Off-line Sampling Research Articles

An Informed-Bi-Quick RRT* Algorithm Based on Offline Sampling: Motion Planning Considering Multiple Constraints for a Dual-Arm Cooperative System

Aiming to address problems such as low sampling success rate and long computation time in the motion planning of a dual-arm cooperative system with multiple constraints, this paper proposes an Informed-Bi-Quick RRT* algorithm based on offline sampling. First, in the process of pre-sampling, the new algorithm relaxes the approximation of constrained manifolds by introducing the idea of incremental construction, and it incorporates the stochastic gradient descent method to replace global random sampling with local random sampling, which enriches the data set and shortens the offline sampling time of the data set. Second, the new algorithm improves the original Quick-RRT* algorithm by combining the two-tree idea and the multi-target bias expansion strategy, and it improves the adaptability of the algorithm to different obstacle environments. In addition, the loosely constrained motion and tightly constrained motion in the two-arm cooperative system are analyzed, and the adaptive planning of the two-arm trajectory in different motions is described in detail. In this paper, the two-arm cooperative model constructed with UR5 and UR10 robot arms is studied, and the ability of the proposed algorithm to deal with multiple constraints is verified by simulating assembly and handling tasks. The experimental results show that compared with other methods, the proposed algorithm has obvious advantages in path quality and planning efficiency.

Contaminant characterization of odor in soil of typical pesticide-contaminated site with shallow groundwater.

Odor emission from the soil of pesticide-contaminated sites is a prominent environmental problem in China, but there are very few researches about the component and spatial distribution of odorous substances in the soil of contaminated sites. In this paper, to investigate the odor pollution condition of an organophosphorus pesticide production site in a city of South China, the odor pollutants in the soil and soil gas were analyzed and the key odor-contributing substances were identified. Besides, the correlation between the concentrations of odorous substances in soil and soil gas was analyzed, and the measured results were compared with the predicted results by the linear model and DED model. An off-line soil gas sampling device was designed to collect the gas emitted from soil because the groundwater level in the site was too shallow to build a soil gas well. The key odor substances were screened from the detection results of soil gas via odor activity value (OAV) analysis, which revealed that the key odorous substances included benzene, ethylbenzene, ammonia, toluene, m,p-xylene, methyl sulfide, dimethyl disulfide, and formaldehyde. Furthermore, the spatial distribution of the odor substances in the soil of the pesticide-contaminated site was closely related to the layout and geologic structure of the site. The odor pollutants in soil were mainly distributed near the phosmet production workshop and the drainage ditch network. As for the deep distribution, the odorous substances were mainly enriched in the silty clay or clay layer (5.6-11m), followed by the sludge layer (1-3.6m). Finally, the predicted model (linear model and DED model) analysis suggested that the linear model was more suitable for predicting the concentration of odorous substances in the soil gas with the detection data of soil in this pesticide-contaminated site.

Real-time environmental surveillance of SARS-CoV-2 aerosols

Real-time surveillance of airborne SARS-CoV-2 virus is a technological gap that has eluded the scientific community since the beginning of the COVID-19 pandemic. Offline air sampling techniques for SARS-CoV-2 detection suffer from longer turnaround times and require skilled labor. Here, we present a proof-of-concept pathogen Air Quality (pAQ) monitor for real-time (5 min time resolution) direct detection of SARS-CoV-2 aerosols. The system synergistically integrates a high flow (~1000 lpm) wet cyclone air sampler and a nanobody-based ultrasensitive micro-immunoelectrode biosensor. The wet cyclone showed comparable or better virus sampling performance than commercially available samplers. Laboratory experiments demonstrate a device sensitivity of 77–83% and a limit of detection of 7-35 viral RNA copies/m3 of air. Our pAQ monitor is suited for point-of-need surveillance of SARS-CoV-2 variants in indoor environments and can be adapted for multiplexed detection of other respiratory pathogens of interest. Widespread adoption of such technology could assist public health officials with implementing rapid disease control measures.

