Boosted black holes play an important role in General Relativity (GR), especially in relation to the binary black hole problem. Solving Einstein vacuum equations in the strong field regime had long been the holy grail of numerical relativity until the significant breakthroughs made in 2005 and 2006. Numerical relativity plays a crucial role in gravitational wave detection by providing numerically generated gravitational waveforms that help search for actual signatures of gravitational radiation exciting laser interferometric detectors such as LIGO, VIRGO and GEO600 here on Earth. Binary black holes orbit each other in an ever tightening adiabatic inspiral caused by energy loss due to gravitational radiation emission. As the orbits shrinks, the holes speed up and eventually move at relativistic speeds in the vicinity of each other (separated by ~ 10M or so where 2M is the Schwarzschild radius). As such, one must abandon the Newtonian notion of a point mass on a circular orbit with tangential velocity and replace it with the concept of black holes, cloaked behind spheroidal event horizons that become distorted due to strong gravity, and further appear distorted because of Lorentz effects from the high orbital velocity. Apparent horizons (AHs) are 2-dimensional boundaries that are trapped surfaces. Conceptually, one can think of them as 'quasi-local' definitions for a black hole horizon. This will be explained in more detail in chapter 2. Apparent horizons are especially important in numerical relativity as they provide a computationally efficient way of describing and locating a black hole horizon. For a stationary spacetime, apparent horizons are 2-dimensional cross-sections of the event horizon, which is itself a 3-dimensional null surface in spacetime. Because an AH is a 2-dimensional cross-section of an event horizon, its area remains invariant under distortions due to Lorentz boosts although its shape changes. This fascinating property of the AH can be attributed to the fact that it is a cross-section of a null surface, which, under the boost, still remains null and the total area does not change. Although this invariance of the area is conceptually easy to see it is less straightforward to derive this result. We present two different ways to show the area invariance. One is based on the spin-boost transformation of the null tetrad and the other a direct coordinate transformation of the boosted metric under the Lorentz boost. Despite yielding identical results the two methods differ significantly and we elaborate on this in much more detail. We furthermore show that the use of the spin-boost transformation is not well-suited for binary black hole spacetime and that the spin-boost is fundamentally different from a Lorentz boost although the transformation equations look very similar. We also provide a way to visualize the distorted horizons and look at the multi-pole moments of these surfaces under small boosts. We finish by summarizing our main results at the end and by commenting on the binding energy of the binary and how the apparent horizon is distorted due to presence of another black hole. / text

No description available.

In this paper, we provide a generalization of binary quadratic residue codes to the cases of higher power prime residues over the finite field of the same order, which we will call qth power residue codes. We find generating polynomials for such codes, define a new notion corresponding to the binary concept of an idempotent, and use this to find square root lower bound for the codeword weight of the duals of such codes, which leads to a lower bound on the weight of the codewords themselves. In addition, we construct a family of asymptotically bad qth power residue codes. / text

This report describes the architecture and implementation of a memory profiler for 3D graphics applications. The memory profiling is done for parts of the program which runs on the graphics processor and is responsible for rendering the image. The shaders are parsed and every memory instruction is instrumented with additional instruction for profiling. The results are then transferred from the video memory to CPU memory. Profiling is done for a frame and completes in less than three minutes. The report also describes various analyses that can be done using the results obtained from this profiler. The report discusses the design of an analytical cache model that can be used to identify candidate memory buffers suitable for caching among all the buffers used by an application. The profiler can segregate results for reads and writes separately, can handle all formats of texture access instructions and predicated instructions. / text

