The Mobile Waterway Monitor seeks to monitor water in an unexplored way. The module is buoyant and will float with the current as well as harvests solar energy. In short, the Mobile Waterway Monitor excels in size constraints, flexibility, extensibility, and capability. This current following monitor can show both measured trends like pH and interpolated trends like water speed, river contours, and elevation drop. The MWM strikes a balance between accuracy, portability, and being multi-purpose.

With the massive multithreading execution feature, graphics processing units (GPUs) have been widely deployed to accelerate general-purpose parallel workloads (GPGPUs). However, using GPUs to accelerate computation does not always gain good performance improvement. This is mainly due to three inefficiencies in modern GPU and system architectures.

First, not all parallel threads have a uniform amount of workload to fully utilize GPU’s computation ability, leading to a sub-optimal performance problem, called warp criticality. To mitigate the degree of warp criticality, I propose a Criticality-Aware Warp Acceleration mechanism, called CAWA. CAWA predicts and accelerates the critical warp execution by allocating larger execution time slices and additional cache resources to the critical warp. The evaluation result shows that with CAWA, GPUs can achieve an average of 1.23x speedup.

Second, the shared cache storage in GPUs is often insufficient to accommodate demands of the large number of concurrent threads. As a result, cache thrashing is commonly experienced in GPU’s cache memories, particularly in the L1 data caches. To alleviate the cache contention and thrashing problem, I develop an instruction aware Control Loop Based Adaptive Bypassing algorithm, called Ctrl-C. Ctrl-C learns the cache reuse behavior and bypasses a portion of memory requests with the help of feedback control loops. The evaluation result shows that Ctrl-C can effectively improve cache utilization in GPUs and achieve an average of 1.42x speedup for cache sensitive GPGPU workloads.

Finally, GPU workloads and the co-located processes running on the host chip multiprocessor (CMP) in a heterogeneous system setup can contend for memory resources in multiple levels, resulting in significant performance degradation. To maximize the system throughput and balance the performance degradation of all co-located applications, I design a scalable performance degradation predictor specifically for heterogeneous systems, called HeteroPDP. HeteroPDP predicts the application execution time and schedules OpenCL workloads to run on different devices based on the optimization goal. The evaluation result shows HeteroPDP can improve the system fairness from 24% to 65% when an OpenCL application is co-located with other processes, and gain an additional 50% speedup compared with always offloading the OpenCL workload to GPUs.

In summary, this dissertation aims to provide insights for the future microarchitecture and system architecture designs by identifying, analyzing, and addressing three critical performance problems in modern GPUs.

Memory systems are becoming increasingly error-prone, and thus guaranteeing their reliability is a major challenge. In this dissertation, new techniques to improve the reliability of both 2D and 3D dynamic random access memory (DRAM) systems are presented. The proposed schemes have higher reliability than current systems but with lower power, better performance and lower hardware cost.

First, a low overhead solution that improves the reliability of commodity DRAM systems with no change in the existing memory architecture is presented. Specifically, five erasure and error correction (E-ECC) schemes are proposed that provide at least Chipkill-Correct protection for x4 (Schemes 1, 2 and 3), x8 (Scheme 4) and x16 (Scheme 5) DRAM systems. All schemes have superior error correction performance due to the use of strong symbol-based codes. In addition, the use of erasure codes extends the lifetime of the 2D DRAM systems.

Next, two error correction schemes are presented for 3D DRAM memory systems. The first scheme is a rate-adaptive, two-tiered error correction scheme (RATT-ECC) that provides strong reliability (10^10x) reduction in raw FIT rate) for an HBM-like 3D DRAM system that services CPU applications. The rate-adaptive feature of RATT-ECC enables permanent bank failures to be handled through sparing. It can also be used to significantly reduce the refresh power consumption without decreasing the reliability and timing performance.

The second scheme is a two-tiered error correction scheme (Config-ECC) that supports different sized accesses in GPU applications with strong reliability. It addresses the mismatch between data access size and fixed sized ECC scheme by designing a product code based flexible scheme. Config-ECC is built around a core unit designed for 32B access with a simple extension to support 64B and 128B accesses. Compared to fixed 32B and 64B ECC schemes, Config-ECC reduces the failure in time (FIT) rate by 200x and 20x, respectively. It also reduces the memory energy by 17% (in the dynamic mode) and 21% (in the static mode) compared to a state-of-the-art fixed 64B ECC scheme.

