Partitioning circuits for improved testability

Network-on-Chip (NoC) links consume a significant fraction of the total NoC power. We present Selective Packet Interleaving (SPI), a flit transmission scheme that reduces power consumption in NoC links. SPI decreases the number of bit transitions in the links by exploiting the multiplicity of virtual channels in a NoC router. SPI multiplexes flits to the router's output link so as to minimize the number of bit transitions from the previously transmitted flit. Analysis and simulations demonstrate a reduction of up to 55% in the number of bit transitions and up to 40% savings in power consumed on the link. SPI benefits grow with the number of virtual channels. SPI works better for links with a small number of bits in parallel. While SPI compares favorably against bus inversion, combining both schemes helps to further reduce bit transitions.

Based on digitized TV frames a computer simulation study was conducted on a motion -adaptive TV data compression scheme, defined as one in which a 2- dimensional transformapplied to subblocks is followed by the transmission to the receiving site of only changedsubblocks. Performance was measured both by calculation of Normalized Mean Square Errorand by subjective viewing of reconstructed frames. Two systems designs leading to 512x512resolution operation (one based on the Hadamard and one on the Discrete Cosine Transform)were prepared.IntroductionThe motion -adaptive TV data compression described herein is more effective than conven-tional intraframe techniques because it encompasses them and furthermore takes advantageof the correlation between frames by updating only those subblocks which have changed.Following a 2 -D transform performed on small (e.g. 8x8) subblocks, intraframe compressionis achieved by assigning a small number of bits (or none) to each transform coefficient.This bit allocation is made adaptive by assigning fewer bits to those subblocks within aframe which exhibit less busyness or activity. Next, motion -adaptivity, a form of inter -frame compression, is provided by having the amount of change in coefficients between thecurrent subblock and the corresponding subblock of the previous frame determine which trans-form subblocks to transmit. This process has been simulated and an approach to hardwareimplementation derived.BackgroundTwo -dimensional transforms combined with frame to frame motion -adaptivity achieve highcompression ratios because they exploit pixel correlation in all three dimensions.

A steganographic technique inspired by rook is presented in this paper to ensure privacy and secrecy. In this approach, the cover image is partitioned into n × 1 pixel blocks and converted equivalent n × 8 binary bit planes. Then the functional output of each block is calculated on the basis of the number of rook positions, which are attacked by opponent rooks. The rook is a chess piece that moves only forward and backward in a straight line. In binary bit plane, 0 and 1 are considered as a black and white opponent rook, respectively. The secret information is considered as stream of binary bits. The binary bits of secret information are compared with the functional output of the corresponding block. If it is equal to the functional output of the corresponding block, then nothing needs to be done. In case of inequality, the small number of bits needs to be flipped in such a way that the functional output of the corresponding block becomes equal to the corresponding secret binary bits and the distortion of the cover is minimized.

The defining characteristic of a networked control system (NCS) is having a feedback loop that passes through a local area computer network. This paper considers nonlinear systems controlled in this manner, and demonstrates that for sufficiently high transmission rates, the network may be considered transparent. Three methods of scheduling data packets are compared: a static scheduler (token ring), the try-once-discard (maximum-error-first) scheduler with continuous priority levels, and the try-once-discard scheduler with discrete priority levels. The third method is of particular interest when only a small number of bits are available for collision resolution. Asymptotic stability is guaranteed in the first two cases, and ultimate uniform boundedness in the third. In the final section, simulations demonstrate the theoretical results. The contributions of this paper are two-fold: first, it extends the earlier results on NCS to nonlinear systems, and second, it allows for finite word-length message identifiers.

This paper introduces a best approach for Least Significant Bit (LSB) based on image steganography that enhances the existing LSB substitution techniques to improve the security level of hidden information. It is a new approach to substitute LSB of RGB true color image. The new security conception hides secret information within the LSB of image where a secret key encrypts the hidden information to protect it from unauthorized users. In general, in LSB methods, hidden information is stored into a specific position of LSB of image. For this reason, knowing the retrieval methods, anyone can extract the hidden information. In our paper, hidden information is stored into different position of LSB of image depending on the secret key. As a result, it is difficult to extract the hidden information knowing the retrieval methods. We have used the Peak Signal-to-Noise Ratio (PSNR) to measure the quality of the stego images. The value of PSNR gives better result because our proposed method changes very small number of bits of the image. The obtained results show that the proposed method results in LSB based image steganography using secret key which provides good security issue and PSNR value than general LSB based image steganography methods.