16 Advanced Solutions for Comprehensive Ambient Air Monitoring

Abstract Growing concerns over harmful exposure to low levels of hazardous air pollutants has led to increasingly demanding volatile organic compound (VOC) monitoring programs. Regulations are evolving to increase monitoring frequency, expand the list of target compounds and set lower detection limits in ambient air. Alongside this ambient and indoor air research continues to identify new, challenging compounds. There have long been specific standard methods and regulations for subsets of compounds, each using different instrumentation. For example method TO-15 for air toxics requires TD-GC-MS, ozone precursors monitoring uses TD-GC-FID and aldehydes are sampled with DNPH cartridges, analyzed by HPLC. Each compound set has its own challenges. Even so there is a drive to combine all three analyses into a single analytical run with the possibility of monitoring in online and offline. Additional considerations are high levels of humidity and large sample volumes, required to meet lower reporting limits. These require sophisticated trapping and water management techniques. As such removal of water from canister and on-line samples without loss of polar species is paramount for successful analysis of the Ozone Precursors, Air Toxics and oxygenated VOCs. This presentation will address the challenges associated with this complex compound list and show how confident identification and quantitation, with maximum sensitivity, can be achieved using state of the art pre-concentration technology. The cryogen-free methodology allows remote, on-line analysis or off-line canister sampling, all with a sample-to-sample cycle time of <60 minutes, significantly reducing equipment overhead and analysis times whilst enhancing insight into air quality.

Off-line measuring sampling data identification parameters for digital twins mirroring load modelling and stability analysis

Currently, the methods used to represent loads do not differ between the characteristics that compose them or the nature of these. Therefore, the purpose of this research is to develop digital twins mirroring load models that can be used for more precise studies on power-flows and stability within the National Transmission Grid (NTG). Off-line sampling data of different electric measurements have been used in six substations of the Guayaquil (Ecuador) area. These values were organized by statistical methods and by time periods, to determine the parameters that make up the static load model. Dynamic models are also constructed for the same six substations using the analysis of current and voltage signals obtained from the substations. All data is organized to show a digital twin mirroring visual representation of the disturbances that may occur in the substation buses. A more accurate description of the static and dynamic responses can be obtained by replacing the general model that is currently used by engineers and planners with off-line sampling data. Digital twins help the electric utility businesses gather, visualise, and contextualise data from different sources, and enable to act on data, and to understand what-if modelling stability scenarios.

Boron isotopes in blue diamond record seawater-derived fluids in the lower mantle

Boron is a quintessential crustal element but its conspicuous presence in diamond – a mantle mineral – raises questions about potential subduction pathways for boron and other volatiles. It has been a long-standing goal to characterize the isotopic composition of boron in blue, boron-bearing (Type IIb) diamonds to reveal its origin. Mineral inclusions indicate that Type IIb diamonds crystallize at transition zone to lower mantle depths, meaning that if the boron is subducted it would trace a pathway of volatile elements into the deep mantle. Here, using off-line laser ablation sampling, we present the first boron isotope compositions, along with trace element contents and carbon isotope compositions, of a suite of blue diamonds mainly sourced from the Cullinan diamond mine in South Africa. The ten analyzed blue diamonds have a wide range in δ11B, between −9.2 ± 2‰ and −0.5 ± 2‰, compared to the more restricted range for mid-ocean ridge basalts (−7.1 ± 0.9‰). Carbon isotope values for the blue diamonds range between −20.6 and −1.8‰, with a mode at −17‰, significantly more negative than the main mantle mode at −5‰. Combined, the boron and carbon isotope compositions require fluid input from subducted oceanic lithosphere down into deep mantle source regions of blue diamonds. This finding highlights a deep subduction pathway for volatiles to stoke the source regions of deeply-derived magmas.

Optimisation of PLS Calibrations for Filtered and Untreated Samples towards In-Line Monitoring of Phenolic Extraction during Red-Wine Fermentations

Infrared spectroscopy provides an efficient, robust, and multivariate means to measure phenolic levels during red-wine fermentations. However, its use is currently limited to off-line sampling. In this study, partial least squares (PLS) regression was used to investigate the possibility of using spectral data from minimally pre-treated or untreated samples for the optimisation of prediction calibrations towards an in-line monitoring set-up. The evaluation of the model performance was conducted using a variety of metrics. Limits of detection and quantification of the PLS calibrations were used to assess the ability of the models to predict lower levels of phenolics from the start of fermentation. The calibrations were shown to be useful for the quantification of phenolic compounds and phenolic parameters with minimal or no sample pre-treatment during red-wine fermentation. Upon evaluation of performance, the calibrations built for attenuated-transmission Fourier-transform mid-infrared (ATR-FT-MIR) and diffuse-reflectance Fourier-transform near-infrared (DR-FT-NIR) were shown to be the most suitable spectroscopy techniques for eventual application in an automated and in-line system with values for limits of detection and quantification being suitable for the entire duration of fermentation.

