Retention Time Variation Research Articles

AVALIAÇÃO DO BALANÇO HÍDRICO E DA EFICIÊNCIA DE UM ALAGADO CONSTRUÍDO COMO ALTERNATIVA DE TRATAMENTO PARA EFLUENTE DOMÉSTICO

This study aimed to evaluate how pluviometric precipitation and evapotranspiration interfere with the efficiency of a horizontal subsurface flow constructed wetland used for domestic wastewater treatment. The study was developed in two phases of six months, the first evaluated without the influence of pluviometric precipitation (SIPP) and the second exposed to pluviometric precipitation (IPP). The constructed wetland is subsurface, with horizontal flow, it is vegetated with macrophytes and occupied with stone chips (“rachao”), gravel and sand. The equipments used for the measurement of water balance were an analogical rain gauge and two water meters, while to evaluate the treatment efficacy, the analysis of OBD (Oxygen Biochemical Demand), NTK (Nitrogen Total Kjeldal) and Pt (Phosphorus total) were realized in laboratory. The results showed that in the phase IPP the water balance faced significant changes due to the pluviometric volumes and evapotranspiration. The average removal efficiencies in the phases SIPP and IPP were 67,3% e 66,8% for DBO, 40% and 21% for NTK and 34,8% and 17,6% for Pt, respectively. The study showed that pluviometric precipitation and evapotranspiration rates influence the water balance of a constructed wetland, providing variations in hydraulic retention time of the efluente, what, as a consequence, interferes with the efficiency of removal of polluting loads.

Trade-off between carbon emission and effluent quality of activated sludge processes under seasonal variations of wastewater temperature and mean cell retention time

Over the seasonal cycles, the mean cell retention time (MCRT) of the activated sludge process is varied to compensate the wastewater temperature variations. The effects of these variations on the carbon footprint (CFP) and effluent quality index (EQI) of a conventional activated sludge (CAS) process and a nitrification/denitrification (NDN) process were quantified. The carbon emission included both biogenic and non-biogenic carbon. Carbon emissions of wasted biosolids management were also addressed. Our results confirmed that the effluent quality indicated by EQI was not necessarily improved by increasing MCRT. Higher MCRT increased the carbon emission and reduced excess sludge production, which decreased the potential for biogas energy recovery. The NDN process was preferable to the CAS process from the perspective of effluent quality. This consideration extended to the whole plant CFP if the N2O emitted during NDN was limited ([N2O]<1% [NH4+]removed) as the carbon emission per unit effluent quality achieved by NDN process is less than that of the CAS process. By putting forward carbon emission intensity (γ) derived from CFP and EQI, our work provides a quantitative tool for decision makers evaluating process alternatives when there is a trade-off between carbon emission and effluent quality.

DRAM Refresh Mechanisms, Penalties, and Trade-Offs

Ever-growing application data footprints demand faster main memory with larger capacity. DRAM has been the technology choice for main memory due to its low latency and high density. However, DRAM cells must be refreshed periodically to preserve their content. Refresh operations negatively affect performance and power. Traditionally, the performance and power overhead of refresh have been insignificant. But as the size and speed of DRAM chips continue to increase, refresh becomes a dominating factor of DRAM performance and power dissipation. In this paper, we conduct a comprehensive study of the issues related to refresh operations in modern DRAMs. Specifically, we describe the difference in refresh operations between modern synchronous DRAM and traditional asynchronous DRAM; the refresh modes and timings; and variations in data retention time. Moreover, we quantify refresh penalties versus device speed, size, and total memory capacity. We also categorize refresh mechanisms based on command granularity, and summarize refresh techniques proposed in research papers. Finally, based on our experiments and observations, we propose guidelines for mitigating DRAM refresh penalties.