Compact binary coalescing systems: binary neutron stars, neutron star black hole pairs and binary black hole systems, represent promising candidates for gravitational wave first detection and have the potential to provide precise tests of the strong-field predictions of general relativity. Observations of binary black hole (BBH) systems will provide a wealth of information relevant to fundamental physics, astrophysics and cosmology. The search for such systems is a major priority of the Laser Interferometer Gravitational-Wave Observatory (LIGO) and Virgo collaborations. A major area of research within LIGO-Virgo analysis groups is incorporation of spin into the search template banks used for binary black hole systems. In this dissertation, I compare the injection efficiency and parameter recovery from three binary black hole searches. One of the searches presented here uses non-spinning templates and represents the standard LIGO search for binary black holes with total masses between 35 and 100M[circle with dot]. The other two use spin aligned and anti-aligned templates representing a future search for black hole binary systems with total masses between 35-100M[circle with dot]. One of the two spinning searches has the spin parameter set to zero, nonspinning, as a check of the spinning method. (Additionally the (anti-)aligned spin searches use a retooling of the standard pipeline taking advantage of a code base designed specifically to handle Advanced LIGO data.) All three searches were run on artificial data created by the Numerical Injection Analysis 2 collaboration (NINJA2) containing Gaussian noise and numerically generated signals modeling aligned and anti-aligned spinning binary black holes. I found that for the analyzed two weeks of data the three searches recover injections with nearly equal efficiency; however, the spinning search recovers the parameters of the injections more accurately than the non-spinning search. Specifically, the parameter recovery of the spins shows a correlation between the injected and recovered spins, and the addition of spin to the template bank improves the recovery of the signal-to-noise ratio and the chirp mass for an injected signal. While spin aligned situations are geometrically low probability configurations, there are plausible astrophysical effects that lead to alignment of spins prior to merger. Therefore my results show that the spin-aligned template bank search represents an improvement on the standard non-spinning search in the highmass region and should be pursued on real data. / text

This thesis focuses on the logical design of binary adders. It covers topics extending from cardinal numbers to carry skip optimization. The conventional adder designs are described in detail, including: carry completion, ripple carry, carry select, carry skip, conditional sum, and carry lookahead. We show that the method of parallel prefix analysis can be used to unify the conventional adder designs under one parameterized model. The parallel prefix model also produces other useful configurations, and can be used with carry operator variations that are associative. Parallel prefix adder parameters include group sizes, tree shape, and device sizes. We also introduce a general algorithm for group size optimization. Code for this algorithm is available on the World Wide Web. Finally, the thesis shows the derivation for some carry operator variations including those originally given by Majerski and Ling. / text

Η μελέτη των διπλών εκλειπτικών συστημάτων παραμένει ένα από τα πιο ενδιαφέροντα ερευνητικά πεδία της αστροφυσικής γιατί παρέχει την πιο ακριβή μέθοδο υπολογισμού των βασικών αστρικών μεγεθών (μαζών, ακτίνων, θερμοκρασιών) αλλά και κάθε μορφής αστρικής δραστηριότητας που αναπτύσσεται (κηλίδες, δίσκοι προαύξησης ύλης, αναπάλσεις) μέσα από τη μακρόχρονη μεταβολή της περιόδου τους. Επιπλέον επιτρέπει τον έλεγχο των θεωριών αστρικής εξέλιξης μέσα από την αλληλεπίδρασή τους αλλά και την πρόβλεψη παρουσίας τρίτων σωμάτων αστρικής ή μη φύσης. Σε αυτήν τη διατριβή μελετάται μία ιδιαίτερη κατηγορία διπλών εκλειπτικών συστημάτων σε επαφή, τα W UMa τα οποία αποτελούνται από αστέρες κύριας ακολουθίας οι οποίοι βρίσκονται σε υπερεπαφή, και χαρακτηρίζονται από τις μικρότερες περιόδους κι άρα από τη μικρότερη στροφορμή. Σκοπός της είναι τόσο η παρατηρησιακή