Economic production lot sizing under imperfect quality, on-line inspection, and inspection errors: Full vs. sampling inspection

A manufacturer’s choice between full inspection and sampling inspection can significantly impact economic production lot sizing decision for items with imperfect quality and inspection errors. This can further change depending on whether the manufacturer opts for on-line or off-line inspection. The present literature does not consider on-line inspection in the context of acceptance sampling for economic production lot sizing decision problem. In this paper, we develop an on-line sampling inspection model and argue that the choice between full inspection and sampling inspection usually depends on the threshold level of unit inspection cost. Our paper leads to several significant findings. Firstly, we establish that such an on-line sampling inspection scheme can lead to significant savings in inventory holding costs and lead to higher optimal expected profit. Secondly, we establish the intuitively anticipated result that for high unit inspection costs, on-line sampling leads to a higher optimal expected profit than a full inspection. Contrary to the common intuition, we also show that, even when inspection costs are low, sampling inspection could still be more profitable than full inspection in many situations. Finally, we establish that on-line sampling inspection provides a higher optimal expected profit than the usually preferred off-line sampling inspection. We also discuss the impact of relaxing some of the standard assumptions. In addition to our theoretical findings, through a detailed numerical analysis, we present the impact of changes in various parameters, particularly true defective fraction, proportion of inspection errors, sample size, rework effectiveness, and unit inspection cost on the choices between full versus sampling inspection.

Design & Characterization of Pb Electrode for Sensing Trichloroethylene in Ionic Liquid/Acetonitrile Co-Solvents

Design & Characterization of Pb Electrode for Sensing Trichloroethylene in Ionic Liquid/Acetonitrile Co-Solvents

Model-based calibration of a gas sensor array for on-line monitoring of ethanol concentration in Saccharomyces cerevisiae batch cultivation

The ethanol concentration in batch cultivation with the yeast Saccharomyces cerevisiae was predicted on-line using a gas sensor array. Headspace samples were pumped past the gas sensors array for 10 s every five minutes and the voltage changes of the sensors were measured. For the calibration procedure no off-line sampling was used. Instead, a theoretical model of the process has been applied to simulate the ethanol production at any given time. However, the kinetic parameters of the simulation model are unknown at the beginning of the calibration. It is demonstrated that these kinetic parameters of the theoretical process model can be acquired from the response of the gas sensor array alone. The calculated parameters result in a simulation model that is at least as accurate as a model whose parameters have been acquired by least squares fitting to off-line measurements. The root mean square error of calibration as well as the percentage error for validation sets were below 0.2 g L−1 and 7%, respectively. The obtained results indicate that the model-based calibrated gas sensor array can be a cheap alternative to other tools, such as spectroscopy-based methods, that are used for monitoring yeast cultivations.

Distribution, Sources, and Health Risk Assessment of Volatile Organic Compounds in Hefei City.

Volatile organic compounds (VOCs) are involved in the formation of ozone formation, which plays a significant role in regional air contamination and poses a great threat to human health. The VOCs were collected from the urban area of Hefei city via an off-line sampling method (SUMMA canister) and determined by gas chromatography-mass spectrometer. The average concentrations of VOCs were 17.65 ± 28.36 ppbv, which were mainly contributed by aromatics (10.02 ± 13.37 ppbv), haloalkane (5.37 ± 8.90 ppbv), ally halide (1.25 ± 3.36 ppbv), and aryl halid (1.02 ± 2.73 ppbv). According to the principal component analysis, three major sources were identified, including solvent use, vehicle exhaust, and industrial release, accounting for 70.6% of the total variance of the data. Health risk assessment was utilized to evaluate the potential adverse health effects of the individual VOC. The total hazard ratio in the selected area was higher than 1, where could pose health threat to exposed population. The cancer risk for benzene, carbon tetrachloride, trichloromethane, and 1, 2-dichloroethane were 4.8 × 10-5, 4.5 × 10-5, 3.3 × 10-5, and 2.5 × 10-5, respectively, indicating definite health risks.

