Deployment of modern TinyML tasks on small battery-constrained IoT devices requires high computational energy efficiency. Analog In-Memory Computing (IMC) using non-volatile memory (NVM) promises major efficiency improvements in deep neural network (DNN) inference and serves as on-chip memory storage for DNN weights. However, IMC's functional flexibility limitations and their impact on performance, energy, and area efficiency are not yet fully understood at the system level. To target practical end-to-end IoT applications, IMC arrays must be enclosed in heterogeneous programmable systems, introducing new system-level challenges which we aim at addressing in this work. We present a heterogeneous tightly-coupled clustered architecture integrating 8 RISC-V cores, an in-memory computing accelerator (IMA), and digital accelerators. We benchmark the system on a highly heterogeneous workload such as the Bottleneck layer from a MobileNetV2, showing 11.5x performance and 9.5x energy efficiency improvements, compared to highly optimized parallel execution on the cores. Furthermore, we explore the requirements for end-to-end inference of a full mobile-grade DNN (MobileNetV2) in terms of IMC array resources, by scaling up our heterogeneous architecture to a multi-array accelerator. Our results show that our solution, on the end-to-end inference of the MobileNetV2, is one order of magnitude better in terms of execution latency than existing programmable architectures and two orders of magnitude better than state-of-the-art heterogeneous solutions integrating in-memory computing analog cores.
Traditional graphics processing units (GPUs) suffer from the low memory capacity and demand for high memory bandwidth. To address these challenges, we propose Ohm-GPU, a new optical network based heterogeneous memory design for GPUs. Specifically, Ohm-GPU can expand the memory capacity by combing a set of high-density 3D XPoint and DRAM modules as heterogeneous memory. To prevent memory channels from throttling throughput of GPU memory system, Ohm-GPU replaces the electrical lanes in the traditional memory channel with a high-performance optical network. However, the hybrid memory can introduc…
The trend in industry is towards heterogeneous multicore processors (HMCs), including chips with CPUs and massively-threaded throughput-oriented processors (MTTOPs) such as GPUs. Although current homogeneous chips tightly couple the cores with cache-coherent shared virtual memory (CCSVM), this is not the communication paradigm used by any current HMC. In this paper, we present a CCSVM design for a CPU/MTTOP chip, as well as an extension of the pthreads programming model, called xthreads, for programming this HMC. Our goal is to evaluate the potential performance benefits of tightly coupling he…
This work presents a heterogeneous communication library for clusters of processors and FPGAs. This library, Shoal, supports the Partitioned Global Address Space (PGAS) memory model for applications. PGAS is a shared memory model for clusters that creates a distinction between local and remote memory access. Through Shoal and its common application programming interface for hardware and software, applications can be more freely migrated to the optimal platform and deployed onto dynamic cluster topologies. The library is tested using a thorough suite of microbenchmarks to establish latency and …
Privatizing data is a useful strategy for increasing parallelism in a shared memory multithreaded program. Independent cores can compute independently on duplicates of shared data, combining their results at the end of their computations. Conventional approaches to privatization, however, rely on explicit static or dynamic memory allocation for duplicated state, increasing memory footprint and contention for cache resources, especially in shared caches. In this work, we describe CCache, a system for on-demand privatization of data manipulated by commutative operations. CCache garners the benef…
Hardware specialization is becoming a key enabler of energyefficient performance. Future systems will be increasingly heterogeneous, integrating multiple specialized and programmable accelerators, each with different memory demands. Traditionally, communication between accelerators has been inefficient, typically orchestrated through explicit DMA transfers between different address spaces. More recently, industry has proposed unified coherent memory which enables implicit data movement and more data reuse, but often these interfaces limit the coherence flexibility available to heterogeneous sy…
In modern heterogeneous MPSoCs, the management of shared memory resources is crucial in delivering end-to-end QoS. Previous frameworks have either focused on singular QoS targets or the allocation of partitionable resources among CPU applications at relatively slow timescales. However, heterogeneous MPSoCs typically require instant response from the memory system where most resources cannot be partitioned. Moreover, the health of different cores in a heterogeneous MPSoC is often measured by diverse performance objectives. In this work, we propose a Self-Aware Resource Allocation (SARA) framewo…
Modern SoCs integrate multiple CPU cores and Hardware Accelerators (HWAs) that share the same main memory system, causing interference among memory requests from different agents. The result of this interference, if not controlled well, is missed deadlines for HWAs and low CPU performance. State-of-the-art mechanisms designed for CPU-GPU systems strive to meet a target frame rate for GPUs by prioritizing the GPU close to the time when it has to complete a frame. We observe two major problems when such an approach is adapted to a heterogeneous CPU-HWA system. First, HWAs miss deadlines because …