μελέτη και ανάλυση συστημάτων W UMa που παρουσιάζουν ιδιαίτερο ενδιαφέρον όσο και η ανάγκη ανάπτυξης μεθοδολογίας μοντελοποίησης για την εξαγωγή της μέγιστης πληροφορίας και ελέγχου των εξαγόμενων φυσικών παραμέτρων κάτω από την εφαρμογή σύγχρονων στατιστικών μεθόδων. Το πρώτο μέρος της επικεντρώνεται στη σκιαγράφηση του θεωρητικού επιστημονικού πλαισίου περιγραφής των διπλών εκλειπτικών αστρικών συστημάτων σε επαφή. Το Κεφάλαιο1 αναφέρεται στη γεωμετρία των τροχιών και στη φυσική των αστέρων του συστήματος, στην περιγραφή του μοντέλου Roche και στην αλληλεπίδραση των μελών καθώς και στη δυναμική των συστημάτων όπως αυτή εκδηλώνεται από την παρατηρούμενη μεταβολή της περιόδου του. Το Κεφάλαιο 2 αναφέρεται στις φωτομετρικές παρατηρήσεις που έγιναν κατά την διάρκεια της παρούσας διατριβής καθώς και στις αυτοματοποιημένες μεθόδους επεξεργασίας και ανάλυσης των παρατηρησιακών αστρονομικών δεδομένων οι οποίες αναπτύχθηκαν. Το Κεφάλαιο 3 περιγράφει τις σύγχρονες τεχνικές μοντελοποίησης και τα προγράμματα ανάλυσης για τηλύση του αντίστροφου προβλήματος. Το Κεφάλαιο 4 περιγράφει την αναζήτηση, εξόρυξη και αυτόματη ανάλυση δεδομένων από παρατηρησιακές αστρονομικές επισκοπήσεις και την εφαρμογή του κώδικα τεχνητών νευρωνικών δικτύων EBAI. Το δεύτερο μέρος επικεντρώνεται στην αναλυτική μελέτη επιλεγμένων στενών διπλών συστημάτων W UMa με ιδιαίτερο ενδιαφέρον. Σε όλα τα επιλεγμένα συστήματα παρουσιάζεται η φωτομετρική τους μελέτη κάτω από το φως νέων BV RcIc παρατηρήσεων, η μελέτη της περιόδου, το εξαγόμενο μοντέλο σε συνδυασμό με φασματοσκοπικά δεδομένα, ο προσδιορισμός τροχιακών και φυσικών παραμέτρων και η θέση τους σε εξελικτικά διαγράμματα. Επίσης γίνεται η διερεύνηση του προτεινόμενου μοντέλου για μοναδικότητα στο χώρο των λύσεων μέσα από σάρωση με εισαγωγή διαταραχών ή άλλων στατιστικών μεθόδων και ο στατιστικός προσδιορισμός των σφαλμάτων των παραμέτρων. Συγκεκριμένα στο Κεφάλαιο5 παρουσιάζεται το ενοποιημένο μοντέλο του συστήματος TY Boo δύο κηλίδων που ερμηνεύει τη συμπεριφορά του την περίοδο 1969-2011 και τη μελέτη της περιόδου του. Σύμφωνα με αυτό το σύστημα TY Boo ανήκει στην υποκατηγορία W των W UMa με μικρό βαθμό επαφής f = (7.6 ±0.8)% και η μακρόχρονη μελέτη της περιόδου του δείχνει μακροχρόνια μείωση (dP/dt = −3.65x10^−8 d yr^−1 ) και μία περιοδικότητα (P3 = 58.9yrs, A = 0.0254 days)η οποία, χωρίς να αποκλείεται η παρουσία τρίτου σώματος, ερμηνεύεται με ενεργό μαγνητικό κύκλο που προκαλεί την εμφάνιση κηλίδων.Στο Κεφάλαιο 6 παρουσιάζεται για πρώτη φορά η λεπτομερής ανάλυση του συστήματος WUMa, FI Boo το οποίο ταξινομείται στην υποκατηγορία W, με βαθμό επαφής f = (50.15 ±8.10) %, κάτω από την παρουσία τρίτου σώματος το οποίο μπορεί να παίζει σημαντικό ρόλο στη δημιουργία και την εξέλιξή του. Σύμφωνα με το προτεινόμενο μοντέλο προσδιορίζονται οι μάζες Mh =0.40 ± 0.05 Msun , Mc =1.07 ± 0.05 Msun , και οι θερμοκρασίες Th = 5746±33Κ, Tc = 5420±46 Κ των αστέρων-μελών του καθώς και των προγεννητόρων τους (1.71 ±0.10 Msun και 0.63 ± 0.01 Msun , αντίστοιχα) και διερευνάται το εξελικτικό στάδιο του αστέρα με τη μεγαλύτερη μάζα μέσα από ισόχρονες.Στο Κεφάλαιο 7 διερευνάται σε βάθος το σύστημα σε επαφή V1003 Her, το οποίο χαρακτηρίζεται από μεταβολή μικρού πλάτους, με πολλαπλά μοντέλα για τα οποία ελέγχεται η σταθερότητα των με ανεξάρτητες μεθόδους. Το επικρατέστερο από αυτά δείχνει ότι το σύστημα είναι W UMa με βαθμό επαφής f = (36±10) %, με ασυνήθιστα άνισες θερμοκρασίες μεταξύ των αστέρων του 550 Κ, και λόγο μαζών q = 0.373. Η μικρή του κλίση i = 38^o ±1^ο το καθιστά ένα από τα μεγαλύτερης μάζας στην υποκατηγορίας W.Στο Κεφάλαιο 8 εξετάζεται η φύση του διπλού συστήματος HI Dra, η λεπτομερής ανάλυση του οποίου δείχνει ότι είναι W UMa με μικρό βαθμό επαφής f = (24 ± 4) % και διαφορά θερμοκρασίας μεταξύ των αστέρων του 330 Κ. Προτείνονται και διερευνώνται εξονυχιστικά δύο μοντέλα με κηλίδες από τα οποία επικρατέστερο, με ψυχρή κηλίδα πάνω στον μικρότερης μάζας και ψυχρότερο αστέρα- τον τοποθετεί στην υποκατηγορία Α. Προσδιορίζονται οι φυσικές παράμετροι των μελών του (M = 1.72 ± 0.08Msun , Mc = 0.43 ± 0.02 Msun , Rh =1.98 ± 0.03Rsun , Rc = 1.08 ± 0.02 Rsun , Lh = 9.6 ± 0.1 Lsun , Lc = 2.4 ± 0.1 Lsun ) και των προγεννητόρων τους (1.11 ± 0.03 Msun και 2.25 ± 0.07 Msun, αντίστοιχα) και προσεγγίζεται η ηλικία του συστήματος σε 2.4 Gyr .