The characteristics and mixing states of PM2.5 during a winter dust storm in Ningbo of the Yangtze River Delta, China

Dust particulates play an essential role for the nucleation, hygroscopicity and also contribute to aerosol mass. We investigated the chemical composition, size distribution and mixing states of PM2.5 using a single-particle aerosol mass spectrometer (SPAMS), Monitor for AeRosols and Gases (MARGA), and off-line membrane sampling from 2018.1.24 to 2018.2.20 at a coastal supersite in Ningbo, a port city in Yangtze River Delta, China. During the study campaign, the eastern part of China had experienced a wide range of cooling, sandstorm, and snowfall processes. The entire sampling campaign was categorized into five sub-periods based on the levels of PM2.5 and the ratios of PM2.5/PM10, namely clean (T1), heavy pollution (T2), light pollution (T3), dust (sandstorm) (T4) and cleaning pollution (T5) period. After comparing the average mass spectrum for each period, it shows that the primary ions, such as Ca2+and SiO3−, rarely coexist with each other within a single particle, but secondary ions generally coexist with these primary ions. Furthermore, the coexistence of each two different ions within a particle does not show distinct variation for the whole study periods. All these suggest that the absorption and partitioning of gaseous contaminants into the surface of primary aerosol through heterogeneous reactions are the major pathways of aging and growth of aerosol; and the merging of particles through collisions usually is insignificant. Although the absolute concentrations of nitrate and sulfate all increased with the PM2.5 concentrations, the relative equivalent concentrations of NO3− and SO42− displayed opposite trends; the relative contribution of sulfate decreased and that of nitrate increased as the increase of pollution. During the dust period, the relative equivalent concentrations of calcium and/or potassium ions in PM2.5 are significantly higher. This study provided deep insights about the mixing states and characteristics of particulate after long-range transport and a visualization tool for aerosol study.

A refined source apportionment study of atmospheric PM2.5 during winter heating period in Shijiazhuang, China, using a receptor model coupled with a source-oriented model

With the intensification of Chinese source control of air pollution, there is an urgent need for refined and rapid source apportionment techniques. A refined source apportionment method was constructed based on an off-line sampling dataset using a receptor model coupled with a source-oriented model, and the method was implemented in Shijiazhuang during the heating period. The refined results for source apportionment mainly included temporal, spatial, and source-category refinement data. The results indicated that the mean concentration of PM2.5 during the heating period was 96 μg/m3. Organic carbon (OC) and NO3− were found to be the dominant species of PM2.5 during the study. A high correlation was detected between elemental carbon (EC) and NO3− on polluted days, which was suggestive of the stagnant condition that accumulates EC and nitrate simultaneously. Secondary particle formation greatly promoted the occurrence of haze events. Secondary sources (34.9%), vehicle exhaust (18.6%), coal combustion (20.0%), industrial emissions (9.2%), crustal dust (9.7%), and biomass burning (7.6%) were the major sources during the heating period. The contributions of secondary sources and vehicle exhaust increased on polluted days, while those of coal combustion, industrial emissions and crustal dust decreased significantly. The contribution percentage of secondary sources from the southeast direction was basically the highest, while those of vehicle exhaust from the northwest or southeast directions were relatively higher as well, likely due to the distribution of traffic arteries. Based on the refined results for the source-category assessment, we found that the heating boilers (17.0%), non-road mobile (13.8%), diesel vehicles (10.4%), residential combustion (6.7%), road dust (5.5%), and architectural material industry (4.9%) were the major contributors to PM2.5. There was some uncertainty in the distribution proportions of the refined results, which were derived based on the emission inventory and the results of CALPUFF model.

In situ microscopy as online tool for detecting microbial contaminations in cell culture

Microbial contamination in mammalian cell cultures causing rejected batches is costly and highly unwanted. Most methods for detecting a contamination are time-consuming and require extensive off-line sampling. To circumvent these efforts and provide a more convenient alternative, we used an online in situ microscope to estimate the cell diameter of the cellular species in the culture to distinguish mammalian cells from microbial cells depending on their size. A warning system was set up to alert the operator if microbial cells were present in the culture. Hybridoma cells were cultured and infected with either Candida utilis or Pichia stipitis as contaminant. The warning system could successfully detect the introduced contamination and alert the operator. The results suggest that in situ microscopy could be used as an efficient online tool for early detection of contaminations in cell cultures.