Halftoning is a process that deliberately injects noise into the original image in order to obtain visually pleasing output images with a smaller number of bits per pixel for displaying or printing purposes. In this paper, a novel inverse halftoning method is proposed to restore a continuous tone image from the given halftone image. A set theoretic formulation is used where three sets are defined using the prior information about the problem. A new space domain projection is introduced assuming the halftoning is performed with error diffusion, and the error diffusion filter kernel is known. The space domain, frequency domain, and space-scale domain projections are used alternately to obtain a feasible solution for the inverse halftoning problem which does not have a unique solution.

This paper presents an efficient method to block unknowns for temporal compactors. The proposed blocking logic can reduce data volume required to control the blocking logic and increase the number of scan cells that are observed by temporal compactors. Control patterns, which specify values required at the control signals of the blocking logic, are compressed by linear feedback shift register reseeding. In this paper, the blocking logic gates for some scan chains that do not capture unknowns are bypassed. Since all scan cells in these scan chains can be observed without specifying the corresponding bits in control patterns, more scan cells are observed while a smaller number of bits are required to be specified. The seed size is further reduced by reducing the numbers of specified bits in the densely specified control patterns. The proposed method can always achieve the same fault coverage that can be achieved by directly observing scan chains without any output compaction. Experiments with large industrial designs clearly demonstrate that the proposed method is scalable to large circuits. Hardware overhead for the proposed unknown blocking scheme is very low.

We propose an AVF-driven parity selection method for protecting modern microprocessor in-core memory arrays against MBUs. As MBUs constitute more than 50% of the upsets in latest technologies, error correcting codes or physical interleaving are typically employed to effectively protect out-of-core memory structures, such as caches. However, such methods are not applicable to high-performance in-core arrays, due to computational complexity, high delay and area overhead. To this end, we revisit parity as an effective mechanism to detect errors and we resort to pipeline flushing and checkpointing for correction. We demonstrate that optimal parity tree construction for MBU detection is a computationally complex problem, which we then formulate as an integer-linear-program (ILP). Experimental results on Alpha 21264 and Intel P6 in-core memory arrays demonstrate that optimal parity tree selection can achieve great vulnerability reduction, even when a small number of bits are added to the parity trees, compared to simple heuristics. Furthermore, the ILP formulation allows us to find better solutions by effectively exploring the solution space in the presence of multiple parity trees; results show that the presence of 2 parity trees offers a vulnerability reduction of more than 50% over a single parity tree.

We consider decentralized estimation of a noise-corrupted deterministic parameter by a bandwidth-constrained sensor network with a fusion center. The sensor noises are assumed to be additive, zero mean, spatially uncorrelated, but otherwise unknown and possibly different across sensors due to varying sensor quality and inhomogeneous sensing environment. The classical best linear unbiased estimator (BLUE) linearly combines the real-valued sensor observations to minimize the mean square error (MSE). Unfortunately, such a scheme cannot be implemented in a practical bandwidth-constrained sensor network due to its requirement to transmit real-valued messages. In this paper, we construct a decentralized estimation scheme (DES) where each sensor compresses its observation to a small number of bits with length proportional to the logarithm of its local signal-to-noise ratio (SNR). The resulting compressed bits from different sensors are then collected and combined by the fusion center to estimate the unknown parameter. The proposed DES is universal in the sense that each sensor compression scheme requires only the knowledge of local SNR, rather than the noise probability distribution functions (pdf), while the final fusion step is also independent of the local noise pdfs. We show that the MSE of the proposed DES is within a constant factor of 25/8 of that achieved by the classical centralized BLUE estimator.

