Matching Items (12)
Filtering by
- All Subjects: Computer architecture
- Creators: Shrivastava, Aviral
- Member of: Theses and Dissertations
Description
In recent years we have witnessed a shift towards multi-processor system-on-chips (MPSoCs) to address the demands of embedded devices (such as cell phones, GPS devices, luxury car features, etc.). Highly optimized MPSoCs are well-suited to tackle the complex application demands desired by the end user customer. These MPSoCs incorporate a constellation of heterogeneous processing elements (PEs) (general purpose PEs and application-specific integrated circuits (ASICS)). A typical MPSoC will be composed of a application processor, such as an ARM Coretex-A9 with cache coherent memory hierarchy, and several application sub-systems. Each of these sub-systems are composed of highly optimized instruction processors, graphics/DSP processors, and custom hardware accelerators. Typically, these sub-systems utilize scratchpad memories (SPM) rather than support cache coherency. The overall architecture is an integration of the various sub-systems through a high bandwidth system-level interconnect (such as a Network-on-Chip (NoC)). The shift to MPSoCs has been fueled by three major factors: demand for high performance, the use of component libraries, and short design turn around time. As customers continue to desire more and more complex applications on their embedded devices the performance demand for these devices continues to increase. Designers have turned to using MPSoCs to address this demand. By using pre-made IP libraries designers can quickly piece together a MPSoC that will meet the application demands of the end user with minimal time spent designing new hardware. Additionally, the use of MPSoCs allows designers to generate new devices very quickly and thus reducing the time to market. In this work, a complete MPSoC synthesis design flow is presented. We first present a technique \cite{leary1_intro} to address the synthesis of the interconnect architecture (particularly Network-on-Chip (NoC)). We then address the synthesis of the memory architecture of a MPSoC sub-system \cite{leary2_intro}. Lastly, we present a co-synthesis technique to generate the functional and memory architectures simultaneously. The validity and quality of each synthesis technique is demonstrated through extensive experimentation.
ContributorsLeary, Glenn (Author) / Chatha, Karamvir S (Thesis advisor) / Vrudhula, Sarma (Committee member) / Shrivastava, Aviral (Committee member) / Beraha, Rudy (Committee member) / Arizona State University (Publisher)
Created2013
Description
Coarse-Grained Reconfigurable Architectures (CGRA) are a promising fabric for improving the performance and power-efficiency of computing devices. CGRAs are composed of components that are well-optimized to execute loops and rotating register file is an example of such a component present in CGRAs. Due to the rotating nature of register indexes in rotating register file, it is very challenging, if at all possible, to hold and properly index memory addresses (pointers) and static values. In this Thesis, different structures for CGRA register files are investigated. Those structures are experimentally compared in terms of performance of mapped applications, design frequency, and area. It is shown that a register file that can logically be partitioned into rotating and non-rotating regions is an excellent choice because it imposes the minimum restriction on underlying CGRA mapping algorithm while resulting in efficient resource utilization.
ContributorsSaluja, Dipal (Author) / Shrivastava, Aviral (Thesis advisor) / Lee, Yann-Hang (Committee member) / Wu, Carole-Jean (Committee member) / Arizona State University (Publisher)
Created2014
Description
Stream processing has emerged as an important model of computation especially in the context of multimedia and communication sub-systems of embedded System-on-Chip (SoC) architectures. The dataflow nature of streaming applications allows them to be most naturally expressed as a set of kernels iteratively operating on continuous streams of data. The kernels are computationally intensive and are mainly characterized by real-time constraints that demand high throughput and data bandwidth with limited global data reuse. Conventional architectures fail to meet these demands due to their poorly matched execution models and the overheads associated with instruction and data movements.
This work presents StreamWorks, a multi-core embedded architecture for energy-efficient stream computing. The basic processing element in the StreamWorks architecture is the StreamEngine (SE) which is responsible for iteratively executing a stream kernel. SE introduces an instruction locking mechanism that exploits the iterative nature of the kernels and enables fine-grain instruction reuse. Each instruction in a SE is locked to a Reservation Station (RS) and revitalizes itself after execution; thus never retiring from the RS. The entire kernel is hosted in RS Banks (RSBs) close to functional units for energy-efficient instruction delivery. The dataflow semantics of stream kernels are captured by a context-aware dataflow execution mode that efficiently exploits the Instruction Level Parallelism (ILP) and Data-level parallelism (DLP) within stream kernels.