OH reactivity of the urban air in Helsinki, Finland, during winter

A new instrument to measure total OH reactivity in ambient air based on the Comparative Reactivity Method (CRM) has been built and characterized at the Finnish Meteorological Institute in Helsinki, Finland. The system is based on the detection of pyrrole by a gas chromatograph with a photoionization detector and designed for long term studies. It was tested in a container close to the SMEAR III semi-urban station in Helsinki during the winter in February 2016. The sampling location next to the delivery area of the institute was influenced by local vehicle emissions and cannot be considered representative of background conditions in Helsinki. However, effects of nitrogen oxides on the measurements could be investigated there. During this campaign, 56 compounds were measured individually by 1) an in-situ gas chromatograph coupled to a mass spectrometer (GC/MS) and by 2) off-line sampling in canisters and on adsorbent filled cartridges taken at the container and subsequently analysed by GC-FID and liquid chromatography, respectively. In addition, nitrogen oxides were measured at the same location, while ozone, carbon monoxide and sulfur dioxide concentrations have been retrieved from the SMEAR III mast data. The comparison between the total OH reactivity measured and the OH reactivity derived from individual compound measurements are in better agreement for lower reactivity levels. Possible explanations for the differences are discussed in detail.

An interpolation-based parametric reduced order model combined with component mode synthesis

This paper presents an efficient interpolation-based parametric reduced order model with component mode synthesis. The off-line sampling of a large-scale structural dynamic system containing many parameters requires many computations to investigate the parameter-dependency of a dynamical system. Therefore, we introduce a dynamic substructuring scheme to execute off-line sampling in the subdomain level. In addition, we suggest discriminating the interpolation of the subsystem depending on the characteristics of each substructural mode. We synthesize the substructures, and we then construct a two-level, semi-parametrized ROM by reducing the degrees of freedom of the interface in the on-line stage. We then demonstrate the efficiency and accuracy of the present method by computing the relative eigenvalue errors and the transient responses of the system with random values for the parameters, and we conduct the design optimization of large-scale systems under dynamic loading conditions. A comparison of the present method with the full order model and a conventional reduced order model indicates that the present method is useful in carrying out an efficient analysis and design optimization for various large-scale structural dynamic systems.

Sheet metal forming optimization by using surrogate modeling techniques

Surrogate assisted optimization has been widely applied in sheet metal forming design due to its efficiency. Therefore, to improve the efficiency of design and reduce the product development cycle, it is important for scholars and engineers to have some insight into the performance of each surrogate assisted optimization method and make them more flexible practically. For this purpose, the state-of-the-art surrogate assisted optimizations are investigated. Furthermore, in view of the bottleneck and development of the surrogate assisted optimization and sheet metal forming design, some important issues on the surrogate assisted optimization in support of the sheet metal forming design are analyzed and discussed, involving the description of the sheet metal forming design, off-line and online sampling strategies, space mapping algorithm, high dimensional problems, robust design, some challenges and potential feasible methods. Generally, this paper provides insightful observations into the performance and potential development of these methods in sheet metal forming design.

Exposure assessment of workplace manufacturing titanium dioxide particles

With the widespread use of titanium dioxide (TiO2) human exposure is inevitable, but the exposure data on TiO2 are still limited. This study adopted off-line filter-based sampling combined with real-time activity-based monitoring to measure the concentrations in a workplace manufacturing TiO2 (primary diameter: 194 ± 108 nm). Mass concentrations (MCs) of aerosol particles in the packaging workshop (total dust: 3.17 mg/m3, nano dust: 1.22 mg/m3) were much higher than those in the milling workshop (total dust: 0.79 mg/m3, nano dust: 0.31 mg/m3) and executive office (total dust: 0.44 mg/m3, nano dust: 0.19 mg/m3). However, the MCs of TiO2 were at a relatively low level in the packaging workshop (total TiO2: 46.4 μg/m3, nano TiO2: 16.7 μg/m3) and milling workshop (total TiO2: 39.4 μg/m3, nano TiO2: 19.4 μg/m3) by ICP-MS. The number concentration (NC), surface area concentration (SAC) of aerosol particles potentially deposited in alveolar (SACA), and tracheobronchial (SACTB) regions of lungs in the packaging workshop were (1.04 ± 0.89) × 105 particles/cm3, 414.49 ± 395.07, and 86.01 ± 83.18 μm2/cm3, respectively, which were all significantly higher than those of the milling workshop [(0.12 ± 0.40) × 105 particles/cm3, 75.38 ± 45.23, and 17.60 ± 9.22 μm2/cm3, respectively] as well as executive office and outdoor background (p < 0.05). Activity-related characteristics were found in both workshops, and the time-variant characteristics showed very similar trends for 3 days in the packaging workshop. Our study provides important data of TiO2 particles exposure in the workplace.