Unidades combinadas RAFA-SAC para tratamento de água residuária de suinocultura – parte I carga orgânica removida

RESUMO Na pecuária brasileira, tem-se destacado a suinocultura, cuja produção de carne foi apontada como a terceira maior do mundo, e como contrapartida os sistemas produtivos são, potencialmente, poluidores do ambiente. Os sistemas anaeróbios têm sido utilizados para tratamento de água residuária suinícola, necessitando de complementação para diminuir as cargas de nutrientes e reúso agrícola. Neste trabalho, avaliou-se o efeito combinado de duas unidades de tratamento na remoção de matéria orgânica, na forma de demanda química de oxigênio (DQO), sendo um reator anaeróbio de manta de lodo e fluxo ascendente tipo RAFA com volume útil de 96 L, seguido de um sistema alagado construído (SAC) com capacidade para 237 L. O experimento foi conduzido em três fases, variando-se o tempo de detenção hidráulica (TDH) no reator anaeróbio, de 59 h, 19,5 h e 5 h, e no SAC, 146 h, 48 h e 13 h, respectivamente, nas fases 1, 2 e 3. A carga orgânica volumétrica (COV) aplicada foi de 1,2; 1,3 e 13,0 kg m-3d-1 de DQO, e as taxas de aplicação superficial (TAS) no SAC foram de 850; 656 e 6335 kg ha-1 d-1 de DQO. As unidades combinadas removeram 0,07; 0,07 e 0,96 kg d-1 de DQO total e 0,03; 0,07 e 0,14 kg d-1 de DQO solúvel. As características apresentadas pelo efluente do sistema, nas diferentes fases, não alcançaram os padrões ambientais para lançamento em cursos de água, dentre as variáveis avaliadas.

Beeromics: from quality control to identification of differentially expressed compounds in beer

A rigorous quality control platform was developed to monitor short- and long-term retention time reproducibility, instrument response and mass accuracy for untargeted differential analysis of beer. The optimized workflow utilized a QC beer and external calibration standards to yield LC retention time variation <3 s, average interday response RSDs <15 % and average mass errors <10 ppm on a quadrupole time of flight mass spectrometer. QC statistics guided data processing, including molecule feature (MF) extraction, MF alignment across samples and recursion, prior to differential analysis of single-hop India Pale Ales produced in two different production years (2010 vs. 2011). Differential analysis and significance testing revealed that the year of production (vs. hop used) accounted for the greatest compositional difference between samples. Negative and positive ion ESI data collectively revealed a minimum 3× fold change (p < 0.05) for ~4000 MFs. Since identification of ~4000 MFs was untenable, focus turned to the identification of ~150 high-abundance MFs. Targeted MS/MS, Metlin database searching and the use of ChemSpider with automated MS/MS spectral prediction and correlation, identified candidate compounds that were confirmed by authentic standards. Most interestingly, the 2010 beers, which were stored longer at room temperature prior to analysis, provided additional evidence for the use of purines, specifically 5-methylthioadenosine with methionine, as nonvolatile molecular indicators of beer oxidation during storage. While this untargeted metabolomics platform was developed with beer in mind, the workflow described here could be applied to any food metabolomics application.

Tracking and quantification of nitrifying bacteria in biofilm and mixed liquor of a partial nitrification MBBR pilot plant using fluorescence in situ hybridization

A moving bead biofilm reactor (MBBR) pilot plant was implemented as a partial nitrification process for pre-treatment of ammonium-rich liquors (676±195mgL−1), and studied for 479days under variations in hydraulic retention time. The main purpose of this work, was the study of dynamics abundance of total bacteria and single-cells nitrifying bacteria belonging to ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB) in biofilms and mixed liquor of the plant. The microbial monitoring was successfully achieved using fluorescence in situ hybridization combined with flocs disaggregation protocol as a useful microbial monitoring tool. A partial nitrification process with a N-NH4+ removal rate of about 38.6±14.8% was successfully achieved at 211days after start-up, with a clear dominance of AOB, which accounted for 11.3±17.0% of total bacterial cells compared with only 2.1±4.0% of NOB. The effluent obtained was subsequently supplied to an Anammox reactor for complete ammonium treatment.