Τέλος στο Κεφάλαιο 8 παρουσιάζονται τα πρώτα αποτελέσματα από τις φωτομετρικές παρατηρήσεις με το τηλεσκόπιο Αρίσταρχος 2.3 m, διπλών εκλειπτικών συστημάτων με ιδιαίτερο ενδιαφέρον που ανακαλύφθηκαν ή παρατηρήθηκαν από την διαστημική αποστολή Kepler με περιόδους περιφοράς < 0.45 d, τα οποία παρουσιάζουν ασυμμετρίες στην καμπύλη φωτός, χρονικές μεταβολές των εκλείψεων των μελών τους ή είναι εν δυνάμει τριπλά συστήματα. Τα πρώτα αποτελέσματα αφορούν τα συστήματα KIC 11246163 και KIC 4563150τα οποία ταξινομούνται ως συστήματα τύπου W UMa, W υποκατηγορίας με πιθανό τριτο συνοδό. / The study of eclipsing binary systems remains one of the most powerful research fields in stellar astrophysics because it provides the primary source of calculating fundamentals properties of stars (masses, radii, temperatures) and every form of stellar activity (spots, accreting discs, pulsations) through photometry and spectroscopy or/and through the study of long-term variability of their period. Additionally it allows testing of stellar structure and evolution theories and the prediction of third body companions of stellar or sub-stellar origin. This thesis is focused on the study of W UMa type eclipsing binaries containing main sequence stars in overcontact configuration with short periods and,therefore the small angular momentum. The aim is both the multiband observational study, analysis and investigation of WUMa systems of particular interest but also the development of modeling methodology in order to extract the maximum information and to determine a detailed assessment of the parameter uncertainties through the application of modern statistical methods. The first part of the Doctoral Thesis outlines the theoretical framework for describing eclipsing binary stars. Chapter 1 refers to the physics and geometry of orbits and components, the description of the Roche model that led to the classification based on equipotential surfaces, the computation of the total radiated flux in the direction of the observer by including all the corrections needed and the description of systems dynamics as manifested by the observed change of their period. Chapter2 presents the photometric observations made during this research and the automated methods of processing and analysis (pipelines) which were developed. Chapter 3 sketches the modern modeling and analysis software techniques for the solution of inverse problem(Differential Corrections, Levenberg-Marquardt, Downhill Simplex, Genetic Algorithms,Heuristic Scanning, Bootstrap resampling, Metropolis –Hasting Markov Chain MonteCarlo), their implementation to programs developed as part of the present thesis and their application to real data. These also represent the methodology of the analyses implemented for the study of the WUMa systems, described in Part 2. Finally, Chapter4 describes the development of a pipeline for the exploration, extraction and automatic analysis of data from astronomical surveys and its application to ASAS database. In the same chapter the first results from the training and validation of the Artificial Neural Network (ANN), EBAI, and the best topology for the ANN are presented.The second part focuses on the detailed study of selected W UMa systems with particular interest. This includes: the new multiband photometric observations, the longterm period variation, the exported model from photometric and spectroscopic data (from the literature), the determination of orbital and physical parameters and their position in evolutionary diagrams, the investigation of the proposed model for uniqueness in the parameter space through heuristic scanning with parameter kicking or other statistical methods- and the statistical determination of the uncertainties of the derived parameters.In Chapter 5, new CCD four-color light curves of TY Boo made on eight nights over2010–2011 were analyzed in comparison with historical light curves obtained from 1969through 2011. The light curves could all be represented by a unique geometry and by wavelength consistent phototometric parameters of a two-spot model on either stellar component. It is confirmed that TY Boo is a shallow W-type contact binary system with a degree of contact factor of f = 7.6 ± 0.8%. A period investigation based on all available data shows a long-term decrease (dP/dt = −3.65 ×10−8 days yr−1 ) and an oscillation (P3 = 58.9 yr, A = 0.0254 days). Without ruling out the presence of a tertiary companion, the weight of evidence points to an active cyclic magnetic activity that causes spot formation rather than an unseen companion. Mass transfer between the components and angular momentum loss are also considered as possible mechanisms.In Chapter 6, a detailed analysis of the interesting W UMa binary FI Boo