Search

Frameworks for writing, compiling, and optimizing deep learning (DL) models have recently enabled progress in areas like computer vision and natural language processing. Extending these frameworks to accommodate the rapidly diversifying landscape of DL models and hardware platforms presents challenging tradeoffs between expressivity, composability, and portability. We present Relay, a new compiler framework for DL. Relay's functional, statically typed intermediate representation (IR) unifies and generalizes existing DL IRs to express state-of-the-art models. The introduction of Relay's expressive IR requires careful design of domain-specific optimizations, addressed via Relay's extension mechanisms. Using these extension mechanisms, Relay supports a unified compiler that can target a variety of hardware platforms. Our evaluation demonstrates Relay's competitive performance for a broad class of models and devices (CPUs, GPUs, and emerging accelerators). Relay's design demonstrates how a unified IR can provide expressivity, composability, and portability without compromising performance.

State of the art deep learning models have made steady progress in the fields of computer vision and natural language processing, at the expense of growing model sizes and computational complexity. Deploying these models on low power and mobile devices poses a challenge due to their limited compute capabilities and strict energy budgets. One solution that has generated significant research interest is deploying highly quantized models that operate on low precision inputs and weights less than eight bits, trading off accuracy for performance. These models have a significantly reduced memory footprint (up to 32x reduction) and can replace multiply-accumulates with bitwise operations during compute intensive convolution and fully connected layers. Most deep learning frameworks rely on highly engineered linear algebra libraries such as ATLAS or Intel's MKL to implement efficient deep learning operators. To date, none of the popular deep learning directly support low precision operators, partly due to a lack of optimized low precision libraries. In this paper we introduce a work flow to quickly generate high performance low precision deep learning operators for arbitrary precision that target multiple CPU architectures and include optimizations such as memory tiling and vectorization. We present an extensive case study on low power ARM Cortex-A53 CPU, and show how we can generate 1-bit, 2-bit convolutions with speedups up to 16x over an optimized 16-bit integer baseline and 2.3x better than handwritten implementations., Comment: 10 pages, 11 figures

When a computational task tolerates a relaxation of its specification or when an algorithm tolerates the effects of noise in its execution, hardware, programming languages, and system software can trade deviations from correct behavior for lower resource usage. We present, for the first time, a synthesis of research results on computing systems that only make as many errors as their users can tolerate, from across the disciplines of computer aided design of circuits, digital system design, computer architecture, programming languages, operating systems, and information theory. Rather than over-provisioning resources at each layer to avoid errors, it can be more efficient to exploit the masking of errors occurring at one layer which can prevent them from propagating to a higher layer. We survey tradeoffs for individual layers of computing systems from the circuit level to the operating system level and illustrate the potential benefits of end-to-end approaches using two illustrative examples. To tie together the survey, we present a consistent formalization of terminology, across the layers, which does not significantly deviate from the terminology traditionally used by research communities in their layer of focus., Comment: 35 pages

Specialized Deep Learning (DL) acceleration stacks, designed for a specific set of frameworks, model architectures, operators, and data types, offer the allure of high performance while sacrificing flexibility. Changes in algorithms, models, operators, or numerical systems threaten the viability of specialized hardware accelerators. We propose VTA, a programmable deep learning architecture template designed to be extensible in the face of evolving workloads. VTA achieves this flexibility via a parametrizable architecture, two-level ISA, and a JIT compiler. The two-level ISA is based on (1) a task-ISA that explicitly orchestrates concurrent compute and memory tasks and (2) a microcode-ISA which implements a wide variety of operators with single-cycle tensor-tensor operations. Next, we propose a runtime system equipped with a JIT compiler for flexible code-generation and heterogeneous execution that enables effective use of the VTA architecture. VTA is integrated and open-sourced into Apache TVM, a state-of-the-art deep learning compilation stack that provides flexibility for diverse models and divergent hardware backends. We propose a flow that performs design space exploration to generate a customized hardware architecture and software operator library that can be leveraged by mainstream learning frameworks. We demonstrate our approach by deploying optimized deep learning models used for object classification and style transfer on edge-class FPGAs., Comment: 6 pages plus references, 8 figures

We introduce a learning-based framework to optimize tensor programs for deep learning workloads. Efficient implementations of tensor operators, such as matrix multiplication and high dimensional convolution, are key enablers of effective deep learning systems. However, existing systems rely on manually optimized libraries such as cuDNN where only a narrow range of server class GPUs are well-supported. The reliance on hardware-specific operator libraries limits the applicability of high-level graph optimizations and incurs significant engineering costs when deploying to new hardware targets. We use learning to remove this engineering burden. We learn domain-specific statistical cost models to guide the search of tensor operator implementations over billions of possible program variants. We further accelerate the search by effective model transfer across workloads. Experimental results show that our framework delivers performance competitive with state-of-the-art hand-tuned libraries for low-power CPU, mobile GPU, and server-class GPU., Comment: NeurIPS 2018