Effect of Operating Parameters on Process Stability of Continuous Biohydrogen Production from Palm Oil Mill Effluent under Thermophilic Condition

Effect of organic loading rate (OLR), hydraulic retention time (HRT) and temperature variation on process stability of fermentative hydrogen production from palm oil mill effluent (POME) under thermophilic condition was investigated. High OLR (>70 gCOD/l/d) and low HRT (4 d) has significant affected on hydrogen productivity, process stability and microbial community. Hydrogen production at OLR of 55, 60, 65 and 70 gCOD/l/d was 2.4, 3.4, 4.9 and 3.7 L H2/L-POME, respectively. The decreasing in pH and alkalinity and increasing in total VFA could cause the instability of process. The microbial community was changed at OLR >70 gCOD/l/d, HRT 4 d and stop feeding. The dominant microbial community changed from Thermoanaerobacterim sp. to Clostridium sp., Thermoanaerobacterim sp. and Bacillus sp. Hydrogen production at HRT of 2, 4, stop feeding and 2 d was 2.7, 0.8, 0, and 2.6 L H2/L-POME, respectively. Hydrogen production, process stability and microbial community was not affected by variation of temperature. Hydrogen production before and after temperature variation was 2.5 and 2.45 L H2/L-POME.

Bioaugmentation of anaerobic sludge digestion with iron-reducing bacteria: process and microbial responses to variations in hydraulic retention time.

Although anaerobic digestion (AD) is a widely used option to manage waste activated sludge (WAS), there are some drawbacks related to its slow reaction rate and low energy productivity. This study examined an anaerobic WAS digester, augmented with an iron-reducing microbial consortium, relative to changes in microbial community structure and process performance at decreasing hydraulic retention times (HRTs) of 20 to 10 days. The enhanced methanation performance (approximately 40 % increase in methane yield) by the bioaugmentation was sustained until the HRT was decreased to 12.5 days, under Fe(3+)-rich conditions (ferric oxyhydroxide, 20 mM Fe). Enhanced iron-reducing activity was evidenced by the increased Fe(2+) to total Fe ratio maintained above 50 % during the stable operational phases. A further decrease in HRT to 10 days resulted in a significant performance deterioration, along with a drop in the Fe(2+) to total Fe ratio to <35 %, after four turnovers of operation. Prevailing existence of putative iron-reducing bacteria (IRBs) was identified by denaturing gradient gel electrophoresis (DGGE), with Spirochaetaceae- and Thauera-related organisms being dominant members, and clear dominance shifts among them with respect to decrease in HRT were observed. Lowering HRT led to evident shifts in bacterial community structure likely associated with washout of IRBs, leading to decreases in iron respiration activity and AD performance at a lower HRT. The bacterial community structure shifted dynamically over phases, and the community transitions correlated well with the changes in process performance. Overall, the combined biostimulation and bioaugmentation investigated in this study proved effective for enhanced methane recovery from anaerobic WAS digestion, which suggests an interesting potential for high-rate AD.

Transport of Escherichia coli through a Thick Vadose Zone

Livestock manure applications on fields can be a source of contamination in water resources, including groundwater. Although fecal indicators like have often been detected in tile drainage systems, few studies have monitored groundwater at depth after manure treatments, especially at sites with a deep, heterogeneous vadose zone. Our hypothesis was that microbial transport through a thick vadose zone would be limited or nonexistent due to attenuation processes, subsurface thickness, and heterogeneity. This study tested this hypothesis by monitoring concentrations beneath a 12-m-thick vadose zone of coarse, heterogeneous glacial sediments after surface application of liquid swine manure. was detected on all 23 sample dates over the 5-mo period (4 Apr. 2012-13 Aug. 2012), with particularly elevated concentrations 1 wk after application and lasting for 5 wk. Variable low-level concentrations before and after the elevated period suggest remobilization and delayed transport of microorganisms to the water table without additional loadings within the flow field. These findings suggest preferential flow pathways allowing deep infiltration of manure bacteria as well as a continued source of bacteria, with variable retention and travel times, over several months. Preferential flow pathways at this site include soil macropores, depression focused infiltration, and pathways related to subsurface heterogeneity and/or fracture flow through finer-grained diamict beds. Further research is needed to confirm the relative contribution of sources, constrain travel times, and define specific transport pathways.

