Constrained Molecule Generation Modelled Using the Grammar Constraint

Many demanding streaming applications share functional and structural similarities with other applications in their respective domain, e.g., video analytics, software-defined radio, and radar. This opens the opportunity for specialization to achieve the needed efficiency and/or performance.Platforms integrating many accelerators (ACCs) are a primary approach for efficient, high-performance stream computing. However, designing one platform for each application is not economical due to the high costs of nonrecurring engineering (NRE) and time-to-market (TTM). To this end, the concept of domain platforms is proposed, which takes advantage of similarities across applications and designs one unified platform to accelerate a domain of applications instead of focusing on a single reference application. This dissertation approaches designing domain platforms from a function-level (kernel-level) acceleration through a heterogeneous ACC-rich platform, where each ACC is specialized to accelerate a particular function. There is a great challenge to select ACCs allocated in the domain platform, considering the large design space and performance balance across many applications. However, current Design Space Exploration (DSE) tools only focus on an individual application in isolation (e.g., one particular vision flow) for allocating a platform, but not a set of similar applications. Multiple applications are only considered in the application bindings to an existing platform. Hence, optimizations that occur due to considering multiple applications simultaneously are missed in platform design. New DSE methodologies and tools are needed with a broader scope of application sets instead of individual applications. This dissertation introduces a novel Domain Design Space Exploration (DmDSE) methodology that broadens the DSE scope from an individual application to allocating a new heterogeneous ACC-rich platform for many applications.This dissertation defines domain characteristics to express the inherent structural and functional similarities across applications and metrics to quantitatively assess domain platforms with a balanced focus across various applications. Two novel algorithms, Dynamic Score Selection (DSS) and GenetIc Domain Exploration (GIDE), are introduced to do DmDSE and allocate domain platforms, which improve performance over application-specific platforms by 58% and 75% when only considering monolithic (single function) ACCs. A finer granularity of ACCs is beneficial for generality and performance under an area constraint. Considering multi-granularity similarities across functions, Multi-Granularity DmDSE (MG-DmDSE) extends DSE through allocating finer granular and configurable ACCs and decomposing functions in the binding. MG-DmDSE produces a more general platform that benefited 87.5% of applications, while only 50% from GIDE. With a huge design space, DmDSE suffers from a long exploration time due to a large number of complex platform evaluations. Heuristics and machine learning methods are proposed to speed up DmDSE.This dissertation introduces Greedy Guided Mutation (GGM) to speed up the mutation in the GIDE algorithm, which calculates an ACC score according to current allocation to guide mutation. Alternatively, Rapid Domain Platform Performance Prediction (RDP^3) methods are introduced to replace a large number of the slow platform assessment in domain DSE, which avoids the complex application binding exploration. In the experiments, GGM reduces 84.8% of exploration time with a 0.23% loss of the final OpenVX domain platform's performance. RDP^3, using a machine learning method, yields an even more significant speedup, saving 90.8% of exploration time with only 0.0003% performance loss. DmDSE is a milestone to broaden DSE scope from individual applications to the domain level. It tremendously pushes the domain platform design from an experience-guided, manual effort into an automated process.--Author's abstract

In recent years, field-programmable gate arrays (FPGAs) have been increasingly deployed in datacenters as programmable accelerators that can offer software-like flexibility and custom-hardware-like efficiency for key datacenter workloads. To improve the efficiency of FPGAs for these new datacenter use cases and data-intensive applications, a new class of reconfigurable acceleration devices (RADs) is emerging. In these devices, the FPGA fine-grained reconfigurable fabric is a component of a bigger monolithic or multi-die system-in-package that can incorporate general-purpose software-programmable cores, domain-specialized accelerator blocks, and high-performance networks-on-chip (NoCs) for efficient communication between these system components. The integration of all these components in a RAD results in a huge design space and requires re-thinking the implementation of applications that need to be migrated from conventional FPGAs to these novel devices. In this work, we introduce RAD-Sim, an architecture simulator that allows rapid design space exploration for RADs and facilitates the study of complex interactions between their various components. We also present a case study that highlights the utility of RAD-Sim in re-designing applications for these novel RADs by mapping a state-of-the-art deep learning (DL) inference FPGA overlay to different RAD instances. Our case study illustrates how RAD-Sim can capture a wide variety of reconfigurable architectures, from conventional FPGAs to devices augmented with hard NoCs, specialized matrix-vector blocks, and 3D-stacked multi-die devices. In addition, we show that our tool can help architects evaluate the effect of specific RAD architecture parameters on end-to-end workload performance. Through RAD-Sim, we also show that novel RADs can potentially achieve 2.6× better performance on average compared to conventional FPGAs in the key DL application domain.

