We introduce a novel tensor contraction processor (TCP) architecture that offers a paradigm shift from traditional architectures that rely on fixed-size matrix multiplications. TCP aims at exploiting the rich parallelism and data locality inherent in tensor contractions, thereby enhancing both efficiency and performance of AI workloads.TCP is composed of coarse-grained processing elements (PEs) to simplify software development. In order to efficiently process operations with diverse tensor shapes, the PEs are designed to be flexible enough to be utilized as a large-scale single unit or a set of small independent compute units.We aim at maximizing data reuse on both levels of inter and intra compute units. To do that, we propose a circuit switch-based fetch network to flexibly connect compute units to enable inter-compute unit data reuse. We also exploit input broadcast to multiple contraction engines and input buffer based reuse to further exploit reuse behavior in tensor contraction. Our compiler explores the design space of tensor contractions considering tensor shapes and the order of their associated loop operations as well as the underlying accelerator architecture.A TCP chip was designed and fabricated in 5nm technology as the second-generation product of Furiosa AI, offering 256/512/1024 TOPS (BF16/FP8 or INT8/INT4) with 256 MB SRAM and 1.5 TB/s 48 GB HBM3 under 150 W TDP. Commercialization will start in August 2024.We performed an extensive case study of running the LLaMA-2 7B model and evaluated its performance and power efficiency on various configurations of sequence length and batch size. For this model, TCP is 2.7 × and 4.1 × better than H100 and L40s, respectively, in terms of performance per watt.

Motivation: Analysis of RNA sequencing (RNA-Seq) data in human saliva is challenging. Lack of standardization and unification of the bioinformatic procedures undermines saliva's diagnostic potential. Thus, it motivated us to perform this study. Results: We applied principal pipelines for bioinformatic analysis of small RNA-Seq data of saliva of 98 healthy Korean volunteers including either direct or indirect mapping of the reads to the human genome using Bowtie1. Analysis of alignments to exogenous genomes by another pipeline revealed that almost all of the reads map to bacterial genomes. Thus, salivary exRNA has fundamental properties that warrant the design of unique additional steps while performing the bioinformatic analysis. Our pipelines can serve as potential guidelines for processing of RNA-Seq data of human saliva. Availability and implementation: Processing and analysis results of the experimental data generated by the exceRpt (v4.6.3) small RNA-seq pipeline (github.gersteinlab.org/exceRpt) are available from exRNA atlas (exrna-atlas.org). Alignment to exogenous genomes and their quantification results were used in this paper for the analyses of small RNAs of exogenous origin. Contact: [email protected].

The demands and interests in non-contact health monitoring have increased rapidly in recent years due to its noninvasive method and easy to use in daily life. In this paper, a real time and non-contact based cardiac monitoring system including a sensor is presented. The sensor is capacitively-coupled with human skin and does not cause impediments in the natural movement of the user. Moreover, low power RF module nRF24L01+ and passive components were used for reducing power consumption. Current monitoring systems for premature infants are not convenient due to the usage of patched or adhesive tape for monitoring vital information. The front end circuit is integrated with multiple stages of amplifier, filters, analog to digital converter and the wireless module. The non-contact capacitive electrode was integrated into the leather belt as well as in the chair to monitor electrocardiography (ECG) signals. In this work, a simple and easily accessible ECG monitoring sensor node for body area networks was designed and tested. The proposed system was compared through experiments with the one with Ag/Agcl based electrodes. The results show that the proposed system could accurately be used as a medical grade equipment for long-term health monitoring.

A miniaturized monopole antenna was designed and fabricated on an organic paper and LCP material for wireless body area network. Compared with previous work, the proposed design has 20% reduction of the antenna size but with enhanced performance. The effects of the compact coplanar antenna under different twisting conditions is described in this paper. The proposed antennas are simulated and designed on an organic paper and a Liquid Crystal Polymer (LCP) substrate with dielectric constant Dr= 3.4 and thickness 15μm and 5μm respectively, occupying the area of 22×30mm2. A detailed discussion about radiation pattern, Gain, antenna efficiency and power pattern is given with the help of experimental and numerical results.

Increasing interests in remote monitoring of vital signs through telecommunication, especially with wireless and mobile communication have enabled a new generation of information system for healthcare applications. The system may include miniature sensor nodes embedded with wireless communications and a mobile computer delivering information to remote locations. In this paper, we introduced an inexpensive and ultra-low power system for measuring ECG and heart rate in contact and contact-less manners. Existing monitoring system uses gel or electrodes for measuring ECG signals. However, in this study, we used unique electric potential EPIC sensors from Plessey's and infrared sensors. The prototype is capable of monitoring both heart rate and ECG signals with a hibernation mode, which would require less power to transmit the data. Our developed prototype can be used to monitor premature infants as their skins are sensitive and current system uses patches or gel to collect biomedical signals. The proposed prototype for monitoring vital signs has been tested in our lab. The results show that it can achieve low power consumption though a hibernation mode.