Optimization of Fat-Soluble Vitamins Separations by Reversed-Phase Liquid Chromatography with the Use of Aliphatic Alcohols as Mobile Phase Additives

In an effort to reduce the total amount of organic solvent required for the elution of fat-soluble vitamins (FSVs) in reversed-phase liquid chromatography separation, a systematic approach was developed using a mixture of methanol, water, and an alkanol as the mobile phase. In the optimization procedure, a five-parameter retention model was developed and the statistical evaluation of model, revealed that the variations of retention times of each vitamins caused by addition of alkanols to the mobile phase were predicted with an average relative deviation of 1.83%. For this model, all the and values for all vitamins were higher than 0.9839 and 0.9798, respectively, and lack of fit (LOF) values were not statistically significant. A mobile phase containing methanol/octanol/water (86:9:5) was proposed as the optimum mobile phase composition using grid search method and a good agreement was observed between predicted and experimental response values (relative error = −0.734%). Using this optimum condition, the simultaneous separation of six FSVs was achieved in 9.94 min. The results showed an improvement in the separation of FSVs and in reduction of the total amount of organic solvent used for separation with addition of low amount of alkanols as additive to the mobile phase.

SECRET

DRAMs are used as the main memory in most computing systems today. Studies show that DRAMs contribute to a significant part of overall system power consumption. One of the main challenges in low-power DRAM design is the inevitable refresh process. Due to process variation, memory cells exhibit retention time variations. Current DRAMs use a single refresh period determined by the cell with the largest leakage. Since prolonging refresh intervals introduces retention errors, a set of previous works adopt conventional error-correcting code (ECC) to correct retention errors. However, these approaches introduce significant area and energy overheads. In this article, we propose a novel error correction framework for retention errors in DRAMs, called SECRET (selective error correction for refresh energy reduction). The key observations we make are that retention errors are hard errors rather than soft errors, and only few DRAM cells have large leakage. Therefore, instead of equipping error correction capability for all memory cells as existing ECC schemes, we only allocate error correction information to leaky cells under a refresh interval. Our SECRET framework contains two parts: an offline phase to identify memory cells with retention errors given a target error rate and a low-overhead error correction mechanism. The experimental results show that among all test cases performed, the proposed SECRET framework can reduce refresh power by 87.2% and overall DRAM power up to 18.57% with negligible area and performance overheads.

Analysis of Microbial Communities in Biofilms from CSTR-Type Hollow Fiber Membrane Biofilm Reactors for Autotrophic Nitrification and Hydrogenotrophic Denitrification.

Two hollow fiber membrane biofilm reactors (HF-MBfRs) were operated for autotrophic nitrification and hydrogenotrophic denitrification for over 300 days. Oxygen and hydrogen were supplied through the hollow fiber membrane for nitrification and denitrification, respectively. During the period, the nitrogen was removed with the efficiency of 82-97% for ammonium and 87-97% for nitrate and with the nitrogen removal load of 0.09-0.26 kg NH4(+)-N/m(3)/d and 0.10-0.21 kg NO3(-)-N/m(3)/d, depending on hydraulic retention time variation by the two HF-MBfRs for autotrophic nitrification and hydrogenotrophic denitrification, respectively. Biofilms were collected from diverse topological positions in the reactors, each at different nitrogen loading rates, and the microbial communities were analyzed with partial 16S rRNA gene sequences in denaturing gradient gel electrophoresis (DGGE). Detected DGGE band sequences in the reactors were correlated with nitrification or denitrification. The profile of the DGGE bands depended on the NH4(+) or NO3(-) loading rate, but it was hard to find a major strain affecting the nitrogen removal efficiency. Nitrospira-related phylum was detected in all biofilm samples from the nitrification reactors. Paracoccus sp. and Aquaspirillum sp., which are an autohydrogenotrophic bacterium and an oligotrophic denitrifier, respectively, were observed in the denitrification reactors. The distribution of microbial communities was relatively stable at different nitrogen loading rates, and DGGE analysis based on 16S rRNA (341f /534r) could successfully detect nitrate-oxidizing and hydrogen-oxidizing bacteria but not ammonium-oxidizing bacteria in the HF-MBfRs.