During the last few decades an unprecedented technological growth has been at the center of the embedded systems design paramount, with Moore’s Law being the leading factor of this trend. Today in fact an ever increasing number of cores can be integrated on the same die, marking the transition from state-of-the-art multi-core chips to the new many-core design paradigm. Despite the extraordinarily high computing power, the complexity of many-core chips opens the door to several challenges. As a result of the increased silicon density of modern Systems-on-a-Chip (SoC), the design space exploration needed to find the best design has exploded and hardware designers are in fact facing the problem of a huge design space. Virtual Platforms have always been used to enable hardware-software co-design, but today they are facing with the huge complexity of both hardware and software systems. In this thesis two different research works on Virtual Platforms are presented: the first one is intended for the hardware developer, to easily allow complex cycle accurate simulations of many-core SoCs. The second work exploits the parallel computing power of off-the-shelf General Purpose Graphics Processing Units (GPGPUs), with the goal of an increased simulation speed. The term Virtualization can be used in the context of many-core systems not only to refer to the aforementioned hardware emulation tools (Virtual Platforms), but also for two other main purposes: 1) to help the programmer to achieve the maximum possible performance of an application, by hiding the complexity of the underlying hardware. 2) to efficiently exploit the high parallel hardware of many-core chips in environments with multiple active Virtual Machines. This thesis is focused on virtualization techniques with the goal to mitigate, and overtake when possible, some of the challenges introduced by the many-core design paradigm.

This paper presents our new developed FOCAP tool (Framework for optimizing the Computer Architecture Performance) in order to gain a better understanding and familiarity of the students with new advanced learning methods and tools in the Microarchitecture Simulation and Optimization. At this stage, FOCAP allows a mono-objective automatic design space exploration (DSE) of a superscalar processor by varying several architectural parameters. Such DSE tools are very useful, since it is impossible to simulate all the configurations of a highly parameterized microarchitecture. Therefore, heuristic methods, local search algorithms and advanced machine learning methods are good candidates to find near-optimal configurations by evaluating a reduced number of configurations from the huge design space. Our application falls in the European trend of Framework projects that aim expanding research towards the advanced education field using modern learning technologies. The FOCAP tool facilitates the comprehending of theoretical questions, thus allowing students to feel more confident when studying microarchitecture optimization and DSE-related issues.

NRC-VABS: Normalized Reparameterized Conditional Variational Autoencoder with applied beam search in latent space for drug molecule design

ERDSE: efficient reinforcement learning based design space exploration method for CNN accelerator on resource limited platform

According to published statistics regarding pharmaceutical industries, there has been a significant increase in the demand for cosmetics and skincare products. As a result, there has also been an increase in interest by researchers to conduct experimental and theoretical studies on these industries. The outermost layer of the skin, which is the stratum corneum, plays a key role in determining the skin permeability. This layer represents the first and main barrier against dehydration and screening from external elements and pathogens. Research efforts have been made to better understand skin permeability. Experimental methodologies play a key role in pharmaceutical and dermatology studies. However, ethical standards of using animal and human models in in-vitro permeability studies are very strict in their admission process and approval time. Furthermore, the lack of available material and financial resources for setting up, operating and repeating experiments has also limited the required progress. These problems can be addressed by developing better mathematical models on skin permeability. Progress, regarding the mathematical modelling of solute transport in skin, has been achieved in the last decades. The mathematical modelling strategy is used as a skin permeability predictor tool based on common physicochemical properties of the skin and penetrated solute molecules. Many factors play an important role in the formulation of a mathematical model and in determining how reliable the predicted results are. Therefore, the more skin physiological details are included in the model, the more computational efforts are required to interpret the solute permeation mechanisms in the dermal region and simulate a drug transport in the transdermal administration. The mathematical complexities associated with formulating a mathematical model of drug transport in skin significantly limit the utilities of such models in practice. Thus, a balance between the model’s complexity and how realistic it is in describing the solute transport processes in skin has to be achieved. Therefore, a mathematical model formulated on principles familiar to the pharmaceutical scientific community could be of value. In this thesis, we have developed a novel mathematical model using a compartmental approach. The proposed model enables for simulating solute transport across the stratum corneum (SC) in a variety of exposure scenarios in percutaneous drug delivery systems. The obtained results were compared with the diffusion model outcomes – which can be considered the current ‘golden standard’ for modelling skin transport – and showed a good agreement. The mathematical foundations of the model are relatively simple and better aligned with the physiology of the stratum corneum. In addition, this work shows that performing a numerical simulation using the compartmental model is less demanding than using the diffusing model, especially for complex exposure scenarios, and can be performed using free software, such as Python. The research findings of this thesis contribute towards a better understanding of transport phenomena across the skin. Also, the compartmental model proposed in this work can be used to describe transport phenomena in the membrane. Due to the similarity between the governing equations of diffusion and heat transfer, the proposed approach also offers a mathematical framework for studying heat transfer problems in biomaterials with minimum model development or computational efforts.