We develop an evolutionary fuzzy proportional-integral-derivative (PID) controller for a permanent magnet synchronous motor (PMSM). We first consider a fuzzy PID control design problem based on the common control engineering knowledge that good transient performances can be obtained by increasing the P and I gains and decreasing the D gain when the transient error is big. Then we give an evolutionary algorithm (EA) to autotune the control parameters of the fuzzy PID controller. We implement the proposed EA-based fuzzy PID control controller in real time on a Texas Instruments TMS320F28335 floating-point DSP. We also give simulation and experimental results to show the effectiveness of the proposed intelligent digital control system under abrupt load torque variation using a prototype PMSM.

In this paper, we propose an indoor positioning system using smartphones. To localize the position of target that is pedestrian in usual indoors, we make use of perturbations of the geomagnetic field caused by structural steel elements in a building. The system based on magnetic field does not need any physical infrastructure, so it is possible to be realized with low cost. To estimate the target's position using the geomagnetic anomaly, the proposed system measures the magnetic field on its own position using a magnetometer embedded in smartphones and compares the sensor measurement with the magnetic map that has been created for the building in advance. The estimated position is calculated by a stochastic system based on the particle filter. To calculate control inputs for the particle filter, such as moving distance and direction, we exploit the inertial measurement unit (IMU) that is composed of 3-axis accelerometer and gyroscope built in the smartphone. The experimental results show that accuracy is within 3 meter. These results imply the potential to track people in the buildings that have geomagnetic map in advance at the meter scale using only smartphone with embedded sensors.

A new methodology for radiation induced real-time fault detection and diagnosis, utilizing FPGA-based architectures was developed. The methodology includes a full test platform to evaluate a circuit while under radiation and an algorithm to detect and diagnose fault locations within a circuit using Triple Design Triple Modular Redundancy (TDTMR). An analysis of the system was established using a fault injection. Additionally a functional gamma irradiation analysis was performed to assess the effectiveness of the method. The detection and diagnosis algorithm was capable of detecting errors by switching dynamically during the analysis of an FPGA. However, only the injected fault test was able to properly diagnose the location of the fault. The results indicate that FPGA radiation induced fault production is dependent upon radiation dose rate. A fully interchangeable and operational testing platform has been established along with an algorithm that detects and diagnoses errors in real-time.

Recent advances in the ubiquitous computing and the context-aware system have driven the intelligent agents for the management of user movements and multi contexts. With the growth of the agent-based system and the advances in networking technologies, the design and development of the multi agent system have risen in importance. However, the conventional models offer some kinds of static services to limited users and operate the sequential structure of multi agent middleware. Relatively few studies have been devoted to the research issues of the convergence in heterogeneous networks, the diversity in service requirements and the efficiency in service management to effectively support next-generation context-aware services. Therefore, we propose a parallel Multi Agent Middleware which provides efficient approaches to manage the seamless service and the reconstruction of services according to user movements and a couple of situations. Our architecture allows us to create a rich set of MsMu (Multi Source Multi Use) services through the extension of OsMu (One Source Multi Use) services for enhancing a service satisfaction. Moreover, we enhance the aspect of the service response time and the success probability of the context-aware schedule by using the determination rules of service reconstruction.

In smart home, people who use of context-aware systems should expect to be more dynamic and adaptive services by analyzing residents along with life styles. Currently, the research effort is focused on the service-oriented structure and ontology-based context model. However, the conventional model did not consider the classified service, service reconstruction and definition of service type by classified service patterns. In this paper, we propose a session manager which manages service registration states and classified service according to the user movements. Moreover, a session manager offers to analyze the service pattern by changing environments and user location. Our system could create the dynamic services according to service pattern analysis. Therefore, our system reduces the offered service delays and improves the maximum 64% performance of service response time compared with non-existence the session manager. We show the feasibility of our system by evaluating the efficiency on a test bed.

This paper presents a framework to address the combination of ubiquitous computing interests as an intelligent assisting location-aware service and development of its hardware and software systems. In our real testbed, the system can dynamically localize the areas of interest and adapt to changes in the ubiquitous intelligence landscape.

This paper proposes a context model which separates the upper and lower layer according to characteristic of each class using the Web Ontology Language (OWL). We also present a context-aware multi-agent system (CAMAS) architecture which provides the solution of policy problem using the classification of rule and implementation of a rule-based engine. We enhance the response time for offering adaptive services. We solve the policy conflict problem among users by using reasoning and privilege delegation.

This paper discusses the Pareto principle as it relates to the distribution of software defects in code. We look at evidence in the context of both the software test team, and users of the software. We also investigate two related principles. The first principle is that the distribution of defects in code relates to the distribution of complexity in code. The second principle is that how we define complexity relates to the distribution of defects in code. We present this work as an empirical study of three general hypotheses investigated for large production-level software; we show that the essence of the principle holds, while precise percentages do not.