Multiple SEs are grouped together to form a StreamCluster (SC) that communicate via a local interconnect. A novel software FIFO virtualization technique with split-join functionality is proposed for efficient and scalable stream communication across SEs. The proposed communication mechanism exploits the Task-level parallelism (TLP) of the stream application. The performance and scalability of the communication mechanism is evaluated against the existing data movement schemes for scratchpad based multi-core architectures. Further, overlay schemes and architectural support are proposed that allow hosting any number of kernels on the StreamWorks architecture. The proposed oevrlay schemes for code management supports kernel(context) switching for the most common use cases and can be adapted for any multi-core architecture that use software managed local memories.
The performance and energy-efficiency of the StreamWorks architecture is evaluated for stream kernel and application benchmarks by implementing the architecture in 45nm TSMC and comparison with a low power RISC core and a contemporary accelerator.
This work presents StreamWorks, a multi-core embedded architecture for energy-efficient stream computing. The basic processing element in the StreamWorks architecture is the StreamEngine (SE) which is responsible for iteratively executing a stream kernel. SE introduces an instruction locking mechanism that exploits the iterative nature of the kernels and enables fine-grain instruction reuse. Each instruction in a SE is locked to a Reservation Station (RS) and revitalizes itself after execution; thus never retiring from the RS. The entire kernel is hosted in RS Banks (RSBs) close to functional units for energy-efficient instruction delivery. The dataflow semantics of stream kernels are captured by a context-aware dataflow execution mode that efficiently exploits the Instruction Level Parallelism (ILP) and Data-level parallelism (DLP) within stream kernels.
Multiple SEs are grouped together to form a StreamCluster (SC) that communicate via a local interconnect. A novel software FIFO virtualization technique with split-join functionality is proposed for efficient and scalable stream communication across SEs. The proposed communication mechanism exploits the Task-level parallelism (TLP) of the stream application. The performance and scalability of the communication mechanism is evaluated against the existing data movement schemes for scratchpad based multi-core architectures. Further, overlay schemes and architectural support are proposed that allow hosting any number of kernels on the StreamWorks architecture. The proposed oevrlay schemes for code management supports kernel(context) switching for the most common use cases and can be adapted for any multi-core architecture that use software managed local memories.
The performance and energy-efficiency of the StreamWorks architecture is evaluated for stream kernel and application benchmarks by implementing the architecture in 45nm TSMC and comparison with a low power RISC core and a contemporary accelerator.
ContributorsPanda, Amrit (Author) / Chatha, Karam S. (Thesis advisor) / Wu, Carole-Jean (Thesis advisor) / Chakrabarti, Chaitali (Committee member) / Shrivastava, Aviral (Committee member) / Arizona State University (Publisher)
Created2014
Description
Coarse Grain Reconfigurable Arrays (CGRAs) are promising accelerators capable of
achieving high performance at low power consumption. While CGRAs can efficiently
accelerate loop kernels, accelerating loops with control flow (loops with if-then-else
structures) is quite challenging. Techniques that handle control flow execution in
CGRAs generally use predication. Such techniques execute both branches of an
if-then-else structure and select outcome of either branch to commit based on the
result of the conditional. This results in poor utilization of CGRA s computational
resources. Dual-issue scheme which is the state of the art technique for control flow
fetches instructions from both paths of the branch and selects one to execute at
runtime based on the result of the conditional. This technique has an overhead in
instruction fetch bandwidth. In this thesis, to improve performance of control flow
execution in CGRAs, I propose a solution in which the result of the conditional
expression that decides the branch outcome is communicated to the instruction fetch
unit to selectively issue instructions from the path taken by the branch at run time.