The possibility of improving dental restorative materials is investigated through the addition of two different types of fillers to a polymeric resin. These fillers, consisting of porous alumina and TiO2 nanotubes, are compared based on their common physicochemical properties on the nanometric scale. The aim was to characterize and compare the surface morphological properties of composite resins with different types of fillers using analytical techniques. Moreover, ways to optimize the mechanical, surface, and aesthetic properties of reinforced polymer composites are discussed for applications in dental treatments. Filler-reinforced polymer composites are the most widely used materials in curing dental pathologies, although it remains necessary to optimize properties such as mechanical resistance, surface characteristics, and biocompatibility. Anodized porous alumina nanoparticles prepared by electrochemical anodization offer a route to improve mechanical properties and biocompatibility as well as to allow for the controlled release of bioactive molecules that can promote tissue integration and regeneration. The inclusion of TiO2 nanotubes prepared by hydrothermal treatment in the resin matrix promotes the improvement of mechanical and physical properties such as strength, stiffness, and hardness, as well as aesthetic properties such as color stability and translucency. The surface morphological properties of composite resins with anodized porous alumina and TiO2 nanotube fillers were characterized by Atomic Force Microscopy (AFM), Scanning Electron Microscopy (SEM), and X-ray chemical analysis. In addition, the stress-strain behavior of the two composite resins is examined in comparison with enamel and dentin.

Occurrence of monoacyl-diglucosyl-diacyl-glycerol and monoacyl-bis-glycerophosphoryl-diglucosyl-diacyl-glycerol in membranes of Acholeplasma laidlawii strain B-PG9

The study of evolutionary algorithms (EAs) has witnessed an impressive increase during the last decades. The need to explore this area is determined by the growing request for design and the optimization of more and more engineering problems in society, such as highway construction processes, food and agri-technologies processes, resource allocation problems, logistics and transportation systems, microarchitectures, suspension systems optimal design, etc. All of these matters refer to specific highly computational problems with a huge design space, hence the obvious need for evolutionary algorithms and frameworks, or platforms that allow for the implementing and testing of such algorithms and methods. This paper aims to comparatively analyze the existing software platforms and state-of-the-art multi-objective optimization algorithms and make a review of what features exist and what features might be included next as further developments in such tools, from a researcher’s perspective. Additionally, it is essential for a framework to be easily extendable with new types of problems and optimization algorithms, metrics and quality indicators, genetic operators or specific solution representations and results analysis and comparison features. After presenting the most relevant existing features in these types of platforms, we suggest some future steps and the developments we have been working on.

Computational design of thermoset nanocomposite coatings: Methodological study on coating development and testing

In this paper, we present an efficient exploration algorithm for architecture/compiler co-designs of application-specific instruction-set processors. The huge design space is spanned by processor architecture parameters as well as different compiler optimization strategies. The objective space is multi-dimensional including conflicting objectives such as hardware cost, execution time and code size. The goal of the presented exploration algorithm is to determine the set of Pareto-optimal designs and compiler settings for a given benchmark program.In a case study, while exploring Pareto-optimal designs for a given DSP benchmark program, we show that for a realistic architecture family, the huge search space may be reduced dramatically using proper techniques to prune search spaces that may not contain Pareto-optimal solutions. Finally, we analyse and present solutions on what is the best architecture for a mixture of benchmark programs, i.e., what are the best architecture/compiler co-designs to execute the DSPstone benchmark.

With the availability of advanced MPSoC and emerging Dynamic RAM (DRAM) interface technologies, an optimal allocation of logical data buffers to physical memory cannot be handled manually anymore due to the huge design space. An allocation does not only need to decide between an on-or off-chip memory, but also needs to take an increasing number of available memory channels, different bandwidth capacities and several routing possibilities into account. We formalize this problem and introduce a Mixed Integer Linear Programming (MILP) model based on two different optimization criteria. We implement the MILP model into a retargetable tool and present a case study with representative data of the Long-Term-Evolution (LTE) standard to show the real-life applicability of our approach.

Recently, an emerging non-volatile memory called Racetrack Memory (RM) becomes promising to satisfy the requirement of increasing on-chip memory capacity. RM can achieve ultra-high storage density by integrating many bits in a tape-like racetrack, and also provide comparable read/write speed with SRAM. However, the lack of circuit-level modeling has limited the design exploration of RM, especially in the system-level. To overcome this limitation, we develop an RM circuit-level model, with careful study of device configurations and circuit layouts. This model introduces Macro Unit (MU) as the building block of RM, and analyzes the interaction of its attributes. Moreover, we integrate the model into NVsim to enable the automatic exploration of its huge design space. Our case study of RM cache demonstrates significant variance under different optimization targets, in respect of area, performance, and energy. In addition, we show that the cross-layer optimization is critical for adoption of RM as on-chip memory.

Recent computer architectures can be configured in lots of different ways. To explore this huge design space, system simulators are typically used. As performance is no longer the only decisive factor but also e.g. power usage or the resource usage of the system it became very hard for designers to select optimal configurations. In this article we use a multi-objective design space exploration tool called FADSE to explore the vast design space of the Grid Alu Processor (GAP) and its post-link optimizer called GAPtimize. We improved FADSE with techniques to make it more robust against failures and to speed up evaluations through parallel processing. For the GAP, we present an approximation of the hardware complexity as second objective besides execution time. Inlining of functions applied as a whole program optimization with GAPtimize is used as example for a code optimization. We show that FADSE is able to thoroughly explore the design space for both GAP and GAPtimize and it can find an approximation of the Pareto frontier consisting of near-optimal individuals in moderate time.