Carbon Nanotube Emissions from Arc Discharge Production: Classification of Particle Types with Electron Microscopy and Comparison with Direct Reading Techniques.

Introduction:An increased production and use of carbon nanotubes (CNTs) is occurring worldwide. In parallel, a growing concern is emerging on the adverse effects the unintentional inhalation of CNTs can have on humans. There is currently a debate regarding which exposure metrics and measurement strategies are the most relevant to investigate workplace exposures to CNTs. This study investigated workplace CNT emissions using a combination of time-integrated filter sampling for scanning electron microscopy (SEM) and direct reading aerosol instruments (DRIs).Material and Methods:Field measurements were performed during small-scale manufacturing of multiwalled carbon nanotubes using the arc discharge technique. Measurements with highly time- and size-resolved DRI techniques were carried out both in the emission and background (far-field) zones. Novel classifications and counting criteria were set up for the SEM method. Three classes of CNT-containing particles were defined: type 1: particles with aspect ratio length:width >3:1 (fibrous particles); type 2: particles without fibre characteristics but with high CNT content; and type 3: particles with visible embedded CNTs.Results:Offline sampling using SEM showed emissions of CNT-containing particles in 5 out of 11 work tasks. The particles were classified into the three classes, of which type 1, fibrous CNT particles contributed 37%. The concentration of all CNT-containing particles and the occurrence of the particle classes varied strongly between work tasks. Based on the emission measurements, it was assessed that more than 85% of the exposure originated from open handling of CNT powder during the Sieving, mechanical work-up, and packaging work task. The DRI measurements provided complementary information, which combined with SEM provided information on: (i) the background adjusted emission concentration from each work task in different particle size ranges, (ii) identification of the key procedures in each work task that lead to emission peaks, (iii) identification of emission events that affect the background, thereby leading to far-field exposure risks for workers other than the operator of the work task, and (iv) the fraction of particles emitted from each source that contains CNTs.Conclusions:There is an urgent need for a standardized/harmonized method for electron microscopy (EM) analysis of CNTs. The SEM method developed in this study can form the basis for such a harmonized protocol for the counting of CNTs. The size-resolved DRI techniques are commonly not specific enough to selective analysis of CNT-containing particles and thus cannot yet replace offline time-integrated filter sampling followed by SEM. A combination of EM and DRI techniques offers the most complete characterization of workplace emissions of CNTs today.

Self-tuning Intel Restricted Transactional Memory

Transactional Memory is now supported in hardware in Intel processors (called RTM).The performance of RTM varies greatly depending on the workload and run-time usage.We provide a solution by relying on lightweight reinforcement learning techniques.We integrate our solution in GCC compiler, achieving transparency to programmers.We obtain quasi-optimal performance whereas static approaches are up to 10 × worse. The Transactional Memory (TM) paradigm aims at simplifying the development of concurrent applications by means of the familiar abstraction of atomic transaction. After a decade of intense research, hardware implementations of TM have recently entered the domain of mainstream computing thanks to Intel's decision to integrate TM support, codenamed RTM (Reduced Transactional Memory), in their last generation of processors.In this work we shed light on a relevant issue with great impact on the performance of Intel's RTM: the correct tuning of the logic that regulates how to cope with failed hardware transactions. We show that the optimal tuning of this policy is strongly workload dependent, and that the relative difference in performance among the various possible configurations can be remarkable (up to 10 × slow-downs).We address this issue by introducing a simple and effective approach that aims to identify the optimal RTM configuration at run-time via lightweight reinforcement learning techniques. The proposed technique requires no off-line sampling of the application, and can be applied to optimize both the cases in which a single global lock or a software TM implementation is used as fall-back synchronization mechanism.We propose and evaluate different designs for the proposed self-tuning mechanisms, which we integrated with GCC in order to achieve full transparency for the programmers. Our experimental study, based on standard TM benchmarks, demonstrates average gains of 60% over any static approach while remaining within 5% from the performance of manually identified optimal configurations.