Temporal dynamics of seed excretion by wild ungulates: implications for plant dispersal.

Dispersal is a key process in metapopulation dynamics as it conditions species’ spatial responses to gradients of abiotic and biotic conditions and triggers individual and gene flows. In the numerous plants that are dispersed through seed consumption by herbivores (endozoochory), the distance and effectiveness of dispersal is determined by the combined effects of seed retention time in the vector’s digestive system, the spatial extent of its movements, and the ability of the seeds to germinate once released. Estimating these three parameters from experimental data is therefore crucial to calibrate mechanistic metacommunity models of plant–herbivore interactions. In this study, we jointly estimated the retention time and germination probability of six herbaceous plants transported by roe deer (Capreolus capreolus), red deer (Cervus elaphus), and wild boar (Sus scrofa) through feeding experiments and a Bayesian dynamic model. Retention time was longer in the nonruminant wild boar (>36 h) than in the two ruminant species (roe deer: 18–36 h, red deer: 3–36 h). In the two ruminants, but not in wild boar, small and round seeds were excreted faster than large ones. Low germination probabilities of the excreted seeds reflected the high cost imposed by endozoochory on plant survival. Trait-mediated variations in retention time and germination probability among animal and plant species may impact plant dispersal distances and interact with biotic and abiotic conditions at the release site to shape the spatial patterns of dispersed plant species.

Targeted Metabolomic Analysis of Head and Neck Cancer Cells Using High Performance Ion Chromatography Coupled with a Q Exactive HF Mass Spectrometer.

In this study, we have demonstrated a targeted metabolomics method for analysis of cancer cells, based on high-performance ion chromatography (IC) separation, Q Exactive HF MS for high-resolution and accurate-mass (HR/AM) measurement and the use of stable isotope-labeled internal standards for absolute quantitation. Our method offers great technical advantages for metabolite analysis, including exquisite sensitivity, high speed and reproducibility, and wide dynamic range. The high-performance IC provided fast separation of cellular metabolites within 20 min and excellent resolving power for polar molecules including many isobaric metabolites. The IC/Q Exactive HF MS achieved wide dynamic ranges of 5 orders of magnitude for six targeted metabolites, pyruvate, succinic acid, malic acid, citric acid, fumaric acid, and alpha-ketoglutaric acid, with R(2) ≈ 0.99. Using this platform, metabolites can be simultaneously quantified from low fmol/μL to nmol/μL levels in cellular samples. The high flow rate IC at 380 μL/min has shown excellent reproducibility for a large set of samples (150 injections), with minimal variations of retention time (SD < ± 0.03 min). In addition, the IC-MS-based approach acquires targeted and global metabolomic data in a same analytical run, and the use of stable isotope-labeled standards facilitates accurate quantitation of targeted metabolites in large-scale metabolomics analysis. This metabolomics approach has been successfully applied to analysis of targeted metabolites in head and neck cancer cells as well as cancer stem-like cells (CSCs), and the findings indicate that the metabolic phenotypes may be distinct between high and low invasive head and neck cancer cells and between CSCs and non-SCCs.