Experimental results show that my solution can achieve 34.6% better performance
and 52.1% improvement in energy efficiency on an average compared to state of the
art dual issue scheme without imposing any overhead in instruction fetch bandwidth.
achieving high performance at low power consumption. While CGRAs can efficiently
accelerate loop kernels, accelerating loops with control flow (loops with if-then-else
structures) is quite challenging. Techniques that handle control flow execution in
CGRAs generally use predication. Such techniques execute both branches of an
if-then-else structure and select outcome of either branch to commit based on the
result of the conditional. This results in poor utilization of CGRA s computational
resources. Dual-issue scheme which is the state of the art technique for control flow
fetches instructions from both paths of the branch and selects one to execute at
runtime based on the result of the conditional. This technique has an overhead in
instruction fetch bandwidth. In this thesis, to improve performance of control flow
execution in CGRAs, I propose a solution in which the result of the conditional
expression that decides the branch outcome is communicated to the instruction fetch
unit to selectively issue instructions from the path taken by the branch at run time.
Experimental results show that my solution can achieve 34.6% better performance
and 52.1% improvement in energy efficiency on an average compared to state of the
art dual issue scheme without imposing any overhead in instruction fetch bandwidth.
ContributorsRajendran Radhika, Shri Hari (Author) / Shrivastava, Aviral (Thesis advisor) / Christen, Jennifer Blain (Committee member) / Cao, Yu (Committee member) / Arizona State University (Publisher)
Created2014
Description
Limited Local Memory (LLM) multicore architectures are promising powerefficient architectures will scalable memory hierarchy. In LLM multicores, each core can access only a small local memory. Accesses to a large shared global memory can only be made explicitly through Direct Memory Access (DMA) operations. Standard Template Library (STL) is a powerful programming tool and is widely used for software development. STLs provide dynamic data structures, algorithms, and iterators for vector, deque (double-ended queue), list, map (red-black tree), etc. Since the size of the local memory is limited in the cores of the LLM architecture, and data transfer is not automatically supported by hardware cache or OS, the usage of current STL implementation on LLM multicores is limited. Specifically, there is a hard limitation on the amount of data they can handle. In this article, we propose and implement a framework which manages the STL container classes on the local memory of LLM multicore architecture. Our proposal removes the data size limitation of the STL, and therefore improves the programmability on LLM multicore architectures with little change to the original program. Our implementation results in only about 12%-17% increase in static library code size and reasonable runtime overheads.
ContributorsLu, Di (Author) / Shrivastava, Aviral (Thesis advisor) / Chatha, Karamvir (Committee member) / Dasgupta, Partha (Committee member) / Arizona State University (Publisher)
Created2012
Description
Thanks to continuous technology scaling, intelligent, fast and smaller digital systems are now available at affordable costs. As a result, digital systems have found use in a wide range of application areas that were not even imagined before, including medical (e.g., MRI, remote or post-operative monitoring devices, etc.), automotive (e.g., adaptive cruise control, anti-lock brakes, etc.), security systems (e.g., residential security gateways, surveillance devices, etc.), and in- and out-of-body sensing (e.g., capsule swallowed by patients measuring digestive system pH, heart monitors, etc.). Such computing systems, which are completely embedded within the application, are called embedded systems, as opposed to general purpose computing systems. In the design of such embedded systems, power consumption and reliability are indispensable system requirements. In battery operated portable devices, the battery is the single largest factor contributing to device cost, weight, recharging time, frequency and ultimately its usability. For example, in the Apple iPhone 4 smart-phone, the battery is $40\%$ of the device weight, occupies $36\%$ of its volume and allows only $7$ hours (over 3G) of talk time. As embedded systems find use in a range of sensitive applications, from bio-medical applications to safety and security systems, the reliability of the computations performed becomes a crucial factor. At our current technology-node, portable embedded systems are prone to expect failures due to soft errors at the rate of once-per-year; but with aggressive technology scaling, the rate is predicted to increase exponentially to once-per-hour. Over the years, researchers have been successful in developing techniques, implemented at different layers of the design-spectrum, to improve system power efficiency and reliability. Among the layers of design abstraction, I observe that the interface between the compiler and processor micro-architecture possesses a unique potential for efficient design optimizations. A compiler designer is able to observe and analyze the application software at a finer granularity; while the processor architect analyzes the system output (power, performance, etc.) for each executed instruction. At the compiler micro-architecture interface, if the system knowledge at the two design layers can be integrated, design optimizations at the two layers can be modified to efficiently utilize available resources and thereby achieve appreciable system-level benefits. To this effect, the thesis statement is that, ``by merging system design information at the compiler and micro-architecture design layers, smart compilers can be developed, that achieve reliable and power-efficient embedded computing through: i) Pure compiler techniques, ii) Hybrid compiler micro-architecture techniques, and iii) Compiler-aware architectures''. In this dissertation demonstrates, through contributions in each of the three compiler-based techniques, the effectiveness of smart compilers in achieving power-efficiency and reliability in embedded systems.