Shared virtual memory (SVM) is key in heterogeneous systems on chip (SoCs), which combine a general-purpose host processor with a many-core accelerator, both for programmability and to avoid data duplication. However, SVM can bring a significant run time overhead when translation lookaside buffer (TLB) entries are missing. Moreover, allowing DMA burst transfers to write SVM traditionally requires buffers to absorb transfers that miss in the TLB. These buffers have to be overprovisioned for the maximum burst size, wasting precious on-chip memory, and stall all SVM accesses once they are full, h…
Heterogeneous systems appear as a viable design alternative for the dark silicon era. In this paradigm, a processor chip includes several different technological alternatives for implementing a certain logical block (e.g., core, on-chip memories) which cannot be used at the same time due to power constraints. The programmer and compiler are then responsible for selecting which of the alternatives should be used for maximizing performance and/or energy efficiency for a given application. This paper presents an initial approach for the exploration of different technological alternatives for the …
Computers continue to diversify with respect to system designs, emerging memory technologies, and application memory demands. Unfortunately, continually adapting the conventional virtual memory framework to each possible system configuration is challenging, and often results in performance loss or requires non-trivial workarounds. To address these challenges, we propose a new virtual memory framework, the Virtual Block Interface (VBI). We design VBI based on the key idea that delegating memory management duties to hardware can reduce the overheads and software complexity associated with virtua…
Modern deep learning models have high memory and computation cost. To make them fast and memory-cost efficient, structured model pruning is commonly used. We find that pruning a model using a common training accelerator with large systolic arrays is extremely performance-inefficient. To make a systolic array efficient for pruning and training, we propose FlexSA, a flexible systolic array architecture. FlexSA dynamically reconfigures the systolic array structure and offers multiple sub-systolic operating modes, which are designed for energy- and memory bandwidth-efficient processing of tensors …
Ultra-fast \& low-power superconductor single-flux-quantum (SFQ)-based CNN systolic accelerators are built to enhance the CNN inference throughput. However, shift-register (SHIFT)-based scratchpad memory (SPM) arrays prevent a SFQ CNN accelerator from exceeding 40\% of its peak throughput, due to the lack of random access capability. This paper first documents our study of a variety of cryogenic memory technologies, including Vortex Transition Memory (VTM), Josephson-CMOS SRAM, MRAM, and Superconducting Nanowire Memory, during which we found that none of the aforementioned technologies made a …
Main memories play an important role in overall energy consumption of embedded systems. Using conventional memory technologies in future designs in nanoscale era causes a drastic increase in leakage power consumption and temperature-related problems. Emerging non-volatile memory (NVM) technologies offer many desirable characteristics such as near-zero leakage power, high density and non-volatility. They can significantly mitigate the issue of memory leakage power in future embedded chip-multiprocessor (eCMP) systems. However, they suffer from challenges such as limited write endurance and high…
Computer Systems Lab, School of Engineering & Applied Sciences, Yale University（耶鲁大学工程与应用科学学院计算机系统实验室）；Center of Theoretical Neuroscience, Columbia University（哥伦比亚大学理论神经科学中心）；Institute of Neuroinformatics, University of Zurich and ETH Zurich（苏黎世大学与苏黎世联邦理工学院神经信息学研究所）
Neuromorphic computing systems comprise networks of neurons that use asynchronous events for both computation and communication. This type of representation offers several advantages in terms of bandwidth and power consumption in neuromorphic electronic systems. However, managing the traffic of asynchronous events in large scale systems is a daunting task, both in terms of circuit complexity and memory requirements. Here we present a novel routing methodology that employs both hierarchical and mesh routing strategies and combines heterogeneous memory structures for minimizing both memory requi…
Multi-port memory controllers (MPMCs) have become increasingly important in many modern applications due to the tremendous growth in bandwidth requirement. Many approaches so far have focused on improving either the memory access latency or the bandwidth utilization for specific applications. Moreover, the application systems are likely to require certain adjustments to connect with an MPMC, since the MPMC interface is limited to a single-clock and single data-width domain. In this paper, we propose efficient techniques to improve the flexibility, latency, and bandwidth of an MPMC. Firstly, MP…
When multiple processor cores (CPUs) and a GPU integrated together on the same chip share the off-chip DRAM, requests from the GPU can heavily interfere with requests from the CPUs, leading to low system performance and starvation of cores. Unfortunately, state-of-the-art memory scheduling algorithms are ineffective at solving this problem due to the very large amount of GPU memory traffic, unless a very large and costly request buffer is employed to provide these algorithms with enough visibility across the global request stream. Previously-proposed memory controller (MC) designs use a single…
