There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

The organization of behavior: A neuropsychological theory

Many architects believe that major improvements in cost-energy-performance must now come from domain-specific hardware. This paper evaluates a custom ASIC---called a Tensor Processing Unit (TPU)---deployed in datacenters since 2015 that accelerates the inference phase of neural networks (NN). The heart of the TPU is a 65,536 8-bit MAC matrix multiply unit that offers a peak throughput of 92 TeraOps/second (TOPS) and a large (28 MiB) software-managed on-chip memory. The TPU's deterministic execution model is a better match to the 99th-percentile response-time requirement of our NN applications than are the time-varying optimizations of CPUs and GPUs (caches, out-of-order execution, multithreading, multiprocessing, prefetching, ...) that help average throughput more than guaranteed latency. The lack of such features helps explain why, despite having myriad MACs and a big memory, the TPU is relatively small and low power. We compare the TPU to a server-class Intel Haswell CPU and an Nvidia K80 GPU, which are contemporaries deployed in the same datacenters. Our workload, written in the high-level TensorFlow framework, uses production NN applications (MLPs, CNNs, and LSTMs) that represent 95% of our datacenters' NN inference demand. Despite low utilization for some applications, the TPU is on average about 15X - 30X faster than its contemporary GPU or CPU, with TOPS/Watt about 30X - 80X higher. Moreover, using the GPU's GDDR5 memory in the TPU would triple achieved TOPS and raise TOPS/Watt to nearly 70X the GPU and 200X the CPU.

In-Datacenter Performance Analysis of a Tensor Processing Unit

Many architects believe that major improvements in cost-energy-performance must now come from domain-specific hardware. This paper evaluates a custom ASIC---called a Tensor Processing Unit (TPU) --- deployed in datacenters since 2015 that accelerates the inference phase of neural networks (NN). The heart of the TPU is a 65,536 8-bit MAC matrix multiply unit that offers a peak throughput of 92 TeraOps/second (TOPS) and a large (28 MiB) software-managed on-chip memory. The TPU's deterministic execution model is a better match to the 99th-percentile response-time requirement of our NN applications than are the time-varying optimizations of CPUs and GPUs that help average throughput more than guaranteed latency. The lack of such features helps explain why, despite having myriad MACs and a big memory, the TPU is relatively small and low power. We compare the TPU to a server-class Intel Haswell CPU and an Nvidia K80 GPU, which are contemporaries deployed in the same datacenters. Our workload, written in the high-level TensorFlow framework, uses production NN applications (MLPs, CNNs, and LSTMs) that represent 95% of our datacenters' NN inference demand. Despite low utilization for some applications, the TPU is on average about 15X -- 30X faster than its contemporary GPU or CPU, with TOPS/Watt about 30X -- 80X higher. Moreover, using the CPU's GDDR5 memory in the TPU would triple achieved TOPS and raise TOPS/Watt to nearly 70X the GPU and 200X the CPU.

/pdf/in-datacenter-performance-analysis-of-a-tensor-processing-4pmp83zqw4.pdf

Deep neural networks (DNNs) are currently widely used for many artificial intelligence (AI) applications including computer vision, speech recognition, and robotics. While DNNs deliver state-of-the-art accuracy on many AI tasks, it comes at the cost of high computational complexity. Accordingly, techniques that enable efficient processing of DNNs to improve energy efficiency and throughput without sacrificing application accuracy or increasing hardware cost are critical to the wide deployment of DNNs in AI systems. This article aims to provide a comprehensive tutorial and survey about the recent advances toward the goal of enabling efficient processing of DNNs. Specifically, it will provide an overview of DNNs, discuss various hardware platforms and architectures that support DNNs, and highlight key trends in reducing the computation cost of DNNs either solely via hardware design changes or via joint hardware design and DNN algorithm changes. It will also summarize various development resources that enable researchers and practitioners to quickly get started in this field, and highlight important benchmarking metrics and design considerations that should be used for evaluating the rapidly growing number of DNN hardware designs, optionally including algorithmic codesigns, being proposed in academia and industry. The reader will take away the following concepts from this article: understand the key design considerations for DNNs; be able to evaluate different DNN hardware implementations with benchmarks and comparison metrics; understand the tradeoffs between various hardware architectures and platforms; be able to evaluate the utility of various DNN design techniques for efficient processing; and understand recent implementation trends and opportunities.