in view of the spectroscopic signature of a third body through photometry, period variation, and a thorough investigation of solution uniqueness is presented. We obtained new BVRcIc photometric data that, when combined with spectroscopic data, enable us to analyze the system FI Boo and determine its basic orbital and physical properties through PHOEBE,as well as the period variation by studying the times of the minima. This combined approach allows us to study the long-term period changes in the system for the first time in order to investigate the presence of a third body and to check extensively the solution uniqueness and the uncertainties of derived parameters. Our modeling indicates that FIBoo is a W-type moderate (f = 50.15% ± 8.10%) overcontact binary with component masses of Mh = 0.40 ± 0.05 M and Mc =1.07 ± 0.05 M , temperatures of Th = 5746 ±33 K and Tc = 5420 ± 56 K, and a third body, which may play an important role in the formation and evolution. The results were tested by heuristic scanning and parameter kicking to provide the consistent and reliable set of parameters that was used to obtain the initial masses of the progenitors (1.71 ± 0.10 M and 0.63 ± 0.01 M , respectively). We also investigated the evolutionary status of massive components with several sets of widely used isochrones.In Chapter 7, an extensive analysis of the low amplitude, contact binary V1003 Her is presented, based on the new VRc Ic , CCD photometric light curves in combination with published radial velocity (RV) curves. We investigate the stable configurations for the system with two independent methods and modeling tools: PHOEBE, ROCHE, via heuristic scanning and genetic algorithms, although the very low inclination of the system can place limitations. All methods indicate that V1003 Her is most likely in overcontact state with unequal components with temperature difference of 550 K, a mass ratio of q= 0.373 and a contact degree of 36 ± 10 %. As it is viewed at the very low inclination of i ∼ 38^◦ ± 1^◦ , if its configuration is confirmed, it will be among the most massive W-subtype of W UMa systems. This conclusion is also supported by other published models.However, in order to conclude reliable physical properties of the system, the high-precise based-ground photometry (or satellite photometry) and spectroscopic follow-up of V1003Her is required.In Chapter 8, a detailed investigation of the low-amplitude contact binary HI Dra is presented, based on the new VRc Ic CCD photometric light curves (LCs) combined with published radial velocity (RV) curves. Our completely covered LCs were analyzed using PHOEBE and revealed that HI Dra is an overcontact binary with low fill-out factor f = 24 ± 4 (%) and temperature difference between the components of 330 K.Two spotted models are proposed to explain the LC symmetry, between which the A subtype of W UMa type eclipsing systems, with a cool spot on the less massive and cooler component, proves to be more plausible on evolutionary grounds. The results and stability of the solutions were explored by heuristic scan and parameter perturbation to provide a consistent and reliable set of parameters and their errors. Our photometric modeling and RV curve solution give the following absolute parameters of the hot and cool components, respectively: Mh = 1.72 ± 0.08 Msun and Mc = 0.43 ± 0.02 Msun , Rh =1.98 ± 0.03 Rsun and Rc = 1.08 ± 0.02 Rsun , and Lh = 9.6 ± 0.1 Lsun and Lc = 2.4 ± 0.1Lsun . Based on these results the initial masses of the progenitors (1.11 ± 0.03 Msun and2.25 ± 0.07 Msun , respectively) and a rough estimate of the age of the system of 2.4 Gyr are discussed.Finally, in Chapter 9, our first results from ground based follow up photometric observation of interesting eclipsing binary systems (EBs) from Kepler field are presented.The program was launched in 2013 with the 2.3 m Aristarchos telescope at Helmos Observatory, Greece including eclipsing binary systems with periods < 0.45d and Kp (mag)=12.6-16 mag. The included targets in this program show light curve asymmetries,Eclipse Timings Variation or they are third body candidates. Modern analysis techniques such as heuristic scanning with parameter perturbation and genetic algorithm(PIKAIA), enable to reveal and optimize the astrophysical parameters of selected EBs.The results from BVRI photometry for the third body candidates, WUMa eclipsing binary systems of W-subtype, KIC11246163 and KIC4563150, are presented for the first time, as well as conclusions derived so far.