Cyber-physical systems and hard real-time systems have strict timing constraints that specify deadlines until which tasks must finish their execution. Missing a deadline can cause unexpected outcome or endanger human lives in safety-critical applications, such as automotive or aeronautical systems. It is, therefore, of utmost importance to obtain and optimize a safe upper bound of each task’s execution time or the worst-case execution time (WCET), to guarantee the absence of any missed deadline. Unfortunately, conventional microarchitectural components, such as caches and branch predictors, are only optimized for average-case performance and often make WCET analysis complicated and pessimistic. Caches especially have a large impact on the worst-case performance due to expensive off- chip memory accesses involved in cache miss handling. In this regard, software-controlled scratchpad memories (SPMs) have become a promising alternative to caches. An SPM is a raw SRAM, controlled only by executing data movement instructions explicitly at runtime, and such explicit control facilitates static analyses to obtain safe and tight upper bounds of WCETs. SPM management techniques, used in compilers targeting an SPM-based processor, determine how to use a given SPM space by deciding where to insert data movement instructions and what operations to perform at those program locations. This dissertation presents several management techniques for program code and stack data, which aim to optimize the WCETs of a given program. The proposed code management techniques include optimal allocation algorithms and a polynomial-time heuristic for allocating functions to the SPM space, with or without the use of abstraction of SPM regions, and a heuristic for splitting functions into smaller partitions. The proposed stack data management technique, on the other hand, finds an optimal set of program locations to evict and restore stack frames to avoid stack overflows, when the call stack resides in a size-limited SPM. In the evaluation, the WCETs of various benchmarks including real-world automotive applications are statically calculated for SPMs and caches in several different memory configurations.

The availability of a wide range of general purpose as well as accelerator cores on

modern smartphones means that a significant number of applications can be executed

on a smartphone simultaneously, resulting in an ever increasing demand on the memory

subsystem. While the increased computation capability is intended for improving

user experience, memory requests from each concurrent application exhibit unique

memory access patterns as well as specific timing constraints. If not considered, this

could lead to significant memory contention and result in lowered user experience.

This work first analyzes the impact of memory degradation caused by the interference

at the memory system for a broad range of commonly-used smartphone applications.

The real system characterization results show that smartphone applications,

such as web browsing and media playback, suffer significant performance degradation.

This is caused by shared resource contention at the application processor’s last-level

cache, the communication fabric, and the main memory.

Based on the detailed characterization results, rest of this thesis focuses on the

design of an effective memory interference mitigation technique. Since web browsing,

being one of the most commonly-used smartphone applications and represents many

html-based smartphone applications, my thesis focuses on meeting the performance

requirement of a web browser on a smartphone in the presence of background processes

and co-scheduled applications. My thesis proposes a light-weight user space frequency

governor to mitigate the degradation caused by interfering applications, by predicting

the performance and power consumption of web browsing. The governor selects an

optimal energy-efficient frequency setting periodically by using the statically-trained

performance and power models with dynamically-varying architecture and system

conditions, such as the memory access intensity of background processes and/or coscheduled applications, and temperature of cores. The governor has been extensively evaluated on a Nexus 5 smartphone over a diverse range of mobile workloads. By

operating at the most energy-efficient frequency setting in the presence of interference,

energy efficiency is improved by as much as 35% and with an average of 18% compared

to the existing interactive governor, while maintaining the satisfactory performance

of web page loading under 3 seconds.

Several decades of transistor technology scaling has brought the threat of soft errors to modern embedded processors. Several techniques have been proposed to protect these systems from soft errors. However, their effectiveness in protecting the computation cannot be ascertained without accurate and quantitative estimation of system reliability. Vulnerability -- a metric that defines the probability of system-failure (reliability) through analytical models -- is the most effective mechanism for our current estimation and early design space exploration needs. Previous vulnerability estimation tools are based around the Sim-Alpha simulator which has been to shown to have several limitations. In this thesis, I present gemV: an accurate and comprehensive vulnerability estimation tool based on gem5. Gem5 is a popular cycle-accurate micro-architectural simulator that can model several different processor models in close to real hardware form. GemV can be used for fast and early design space exploration and also evaluate the protection afforded by commodity processors. gemV is comprehensive, since it models almost all sequential components of the processor. gemV is accurate because of fine-grain vulnerability tracking, accurate vulnerability modeling of squashed instructions, and accurate vulnerability modeling of shared data structures in gem5. gemV has been thoroughly validated against extensive fault injection experiments and achieves a 97\% accuracy with 95\% confidence. A micro-architect can use gemV to discover micro-architectural variants of a processor that minimize vulnerability for allowed performance penalty. A software developer can use gemV to explore the performance-vulnerability trade-off by choosing different algorithms and compiler optimizations, while the system designer can use gemV to explore the performance-vulnerability trade-offs of choosing different Insruction Set Architectures (ISA).

As persistent non-volatile memory solutions become integrated in the computing ecosystem and landscape, traditional commodity file systems architected and developed for traditional block I/O based memory solutions must be reevaluated. A majority of commodity file systems have been architected and designed with the goal of managing data on non-volatile storage devices such as hard disk drives (HDDs) and solid state drives (SSDs). HDDs and SSDs are attached to a computing system via a controller or I/O hub, often referred to as the southbridge. The point of HDD and SSD attachment creates multiple levels of translation for any data managed by the CPU that must be stored in non-volatile memory (NVM) on an HDD or SSD. Storage Class Memory (SCM) devices provide the ability to store data at the CPU and DRAM level of a computing system. A novel set of modifications to the ext2 and ext4 commodity file systems to address the needs of SCM will be presented and discussed. An in-depth analysis of many existing file systems, from multiple sources, will be presented along with an analysis to identify key modifications and extensions that would be necessary to execute file system on SCM devices. From this analysis, modifications and extensions have been applied to the FAT commodity file system for key functional tests that will be presented to demonstrate the operation and execution of the file system extensions.

