In recent years, we are witnessing a trend toward in-memory computing for future generations of computers that differs from traditional von-Neumann architecture in which there is a clear distinction between computing and memory units. Considering that data movements between the central processing unit (CPU) and memory consume several orders of magnitude more energy compared to simple arithmetic operations in the CPU, in-memory computing will lead to huge energy savings as data no longer needs to be moved around between these units. In an initial step toward this goal, new non-volatile memory technologies, e.g., resistive RAM (ReRAM) and phase-change memory (PCM), are being explored. This has led to a large body of research that mainly focuses on the design of the memory array and its peripheral circuitry. In this article, we mainly focus on the tile architecture (comprising a memory array and peripheral circuitry) in which storage and compute operations are performed in the (analog) memory array and the results are produced in the (digital) periphery. Such an architecture is termed compute-in-memory-periphery (CIM-P). More precisely, we derive an abstract CIM-tile architecture and define its main building blocks. To bridge the gap between higher-level programming languages and the underlying (analog) circuit designs, an instruction-set architecture is defined that is intended to control and, in turn, sequence the operations within this CIM tile to perform higher-level more complex operations. Moreover, we define a procedure to pipeline the CIM-tile operations to further improve the performance. To simulate the tile and perform design space exploration considering different technologies and parameters, we introduce the fully parameterized first-of-its-kind CIM tile simulator and compiler. Furthermore, the compiler is technology-aware when scheduling the CIM-tile instructions. Finally, using the simulator, we perform several preliminary design space explorations regarding the three competing technologies, ReRAM, PCM, and STT-MRAM concerning CIM-tile parameters, e.g., the number of ADCs. Additionally, we investigate the effect of pipelining in relation to the clock speeds of the digital periphery assuming the three technologies. In the end, we demonstrate that our simulator is also capable of reporting energy consumption for each building block within the CIM tile after the execution of in-memory kernels considering the data-dependency on the energy consumption of the memory array. All the source codes are publicly available.

@en

Emerging computation-in-memory (CIM) paradigm offers processing and storage of data at the same physical location, thus alleviating critical memory-processor communication bottlenecks suffered by conventional von-Neumann architecture. Storage of data in a CIM architecture is analog in nature and therefore computation is performed in analog domain i.e. inputs and outputs are analog values. Since the outside computing environment is digital, analog-to-digital converters (ADC) are utilized to perform the output data conversion. However, ADC designs are bulky, power-hungry circuits that are prone to design variations and therefore, play an important role in determining the computing efficiency of CIM architectures. In this paper, we present a scalable and reliable integrate and fire circuit ADC (SRIF-ADC) design for CIM architectures, suitable for stringent power and area constraints. We devise a technique to stabilize the node receiving analog inputs that allows more rows to be activated at the same time, thereby increasing the operand size of input vectors. This allows better scalability in terms of higher parallelism of operations. We employ a self-timed variation-aware design approach and design measures to drastically reduce read disturb of memristor devices that address reliability issues related to the ADC design. In addition, we present a compact, built-in sample-and-hold circuit to replace the large-sized capacitance and built-in weighting technique to alleviate the need for post-processing. For multiply-and-accumulate (MAC) operation, our simulation results show that we can improve the computational parallelism by 3X as well as ADC conversion speed and energy efficiency are improved by 2X and 11.6X, respectively, compared to the state-of-the-art design.

@en

A novel type of hardware accelerators called automata processors (APs) have been proposed to accelerate finite-state automata. The bone structure of an AP is a hierarchical routing matrix that connects many memory arrays. With this structure, an AP can process an input symbol every clock cycle, and hence achieve much higher performance compared to conventional architectures. However, the design automation for the APs is not well researched. This article proposes a fully automated tool named APmap for mapping the automata to APs that use a two-level routing matrix. APmap first partitions a large automaton into small graphs and then maps them. Multiple transformations are applied to the automaton by APmap to meet hardware constraints. The experiments on a standard benchmark suite show that our approach leads to around 19% less storage utilization compared to state-of-the-art.

@en

Conventional computing architectures and the CMOS technology that they are based on are facing major challenges such as the memory bottleneck making the memory access for data transfer a major killer of energy and performance. Computation-in-memory (CIM) paradigm is seen as a potential alternative that could alleviate such problems by adding computational resources to the memory, and significantly reducing the communication. Memristive devices are promising enablers of a such CIM paradigm, as they are able to support both storage and computing. This paper shows the power of memristive device based CIM paradigm in enabling new efficient application-specific architectures as well as efficient implementations of some known domain-specific architectures. In addition, the paper discusses the potential applications that could benefit from such paradigm and highlights the major challenges.@en

