Integer division is different from floating-point division in that (1) the execution time of an integer division is highly dependent on the leading 1 locations of operands, (2) x and −x have different magnitude parts if x is a two’s complement integer, and (3) rounding is not necessary. In this paper, we apply the interval-analysis-based division algorithm proposed in "Dual-Purpose Hardware Algorithms and Architecture – Part 1: Floating-Point Division" [1] to offline and online integer division. We implement four online integer dividers using the algorithm, compare them with other dividers, and present detailed simulation results with in-depth analysis of the dividers. We find that the online dividers outperform the offline dividers when the waiting time for online operands goes up and the number of quotient bits to obtain goes down.

Division is a time-consuming, but frequently-used arithmetic operation, so an enormous amount of effort has been made to improve the performance of dividers. Most of the division algorithms in the literature are offline algorithms that minimize the execution time of a single division, whereas some others are online algorithms that maximize the throughput (# divisions executed per time). In this paper, we propose an interval-analysis-based normal-binary floating-point division algorithm that can be used for both offline and online division. We implement two offline and four online dividers using the algorithm and compare them with recently-proposed offline and online dividers. The simulation results show that the offline versions are the best for a Binary64 offline division, whereas the online versions are the best for a Binary64 online division.

In this research, we derived indicator spectral bands and classification models for detecting diesel or gasoline pollution in soil using a near-and shortwave-infrared (NIR-SWIR) hyperspectral camera under a close range and laboratory condition. The soils samples were collected from temperate climate soil with spectral characteristics manifested by secondary minerals. The hyperspectral images show that the diesel and gasoline polluted soil samples have distinctive spectral differences from clean soil. Different from moisture soil, the spectral absorption features of petroleum hydrocarbons (PHCs) are preserved with an increase in gravimetric content. The more PHCs contents, the stronger the depths at the spectral absorption features. In diesel polluted soils, the absorption features were observed in various content levels. However, we found a detection limit for gasoline content in soil, because the absorption features by PHCs disappeared at 8 wt.%. To derive the indicator bands, the images were classified by the random forest algorithm (RF) with an accuracy and kappa coefficient of 94.3% and 0.92 using three groups of bands corresponding to ferric ion, C-H stretch/bending, and BTEX C-H absorptions. The detection models derived from a logistic regression achieved an overall accuracy of 91.82%. The field test of the models on unprocessed soils achieved an accuracy of 83.36%. Because of their simple forms, the logistic detection models can be transferred to remote sensing applications of soil PHCs pollution under a close-range condition such as drone-based projects.

This study developed case-adaptive spectral indices for detecting heavy metal pollution in skarn soil, hydrothermal soil, and acid mine drainage (AMD) soil, and a universal index for general cases by merging data from all cases. Sequential analyses were conducted including heavy metal concentration, mineral identification, grain size, and spectral characteristics. A factor analysis was used to combine multiple heavy metal elements into representative factors and used for establishing the spectral indices using the stepwise multiple linear regression (SMLR), which is compared with random forest regression (RFR). The spectral indices of skarn soil predict Zn, As, and Pb pollution from moderate to ultrahigh levels using absorption bands of skarn and supergene minerals. The spectral index for hydrothermal soil predicts pollution of Cr and Ni from very low to very high using absorption bands of hydrothermal alteration or supergene minerals. The index for the AMD soil predicts Al, Fe, and As pollution from very low to very high levels by employing the absorption features of iron hydroxide. The universal index predicts Fe, Zn, and Pb pollution levels from very low to ultrahigh by utilizing the clay and iron oxide absorptions, commonly observed in general soils. All models showed statistical significance with R2= 0.66-0.92 for SMLR and R2= 0.66-0.95 for RFR. Because our models were derived from large number of samples (600s) for detecting the pollution levels of multiple heavy metal elements for the first time, the models could provide a decision-making tool for soil survey in heavy metal pollution.

In this paper, we propose a pipelined multiplier architecture that can resolve data dependencies. The proposed architecture generates partial results in the pipeline stages of the multiplier and forwards the partial results back to the pipeline stages through so-called fast-forwarding paths, thereby enabling an execution of dependent multiplications with a minimum delay penalty. We apply the architecture to a normal binary multiplier (NBBE-2) and two redundant binary multipliers (RBBE-4 and CRBBE-4) and compare the execution time, clock period, area, and power consumption of the multipliers. The simulation results show that the proposed architecture achieves up to 30% execution time reduction.

In this paper, we propose two dependency-resolving intra-unit pipeline architectures to design high-throughput multipliers. Simulation results show that the proposed multipliers achieve approximately 2.4× to 4.3× execution time reduction at a cost of 4.4% area and 3.7% power overheads for highly-dependent multiplications.