Semi-continuous mesophilic anaerobic digester performance under variations in solids retention time and feeding frequency

The goal of this research was to understand the effect of solids retention time (SRT) and feeding frequency on the performance of anaerobic digesters used to recover bioenergy from swine waste. Semi-continuous mesophilic anaerobic digesters were operated at varying SRTs and feeding frequencies. Performance metrics included biogas and methane production rates, biomass robustness and functionality and removals of volatile solids, soluble chemical oxygen demand, the fecal-indicator bacteria Escherichia coli, and the human pathogen Salmonella. Biochemical methane formation potential and specific methanogenic activity assays were used to demonstrate biomass robustness and functionality. Results indicated that anaerobic digesters fed weekly had higher average methane yields (0.20 vs. 0.18m3CH4/kg-VSadded), specific methanogenic activities (40 vs. 35ml/day), and fecal indicator bacteria destruction (99.9% vs. 99.4%) than those fed every-other day. Salmonella, soluble COD, and VS destruction did not change with varied feeding frequency; however, higher removals were observed with longer SRT.

The effects of river inflow and retention time on the spatial heterogeneity of chlorophyll and water–air CO&amp;lt;sub&amp;gt;2&amp;lt;/sub&amp;gt; fluxes in a tropical hydropower reservoir

Abstract. Abundant research has been devoted to understanding the complexity of the biogeochemical and physical processes that are responsible for greenhouse gas (GHG) emissions from hydropower reservoirs. These systems may have spatially complex and heterogeneous GHG emissions due to flooded biomass, river inflows, primary production and dam operation. In this study, we investigated the relationships between the water–air CO2 fluxes and the phytoplanktonic biomass in the Funil Reservoir, which is an old, stratified tropical reservoir that exhibits intense phytoplankton blooms and a low partial pressure of CO2 (pCO2). Our results indicated that the seasonal and spatial variability of chlorophyll concentrations (Chl) and pCO2 in the Funil Reservoir are related more to changes in the river inflow over the year than to environmental factors such as air temperature and solar radiation. Field data and hydro\\-dynamic simulations revealed that river inflow contributes to increased heterogeneity during the dry season due to variations in the reservoir retention time and river temperature. Contradictory conclusions could be drawn if only temporal data collected near the dam were considered without spatial data to represent CO2 fluxes throughout the reservoir. During periods of high retention, the average CO2 fluxes were 10.3 mmol m−2 d−1 based on temporal data near the dam versus −7.2 mmol m−2 d−1 with spatial data from along the reservoir surface. In this case, the use of solely temporal data to calculate CO2 fluxes results in the reservoir acting as a CO2 source rather than a sink. This finding suggests that the lack of spatial data in reservoir C budget calculations can affect regional and global estimates. Our results support the idea that the Funil Reservoir is a dynamic system where the hydrodynamics represented by changes in the river inflow and retention time are potentially a more important force driving both the Chl and pCO2 spatial variability than the in-system ecological factors.

Sludge minimization in municipal wastewater treatment by polyhydroxyalkanoate (PHA) production.

An innovative approach has been recently proposed in order to link polyhydroxyalkanoates (PHA) production with sludge minimization in municipal wastewater treatment, where (1) a sequencing batch reactor (SBR) is used for the simultaneous municipal wastewater treatment and the selection/enrichment of biomass with storage ability and (2) the acidogenic fermentation of the primary sludge is used to produce a stream rich in volatile fatty acids (VFAs) as the carbon source for the following PHA accumulation stage. The reliability of the proposed process has been evaluated at lab scale by using substrate synthetic mixtures for both stages, simulating a low-strength municipal wastewater and the effluent from primary sludge fermentation, respectively. Six SBR runs were performed under the same operating conditions, each time starting from a new activated sludge inoculum. In every SBR run, despite the low VFA content (10% chemical oxygen demand, COD basis) of the substrate synthetic mixture, a stable feast-famine regime was established, ensuring the necessary selection/enrichment of the sludge and soluble COD removal to 89%. A good process reproducibility was observed, as also confirmed by denaturing gradient gel electrophoresis (DGGE) analysis of the microbial community, which showed that a high similarity after SBR steady-state had been reached. The main variation factors of the storage properties among different runs were uncontrolled changes of settling properties which in turn caused variations of both sludge retention time and specific organic loading rate. In the following accumulation batch tests, the selected/enriched consortium was able to accumulate PHA with good rate (63 mg CODPHA g CODXa(-1) h(-1)) and yield (0.23 CODPHA CODΔS(-1)) in spite that the feeding solution was different from the acclimation one. Even though the PHA production performance still requires optimization, the proposed process has a good potential especially if coupled to minimization of both primary sludge (by its use as the VFA source for the PHA accumulation, via previous fermentation) and excess secondary sludge (by its use as the biomass source for the PHA accumulation).