ContributorsJeyapaul, Reiley (Author) / Shrivastava, Aviral (Thesis advisor) / Vrudhula, Sarma (Committee member) / Clark, Lawrence (Committee member) / Colbourn, Charles (Committee member) / Arizona State University (Publisher)
Created2012
Description
Performance improvements have largely followed Moore's Law due to the help from technology scaling. In order to continue improving performance, power-efficiency must be reduced. Better technology has improved power-efficiency, but this has a limit. Multi-core architectures have been shown to be an additional aid to this crusade of increased power-efficiency. Accelerators are growing in popularity as the next means of achieving power-efficient performance. Accelerators such as Intel SSE are ideal, but prove difficult to program. FPGAs, on the other hand, are less efficient due to their fine-grained reconfigurability. A middle ground is found in CGRAs, which are highly power-efficient, but largely programmable accelerators. Power-efficiencies of 100s of GOPs/W have been estimated, more than 2 orders of magnitude greater than current processors. Currently, CGRAs are limited in their applicability due to their ability to only accelerate a single thread at a time. This limitation becomes especially apparent as multi-core/multi-threaded processors have moved into the mainstream. This limitation is removed by enabling multi-threading on CGRAs through a software-oriented approach. The key capability in this solution is enabling quick run-time transformation of schedules to execute on targeted portions of the CGRA. This allows the CGRA to be shared among multiple threads simultaneously. Analysis shows that enabling multi-threading has very small costs but provides very large benefits (less than 1% single-threaded performance loss but nearly 300% CGRA throughput increase). By increasing dynamism of CGRA scheduling, system performance is shown to increase overall system performance of an optimized system by almost 350% over that of a single-threaded CGRA and nearly 20x faster than the same system with no CGRA in a highly threaded environment.
ContributorsPager, Jared (Author) / Shrivastava, Aviral (Thesis advisor) / Gupta, Sandeep (Committee member) / Speyer, Gil (Committee member) / Arizona State University (Publisher)
Created2011
Description
The quality and quantity of talented members of the US STEM workforce has
been a subject of great interest to policy and decision makers for the past 40 years.
Recent research indicates that while there exist specific shortages in specific disciplines
and areas of expertise in the private sector and the federal government, there is no
noticeable shortage in any STEM academic discipline, but rather a surplus of PhDs
vying for increasingly scarce tenure track positions. Despite the seeming availability
of industry and private sector jobs, recent PhDs still struggle to find employment in
those areas. I argue that the decades old narrative suggesting a shortage of STEM
PhDs in the US poses a threat to the value of the natural science PhD, and that
this narrative contributes significantly to why so many PhDs struggle to find career
employment in their fields. This study aims to address the following question: what is
the value of a STEM PhD outside academia? I begin with a critical review of existing
literature, and then analyze programmatic documents for STEM PhD programs at
ASU, interviews with industry employers, and an examination the public face of value
for these degrees. I then uncover the nature of the value alignment, value disconnect,
and value erosion in the ecosystem which produces and then employs STEM PhDs,
concluding with specific areas which merit special consideration in an effort to increase
the value of these degrees for all stakeholders involved.
been a subject of great interest to policy and decision makers for the past 40 years.