Although forest properties are known to influence snowpack accumulation and spring runoff, the processes underlying the impacts of forest canopy cover on the input of snowmelt to the catchment remain poorly characterized. In this study I show that throughfall and canopy shading can combine to result in maximal snowpacks in forests of moderate canopy density. Snow depth and density data taken shortly before spring melt in the Jemez Mountains of New Mexico show strong correlation between forest canopy density and snow water equivalent, with maximal snow accumulation in forests with density between 25 and 45%. Forest edges are also shown to be highly influential on local snow depth variability, with shaded open areas holding significantly deeper snow than either unshaded open or deep forest areas. These results are broadly applicable in improving estimates of water resource availability, predicting the ecohydrological implications of vegetation change, and informing integrated water resources management.

Multiframe image superresolution has been an active research area for many years. In this approach image processing techniques are used to combine multiple low-resolution (LR) images capturing different views of an object. These multiple images are generally under-sampled, degraded by optical and pixel blurs, and corrupted by measurement noise. We exploit diversities in the imaging channels, namely, the number of cameras, magnification, position, and rotation, to undo degradations. Using an iterative back-projection (IBP) algorithm we quantify the improvements in image fidelity gained by using multiple frames compared to single frame, and discuss effects of system parameters on the reconstruction fidelity. As an example, for a system in which the pixel size is matched to optical blur size at a moderate detector noise, we can reduce the reconstruction root-mean-square-error by 570% by using 16 cameras and a large amount of diversity in deployment.We develop a new technique for superresolving binary imagery by incorporating finite-alphabet prior knowledge. We employ a message-passing based algorithm called two-dimensional distributed data detection (2D4) to estimate the object pixel likelihoods. We present a novel complexity-reduction technique that makes the algorithm suitable even for channels with support size as large as 5x5 object pixels. We compare the performance and complexity of 2D4 with that of IBP. In an imaging system with an optical blur spot matched to pixel size, and four 2x2 undersampled LR images, the reconstruction error for 2D4 is 300 times smaller than that for IBP at a signal-to-noise ratio of 38dB.We also present a transform-domain superresolution algorithm to efficiently incorporate sparsity as a form of prior knowledge. The prior knowledge that the object is sparse in some domain is incorporated in two ways: first we use the popular L1 norm as the regularization operator. Secondly we model wavelet coefficients of natural objects using generalized Gaussian densities. The model parameters are learned from a set of training objects and the regularization operator is derived from these parameters. We compare the results from our algorithms with an expectation-maximization (EM) algorithm for L1 norm minimization and also with the linear minimum mean squared error (LMMSE) estimator.

This research pursues the use of powerful BDD-based functional circuit analysis to evaluate some approaches to test set generation. Functional representations of the circuit allow the measurement of information about faults that is not directly available through circuit simulation methods, such as probability of random detection and test-space overlap between faults. I have created a software tool that performs experiments to make such measurements and augments existing test generation strategies with this new information. Using this tool, I explored the relationship of fault model difficulty to test set length through fortuitous detection, and I experimented with the application of function-based methods to help reconcile the traditionally opposed goals of making test sets that are both smaller and more effective.