Phytoplankton chlorophyllain Lake Poyang and its tributaries during dry, mid-dry and wet seasons: a 4-year study

Variations in phytoplankton chlorophyll a were studied in Lake Poyang, and these variations were compared with those observed in its tributaries. Samples were collected from January 2009 to January 2013 during the dry, mid-dry and wet seasons. The study showed that chlorophyll a concentrations were significantly higher in the wet season than in the dry and mid-dry seasons (p between the two ecosystems during the wet season but not during the dry and mid-dry seasons. During the wet season, the underwater light conditions were significantly elevated and water retention time increased substantially (25.5 d). In combination, these factors contributed to a significantly higher chlorophyll a concentration during the wet season than that during the dry and mid-dry seasons in the lake; the effect of nutrients on the phytoplankton was not obvious. The large variations in water retention time during the three seasons indicated that Lake Poyang has the combined characteristics of a river and a lake. The short retention times during the dry and mid-dry seasons produced no difference in chlorophyll a between the lake and its tributaries, whereas the long retention time during the wet season facilitated phytoplankton growth in the lake.

A new extension method of retention time for memory cell on dynamic random access memory

Demands have been placed on dynamic random access memory (DRAM) to not only increase memory capacity and data transfer speed but also to reduce operating and standby currents. When a system uses DRAM, the restricted data retention time necessitates a refresh operation because each bit of the DRAM is stored as an amount of electrical charge in a storage capacitor. Power consumption for the refresh operation increases in proportion to memory capacity. A new method is proposed to reduce the refresh power consumption dynamically, when full memory capacity is not required, by effectively extending the memory cell retention time. Conversion from 1 cell/bit to 2N cells/bit reduces the variation of retention times among memory cells. The proposed method reduces the frequency of disturbance and power consumption by two orders of magnitude. Furthermore, the conversion itself can be realized very simply from the structure of the DRAM array circuit, while maintaining all conventional functions and operations in the full array access mode.

Process Variation-Aware Nonuniform Cache Management in a 3D Die-Stacked Multicore Processor

Process variations in integrated circuits have significant impact on their performance, leakage, and stability. This is particularly evident in large, regular, and dense structures such as DRAMs. DRAMs are built using minimized transistors with presumably uniform speed in an organized array structure. Process variation can introduce latency disparity among different memory arrays. With the proliferation of 3D stacking technology, DRAMs become a favorable choice for stacking on top of a multicore processor as a last level cache for large capacity, high bandwidth, and low power. Hence, variations in bank speed create a unique problem of nonuniform cache accesses in 3D space. In this paper, we investigate cache management techniques for tolerating process variation in a 3D DRAM stacked onto a multicore processor. We modeled the process variation in a four-layer DRAM memory, including cell transistor, capacitor trench, and peripheral circuit, to characterize the latency and retention time variations among different banks. As a result, the notion of fast and slow banks from the core's standpoint is no longer associated with their physical distances with the banks. They are determined by the different bank latencies due to process variation. We develop cache migration schemes that utilize fast banks while limiting the cost due to migration. Our experiments show that there is a great performance benefit in exploiting fast memory banks through migration. On average, a variation-aware management can improve the performance of a workload over the baseline (where one of the slowest bank speed is assumed for all banks) by 16.5 percent. We are also only 0.8 percent away in performance from an ideal memory where no process variation is present.