Efficient Processing of Deep Neural Networks: A Tutorial and Survey

Coronary constriction in constrictive pericarditis.

This paper presents an on-chip digitally programmable test structure, referred to as the field programmable thermal emulator (FPTE), for on-line characterization of power pattern, time-varying thermal field, and associated changes in the electrical characteristics of the transistors. A test chip was designed in 130nm CMOS to validate the test structure. The measurement results demonstrated the ability of FPTE to emulate various power patterns and capture the effects on temperature and circuit performance.

Field programmable thermal emulator (FPTE): An all-silicon test structure for thermal characterization of integrated circuits

A test circuit is presented for post-silicon and on-line characterization of the energy-inflection activity of power-gated circuits (the activity when overhead energy is equal to leakage savings) under static (process) and dynamic (voltage/temperature/input) variations. The test circuit is applied to design self-adaptive power-gating for energy-efficient SRAM.

Self-adaptive power gating with test circuit for on-line characterization of energy inflection activity

Energy efficiency and security is a critical requirement for computing at edge nodes. Unrolled architectures for lightweight cryptographic algorithms have been shown to be energy-efficient, providing higher performance while meeting resource constraints. Hardware implementations of unrolled datapaths have also been shown to be resistant to side channel analysis (SCA) attacks due to a reduction in signal-to-noise ratio (SNR) and an increased complexity in the leakage model. This paper demonstrates optimal leakage models and an improved CFA attack which makes it feasible to extract first-order side-channel leakages from combinational logic in the initial rounds of unrolled datapaths. Several leakage models, targeting initial rounds, are explored and 1-bit hamming weight (HW) based leakage model is shown to be an optimal choice. Additionally, multi-band narrow bandpass filtering techniques in conjunction with correlation frequency analysis (CFA) is demonstrated to improve SNR by up to 4×, attributed to the removal of the misalignment effect in combinational logics and signal isolation. The improved CFA attack is performed on side channel signatures acquired for 7-round unrolled SIMON datapaths, implemented on Sakura-G (XILINX spartan 6, 45nm) based FPGA platform and a 24× reduction in minimum-traces-to-disclose (MTD) for revealing 80% of the key bits is demonstrated with respect to conventional time domain correlation power analysis (CPA). Finally, the proposed method is successfully applied to a fully-unrolled datapath for PRINCE and a parallel round-based datapath for Advanced Encryption Standard (AES) algorithm to demonstrate its general applicability.

/pdf/extracting-side-channel-leakage-from-round-unrolled-3aj2shka9t.pdf

Extracting Side-Channel Leakage from Round Unrolled Implementations of Lightweight Ciphers

Learning the evolution of real-time strategy (RTS) game is a challenging problem in artificial intelligent (AI) system. In this paper, we present a novel Hebbian learning method to extract the global feature of a point set in StarCraft II game units, and its application to predict the movement of the points. Our model includes encoder, LSTM, and decoder, and we train the encoder with the unsupervised learning method. We introduce the concept of neuron activity aware learning combined with k-Winner-Takes-All. The optimal value of neuron activity is mathematically derived, and experiments support the effectiveness of the concept over the downstream task. Our Hebbian learning rule benefits the prediction with lower loss compared to self-supervised learning. Also, our model significantly saves the computational cost such as activations and FLOPs compared to a frame-based approach.

Saibal Mukhopadhyay

Papers

Coronary constriction in constrictive pericarditis.

Field programmable thermal emulator (FPTE): An all-silicon test structure for thermal characterization of integrated circuits

Self-adaptive power gating with test circuit for on-line characterization of energy inflection activity

Extracting Side-Channel Leakage from Round Unrolled Implementations of Lightweight Ciphers

Unsupervised Hebbian Learning on Point Sets in StarCraft II