Memristive devices have the potential to reduce the memory access bottleneck in conventional computer architectures. However, memristive devices also suffer from low endurance and large resistance variation. To address these problems, we present a robust logic scheme named Enhanced Scouting Logic (ESL). It produces logic operation results within the peripheral circuit of the memory array. During the execution of logic operations, the resistance states of memristive devices do not change and hence do not affect the memristor lifetime. ESL senses the resistance of input memristive devices via two different paths when different operations such as AND and OR are performed. These different paths guarantee the operation correctness even under large resistance variations. We verified ESL using SPICE simulations and Monte Carlo analysis. Our simulation results show that ESL is more robust as compared with state-of-the-art logic schemes. @en

Von Neumann-based architectures suffer from costly communication between CPU and memory. This communication imposes several orders of magnitude more power and performance overheads compared to the arithmetic operations performed by the processor. This overhead becomes critical for applications that require processing a large amount of data. Computation-in-Memory (CIM) leveraging memristor devices in the crossbar structure offers a potential solution to tackle this challenge. However, support for the integer data type is lacking in CIM approaches as most solutions operate on a single/few bits only. This paper proposes a new organization of the periphery (next to memristor crossbar) to compute matrix-matrix multiplication (MMM) at the tile level. More precisely, the analog additions performed in the crossbar is complemented with additions performed in the digital periphery. In this mixed analog-digital system, digital additions are performed in a way that only the minimum size of adders are required-this is to reduce the latency of the digital periphery as much as possible. In addition, the design is customized to the number of ADCs as well as datatype sizes to support different possible scenarios. The results show that our organization reduces energy and latency up to 50x and 3x, respectively, compared to the reference design.

@en

Technological and architectural improvements have been constantly required to sustain the demand of faster and cheaper computers. However, CMOS down-scaling is suffering from three technology walls: leakage wall, reliability wall, and cost wall. On top of that, a performance increase due to architectural improvements is also
gradually saturating due to three well-known architecture walls: memory wall, power wall, and instruction level parallelism (ILP) wall. Hence, a lot of research is focusing on proposing and developing new technologies and architectures. In this article, we present a comprehensive classification of memory-centric computing architectures; it is based on three metrics: computation location, level of parallelism, and used memory technology. The classification not only provides an overview of existing architectures with their pros and cons but also unifies the terminology that uniquely identifies these architectures and highlights the potential future architectures that can be further explored. Hence, it sets up a direction for future research in the field.@en

Computation-in-memory reverses the trend in von-Neumann processors by bringing the computation closer to the data, to even within the memory array, as opposed to introducing new memory hierarchies to keep (frequently used) data closer to a central processing unit (CPU). In recent years, new non-volatile memory (NVM) technologies, e.g., memristor, PCM, etc., have proven that they can function as memories and perform computations on the stored data as well. In particular, when they are combined with a modest set of (digital) peripheral modules, a wider range of operations can be supported, e.g., vector matrix multiply and Boolean logic. In this paper, we are introducing the CIM-SIM, an open source simulator written in SystemC, which is capable of simulating the functional behaviour of such architectures. The architecture includes the definition of a set of technology-agnostic nano-instructions.

@en

Emerging computing applications (such as big-data and Internet-of-things) are extremely demanding in terms of storage, energy and computational efficiency, while today’s architectures and device technologies are facing major challenges making them incapable to meet these demands. Computation-in-Memory (CIM) architecture based on memristive devices is one of the alternative computing architectures being explored to address these limitations. Enabling such architectures relies on the development of efficient memristive circuits being able to perform logic and arithmetic operations within the non-volatile memory core. This paper addresses memristive circuit designs for CIM architectures. It gives a complete overview of all designs, both for logic as well as arithmetic operations, and presents the most popular designs in details. In addition, it analyzes and classifies them, shows how they result in different CIM flavours and how these architectures distinguish themselves from traditional ones. The paper also presents different potential applications that could significantly benefit from CIM architectures, based on their kernel that could be accelerated. @en

Today's computing architectures suffer from the three well-known bottlenecks, which are the memory, the power and the instruction-level parallelism walls. Emerging non-volatile technologies, such as memristor, enable new resistive architectures that alleviate at least two of such bottlenecks, as they can process data within the memory with almost no leakage. In this paper, we propose a novel resistive computing architecture by extending a conventional architecture with a resistive based Computation-In-Memory accelerator (CIMX). We evaluate the delay, energy and area of the conventional and CIMX architecture using an analytical model and a simulation framework. The results (both based on the analytical model and simulation framework) show that the proposed architecture achieves at least one order of magnitude improvement in terms of performance, area, and energy efficiency for the considered benchmarks.@en

Automata Processor (AP) is a special implementation of non-deterministic finite automata that performs pattern matching by exploring parallel state transitions. The implementation typically contains a hierarchical switching network, causing long latency. This paper proposes a methodology to split such a hierarchical switching network into multiple pipelined stages, making it possible to process several input sequences in parallel by using time-division multiplexing. We use a new resistive RAM based AP (instead of known DRAM or SRAM based) to illustrate the potential of our method. The experimental results show that our approach increases the throughput by almost a factor of 2 at a cost of marginal area overhead.@en