There is an increasing need to bring machine learning to a wide diversity of hardware devices. Current frameworks rely on vendor-specific operator libraries and optimize for a narrow range of server-class GPUs. Deploying workloads to new platforms -- such as mobile phones, embedded devices, and accelerators (e.g., FPGAs, ASICs) -- requires significant manual effort. We propose TVM, a compiler that exposes graph-level and operator-level optimizations to provide performance portability to deep learning workloads across diverse hardware back-ends. TVM solves optimization challenges specific to deep learning, such as high-level operator fusion, mapping to arbitrary hardware primitives, and memory latency hiding. It also automates optimization of low-level programs to hardware characteristics by employing a novel, learning-based cost modeling method for rapid exploration of code optimizations. Experimental results show that TVM delivers performance across hardware back-ends that are competitive with state-of-the-art, hand-tuned libraries for low-power CPU, mobile GPU, and server-class GPUs. We also demonstrate TVM's ability to target new accelerator back-ends, such as the FPGA-based generic deep learning accelerator. The system is open sourced and in production use inside several major companies., Comment: Significantly improved version, add automated optimization

Co-designing efficient machine learning based systems across the whole hardware/software stack to trade off speed, accuracy, energy and costs is becoming extremely complex and time consuming. Researchers often struggle to evaluate and compare different published works across rapidly evolving software frameworks, heterogeneous hardware platforms, compilers, libraries, algorithms, data sets, models, and environments. We present our community effort to develop an open co-design tournament platform with an online public scoreboard. It will gradually incorporate best research practices while providing a common way for multidisciplinary researchers to optimize and compare the quality vs. efficiency Pareto optimality of various workloads on diverse and complete hardware/software systems. We want to leverage the open-source Collective Knowledge framework and the ACM artifact evaluation methodology to validate and share the complete machine learning system implementations in a standardized, portable, and reproducible fashion. We plan to hold regular multi-objective optimization and co-design tournaments for emerging workloads such as deep learning, starting with ASPLOS'18 (ACM conference on Architectural Support for Programming Languages and Operating Systems - the premier forum for multidisciplinary systems research spanning computer architecture and hardware, programming languages and compilers, operating systems and networking) to build a public repository of the most efficient machine learning algorithms and systems which can be easily customized, reused and built upon., Comment: ReQuEST tournament website: http://cKnowledge.org/request

As a result of the increasing demand for deep neural network (DNN)-based services, efforts to develop dedicated hardware accelerators for DNNs are growing rapidly. However,while accelerators with high performance and efficiency on convolutional deep neural networks (Conv-DNNs) have been developed, less progress has been made with regards to fully-connected DNNs (FC-DNNs). In this paper, we propose MATIC (Memory Adaptive Training with In-situ Canaries), a methodology that enables aggressive voltage scaling of accelerator weight memories to improve the energy-efficiency of DNN accelerators. To enable accurate operation with voltage overscaling, MATIC combines the characteristics of destructive SRAM reads with the error resilience of neural networks in a memory-adaptive training process. Furthermore, PVT-related voltage margins are eliminated using bit-cells from synaptic weights as in-situ canaries to track runtime environmental variation. Demonstrated on a low-power DNN accelerator that we fabricate in 65 nm CMOS, MATIC enables up to 60-80 mV of voltage overscaling (3.3x total energy reduction versus the nominal voltage), or 18.6x application error reduction., Comment: 6 pages, 12 figures, 3 tables. Published at Design, Automation and Test in Europe Conference and Exhibition (DATE) 2018

Cameras are the defacto sensor. The growing demand for real-time and low-power computer vision, coupled with trends towards high-efficiency heterogeneous systems, has given rise to a wide range of image processing acceleration techniques at the camera node and in the cloud. In this paper, we characterize two novel camera systems that use acceleration techniques to push the extremes of energy and performance scaling, and explore the computation-communication tradeoffs in their design. The first case study targets a camera system designed to detect and authenticate individual faces, running solely on energy harvested from RFID readers. We design a multi-accelerator SoC design operating in the sub-mW range, and evaluate it with real-world workloads to show performance and energy efficiency improvements over a general purpose microprocessor. The second camera system supports a 16-camera rig processing over 32 Gb/s of data to produce real-time 3D-360 degree virtual reality video. We design a multi-FPGA processing pipeline that outperforms CPU and GPU configurations by up to 10x in computation time, producing panoramic stereo video directly from the camera rig at 30 frames per second. We find that an early data reduction step, either before complex processing or offloading, is the most critical optimization for in-camera systems.

We conduct a review to assess how the simulation of repeated or recurrent events are planned. For such multivariate time-to-events, it is well established that the underlying mechanism is likely to be complex and to involve in particular both heterogeneity in the population and event-dependence. In this respect, we particularly focused on these two dimensions of events dynamic when mimicking actual data. Next, we investigate whether the processes generated in the simulation studies have similar properties to those expected in the clinical data of interest. Finally we describe a simulation scheme for generating data according to the timescale of choice (gap time/ calendar) and to whether heterogeneity and/or event-dependence are to be considered. The main finding is that event-dependence is less widely considered in simulation studies than heterogeneity. This is unfortunate since the occurrence of an event may alter the risk of occurrence of new events.