Adversarial attacks on Neural Network weights, such as the progressive bit-flip attack (PBFA), can cause a catastrophic degradation in accuracy by flipping a very small number of bits. Furthermore, PBFA can be conducted at run time on the weights stored in DRAM main memory. In this work, we propose RADAR, a Run-time adversarial weight Attack Detection and Accuracy Recovery scheme to protect DNN weights against PBFA. We organize weights that are interspersed in a layer into groups and employ a checksum-based algorithm on weights to derive a 2-bit signature for each group. At run time, the 2-bit signature is computed and compared with the securely stored golden signature to detect the bit-flip attacks in a group. After successful detection, we zero out all the weights in a group to mitigate the accuracy drop caused by malicious bit-flips. The proposed scheme is embedded in the inference computation stage. For the ResNet-18 ImageNet model, our method can detect 9.6 bit-flips out of 10 on average. For this model, the proposed accuracy recovery scheme can restore the accuracy from below 1% caused by 10 bit flips to above 69%. The proposed method has extremely low time and storage overhead. System-level simulation on gem5 shows that RADAR only adds <1% to the inference time, making this scheme highly suitable for run-time attack detection and mitigation.

We consider the problem of constructing randomness extractors that are locally computable; that is, read only a small number of bits from their input. As recently shown by Lu locally computable extractors directly yield secure private-key cryptosystems in Maurer’s bounded-storage model. We suggest a general “sample-then-extract” approach to constructing locally computable extractors: use essentially any randomness-efficient sampler to select bits from the input and then apply any extractor to the selected bits. Plugging in known sampler and extractor constructions, we obtain locally computable extractors, and hence cryptosystems in the bounded-storage model, whose parameters improve upon previous constructions. We also provide lower bounds showing that the parameters we achieve are nearly optimal. The correctness of the sample-then-extract approach follows from a fundamental lemma of Nisan and Zuckerman, which states that sampling bits from a weak random source roughly preserves the min-entropy rate. We also present a refinement of this lemma, showing that the min-entropy rate is preserved up to an arbitrarily small additive loss, whereas the original lemma loses a logarithmic factor.

Many applications, including program optimizations, debugging tools, and event loggers, rely on calling context to gain additional insight about how a program behaves during execution. One common strategy for determining calling contexts is to use compiler instrumentation at each function call site and return sites to encode the call paths and store them in a designated area of memory. While recent works have shown that this approach can generate precise calling context encodings with low overhead, the encodings can grow to hundreds or even thousands of bytes to encode a long call path, for some applications. Such lengthy encodings increase the costs associated with storing, detecting, and decoding call path contexts, and can limit the effectiveness of this approach for many usage scenarios.

One method of synthesizing a digital filter involves the simulation of an LC analog filter. This method retains the low element sensitivity that is characteristic of an LC filter and is generally attractive in that it requires only a small number of bits in the multiplier that is used to produce the filter coefficients. The web digital filter is a typical application of this design method and is known to possess low‐element sensitivity, but it involves a somewhat complex configuration and the transfer functions which can be realized are therefore limited. This paper discusses a method of designing digital filters by simulation of the node equation of LC filters. Using this method, the transfer function of an analog filter which can be implemented by means of LCR elements, including transformers, can always be realized. With this method, we first transform the LC filter into an equivalent circuit and determine the applicable node equation. Then bilinear transformation is used to transform the s‐variable into the z‐variable and, to prevent the occurrence of a delay‐free loop within the circuit, after rearranging the circuit equation it is realized as a digital filter. In addition, a study is made of methods reducing the number of multipliers without degrading the coefficient sensitivity. Finally, the effectiveness of this design method is verified by means of computer simulation.

We consider the problem of encoding a finite set of vectors into a small number of bits while approximately retaining information on the angular distances between the vectors. By deriving improved variance bounds related to binary Gaussian circulant embeddings, we largely fix a gap in the proof of the best known fast binary embedding method. Our bounds also show that well-spreadness assumptions on the data vectors, which were needed in earlier work on variance bounds, are unnecessary. In addition, we propose a new binary embedding with a faster running time on sparse data.

To enhance the payload of secret data a scheme on the basis of least significant bit (LSB) substitution cascaded with pixel value differencing (PVD) has been proposed in this paper. Firstly secret bits are embedded using PVD method where smooth areas obtain small number of bits and higher bits in the edged area. Then bits are also embedded in the same pixel block by applying LSB replacement method as a cascading manner. The technique ensures more hiding capacity along with security as two methods are applied as cascaded manner on the same pixel block. Experimental result shows that proposed method can hide nearly double amount of secret data, compared to PVD method with good visual quality of stego-images.