Widespread adoption of smartphone based Mobile Medical Apps (MMAs) is opening new avenues for innovation, bringing MMAs to the forefront of low cost healthcare delivery. These apps often control human physiology and work on sensitive data. Thus it is necessary to have evidences of their trustworthiness i.e. maintaining privacy of health data, long term operation of wearable sensors and ensuring no harm to the user before actual marketing. Traditionally, clinical studies are used to validate the trustworthiness of medical systems. However, they can take long time and could potentially harm the user. Such evidences can be generated using simulations and mathematical analysis. These methods involve estimating the MMA interactions with human physiology. However, the nonlinear nature of human physiology makes the estimation challenging.

This research analyzes and develops MMA software while considering its interactions with human physiology to assure trustworthiness. A novel app development methodology is used to objectively evaluate trustworthiness of a MMA by generating evidences using automatic techniques. It involves developing the Health-Dev β tool to generate a) evidences of trustworthiness of MMAs and b) requirements assured code generation for vulnerable components of the MMA without hindering the app development process. In this method, all requests from MMAs pass through a trustworthy entity, Trustworthy Data Manager which checks if the app request satisfies the MMA requirements. This method is intended to expedite the design to marketing process of MMAs. The objectives of this research is to develop models, tools and theory for evidence generation and can be divided into the following themes:

• Sustainable design configuration estimation of MMAs: Developing an optimization framework which can generate sustainable and safe sensor configuration while considering interactions of the MMA with the environment.

• Evidence generation using simulation and formal methods: Developing models and tools to verify safety properties of the MMA design to ensure no harm to the human physiology.

• Automatic code generation for MMAs: Investigating methods for automatically

• Performance analysis of trustworthy data manager: Evaluating response time generating trustworthy software for vulnerable components of a MMA and evidences.performance of trustworthy data manager under interactions from non-MMA smartphone apps.

Virtual machines and containers have steadily improved their performance over time as a result of innovations in their architecture and software ecosystems. Network functions and workloads are increasingly migrating to virtual environments, supported by developments in software defined networking (SDN) and network function virtualization (NFV). Previous performance analyses of virtual systems in this context often ignore significant performance gains that can be acheived with practical modifications to hypervisor and host systems. In this thesis, the network performance of containers and virtual machines are measured with standard network performance tools. The performance of these systems utilizing a standard 3.18.20 Linux kernel is compared to that of a realtime-tuned variant of the same kernel. This thesis motivates improving determinism in virtual systems with modifications to host and guest kernels and thoughtful process isolation. With the system modifications described, the median TCP bandwidth of KVM virtual machines over bridged network interfaces, is increased by 10.8% with a corresponding reduction in standard deviation of 87.6%. Docker containers see a 8.8% improvement in median bandwidth and 4.4% reduction in standard deviation of TCP measurements using similar bridged networking. System tuning also reduces the standard deviation of TCP request/response latency (TCP RR) over bridged interfaces by 86.8% for virtual machines and 97.9% for containers. Hardware devices assigned to virtual systems also see reductions in variance, although not as noteworthy.

The holy grail of computer hardware across all market segments has been to sustain performance improvement at the same pace as silicon technology scales. As the technology scales and the size of transistors shrinks, the power consumption and energy usage per transistor decrease. On the other hand, the transistor density increases significantly by technology scaling. Due to technology factors, the reduction in power consumption per transistor is not sufficient to offset the increase in power consumption per unit area. Therefore, to improve performance, increasing energy-efficiency must be addressed at all design levels from circuit level to application and algorithm levels.

At architectural level, one promising approach is to populate the system with hardware accelerators each optimized for a specific task. One drawback of hardware accelerators is that they are not programmable. Therefore, their utilization can be low as they perform one specific function. Using software programmable accelerators is an alternative approach to achieve high energy-efficiency and programmability. Due to intrinsic characteristics of software accelerators, they can exploit both instruction level parallelism and data level parallelism.

Coarse-Grained Reconfigurable Architecture (CGRA) is a software programmable accelerator consists of a number of word-level functional units. Motivated by promising characteristics of software programmable accelerators, the potentials of CGRAs in future computing platforms is studied and an end-to-end CGRA research framework is developed. This framework consists of three different aspects: CGRA architectural design, integration in a computing system, and CGRA compiler. First, the design and implementation of a CGRA and its instruction set is presented. This design is then modeled in a cycle accurate system simulator. The simulation platform enables us to investigate several problems associated with a CGRA when it is deployed as an accelerator in a computing system. Next, the problem of mapping a compute intensive region of a program to CGRAs is formulated. From this formulation, several efficient algorithms are developed which effectively utilize CGRA scarce resources very well to minimize the running time of input applications. Finally, these mapping algorithms are integrated in a compiler framework to construct a compiler for CGRA