Recent research indicates that while there exist specific shortages in specific disciplines
and areas of expertise in the private sector and the federal government, there is no
noticeable shortage in any STEM academic discipline, but rather a surplus of PhDs
vying for increasingly scarce tenure track positions. Despite the seeming availability
of industry and private sector jobs, recent PhDs still struggle to find employment in
those areas. I argue that the decades old narrative suggesting a shortage of STEM
PhDs in the US poses a threat to the value of the natural science PhD, and that
this narrative contributes significantly to why so many PhDs struggle to find career
employment in their fields. This study aims to address the following question: what is
the value of a STEM PhD outside academia? I begin with a critical review of existing
literature, and then analyze programmatic documents for STEM PhD programs at
ASU, interviews with industry employers, and an examination the public face of value
for these degrees. I then uncover the nature of the value alignment, value disconnect,
and value erosion in the ecosystem which produces and then employs STEM PhDs,
concluding with specific areas which merit special consideration in an effort to increase
the value of these degrees for all stakeholders involved.
ContributorsGarbee, Elizabeth (Author) / Maynard, Andrew D. (Thesis advisor) / Wetmore, Jameson (Committee member) / Anderson, Derrick (Committee member) / Arizona State University (Publisher)
Created2018
Description
General-purpose processors propel the advances and innovations that are the subject of humanity’s many endeavors. Catering to this demand, chip-multiprocessors (CMPs) and general-purpose graphics processing units (GPGPUs) have seen many high-performance innovations in their architectures. With these advances, the memory subsystem has become the performance- and energy-limiting aspect of CMPs and GPGPUs alike. This dissertation identifies and mitigates the key performance and energy-efficiency bottlenecks in the memory subsystem of general-purpose processors via novel, practical, microarchitecture and system-architecture solutions.
Addressing the important Last Level Cache (LLC) management problem in CMPs, I observe that LLC management decisions made in isolation, as in prior proposals, often lead to sub-optimal system performance. I demonstrate that in order to maximize system performance, it is essential to manage the LLCs while being cognizant of its interaction with the system main memory. I propose ReMAP, which reduces the net memory access cost by evicting cache lines that either have no reuse, or have low memory access cost. ReMAP improves the performance of the CMP system by as much as 13%, and by an average of 6.5%.
Rather than the LLC, the L1 data cache has a pronounced impact on GPGPU performance by acting as the bandwidth filter for the rest of the memory subsystem. Prior work has shown that the severely constrained data cache capacity in GPGPUs leads to sub-optimal performance. In this thesis, I propose two novel techniques that address the GPGPU data cache capacity problem. I propose ID-Cache that performs effective cache bypassing and cache line size selection to improve cache capacity utilization. Next, I propose LATTE-CC that considers the GPU’s latency tolerance feature and adaptively compresses the data stored in the data cache, thereby increasing its effective capacity. ID-Cache and LATTE-CC are shown to achieve 71% and 19.2% speedup, respectively, over a wide variety of GPGPU applications.
Complementing the aforementioned microarchitecture techniques, I identify the need for system architecture innovations to sustain performance scalability of GPG- PUs in the face of slowing Moore’s Law. I propose a novel GPU architecture called the Multi-Chip-Module GPU (MCM-GPU) that integrates multiple GPU modules to form a single logical GPU. With intelligent memory subsystem optimizations tailored for MCM-GPUs, it can achieve within 7% of the performance of a similar but hypothetical monolithic die GPU. Taking a step further, I present an in-depth study of the energy-efficiency characteristics of future MCM-GPUs. I demonstrate that the inherent non-uniform memory access side-effects form the key energy-efficiency bottleneck in the future.
In summary, this thesis offers key insights into the performance and energy-efficiency bottlenecks in CMPs and GPGPUs, which can guide future architects towards developing high-performance and energy-efficient general-purpose processors.
Addressing the important Last Level Cache (LLC) management problem in CMPs, I observe that LLC management decisions made in isolation, as in prior proposals, often lead to sub-optimal system performance. I demonstrate that in order to maximize system performance, it is essential to manage the LLCs while being cognizant of its interaction with the system main memory. I propose ReMAP, which reduces the net memory access cost by evicting cache lines that either have no reuse, or have low memory access cost. ReMAP improves the performance of the CMP system by as much as 13%, and by an average of 6.5%.
