. Gel loading buffer was also added to the lyophilized DNA to provide additional convenience such that DNA size marker was the "ready-to-use" followed by simply reconstituting with distilled water.