Rather than the LLC, the L1 data cache has a pronounced impact on GPGPU performance by acting as the bandwidth filter for the rest of the memory subsystem. Prior work has shown that the severely constrained data cache capacity in GPGPUs leads to sub-optimal performance. In this thesis, I propose two novel techniques that address the GPGPU data cache capacity problem. I propose ID-Cache that performs effective cache bypassing and cache line size selection to improve cache capacity utilization. Next, I propose LATTE-CC that considers the GPU’s latency tolerance feature and adaptively compresses the data stored in the data cache, thereby increasing its effective capacity. ID-Cache and LATTE-CC are shown to achieve 71% and 19.2% speedup, respectively, over a wide variety of GPGPU applications.
Complementing the aforementioned microarchitecture techniques, I identify the need for system architecture innovations to sustain performance scalability of GPG- PUs in the face of slowing Moore’s Law. I propose a novel GPU architecture called the Multi-Chip-Module GPU (MCM-GPU) that integrates multiple GPU modules to form a single logical GPU. With intelligent memory subsystem optimizations tailored for MCM-GPUs, it can achieve within 7% of the performance of a similar but hypothetical monolithic die GPU. Taking a step further, I present an in-depth study of the energy-efficiency characteristics of future MCM-GPUs. I demonstrate that the inherent non-uniform memory access side-effects form the key energy-efficiency bottleneck in the future.
In summary, this thesis offers key insights into the performance and energy-efficiency bottlenecks in CMPs and GPGPUs, which can guide future architects towards developing high-performance and energy-efficient general-purpose processors.
ContributorsArunkumar, Akhil (Author) / Wu, Carole-Jean (Thesis advisor) / Shrivastava, Aviral (Committee member) / Lee, Yann-Hang (Committee member) / Bolotin, Evgeny (Committee member) / Arizona State University (Publisher)
Created2018
Description
Coarse-grained Reconfigurable Arrays (CGRAs) are promising accelerators capable
of accelerating even non-parallel loops and loops with low trip-counts. One challenge
in compiling for CGRAs is to manage both recurring and nonrecurring variables in
the register file (RF) of the CGRA. Although prior works have managed recurring
variables via rotating RF, they access the nonrecurring variables through either a
global RF or from a constant memory. The former does not scale well, and the latter
degrades the mapping quality. This work proposes a hardware-software codesign
approach in order to manage all the variables in a local nonrotating RF. Hardware
provides modulo addition based indexing mechanism to enable correct addressing
of recurring variables in a nonrotating RF. The compiler determines the number of
registers required for each recurring variable and configures the boundary between the
registers used for recurring and nonrecurring variables. The compiler also pre-loads
the read-only variables and constants into the local registers in the prologue of the
schedule. Synthesis and place-and-route results of the previous and the proposed RF
design show that proposed solution achieves 17% better cycle time. Experiments of
mapping several important and performance-critical loops collected from MiBench
show proposed approach improves performance (through better mapping) by 18%,
compared to using constant memory.
of accelerating even non-parallel loops and loops with low trip-counts. One challenge
in compiling for CGRAs is to manage both recurring and nonrecurring variables in
the register file (RF) of the CGRA. Although prior works have managed recurring
variables via rotating RF, they access the nonrecurring variables through either a
global RF or from a constant memory. The former does not scale well, and the latter
degrades the mapping quality. This work proposes a hardware-software codesign
approach in order to manage all the variables in a local nonrotating RF. Hardware
provides modulo addition based indexing mechanism to enable correct addressing
of recurring variables in a nonrotating RF. The compiler determines the number of
registers required for each recurring variable and configures the boundary between the
registers used for recurring and nonrecurring variables. The compiler also pre-loads
the read-only variables and constants into the local registers in the prologue of the
schedule. Synthesis and place-and-route results of the previous and the proposed RF
design show that proposed solution achieves 17% better cycle time. Experiments of
mapping several important and performance-critical loops collected from MiBench
show proposed approach improves performance (through better mapping) by 18%,
compared to using constant memory.
ContributorsDave, Shail (Author) / Shrivastava, Aviral (Thesis advisor) / Ren, Fengbo (Committee member) / Ogras, Umit Y. (Committee member) / Arizona State University (Publisher)
Created2016