Hyejeong Hong

Total Energy Minimization of Real-Time Tasks in an On-Chip Multiprocessor Using Dynamic Voltage Scaling Efficiency Metric

This paper proposes an algorithm that provides both dynamic voltage scaling and power shutdown to minimize the total energy consumption of an application executed on an on-chip multiprocessor. The proposed algorithm provides an extended schedule and stretch method, where task computations are iteratively stretched within the slack of a time-constrained dependent task set. In addition, the break-even threshold interval for amortizing the shutdown overhead is considered. By evaluating each set of stretched task computations, an energy-efficient set is obtained. The proposed dynamic voltage scaling efficiency metric is the ratio of the reduced energy to the increased cycle time when the supply voltage is scaled, which can be used to determine the task computation cycle to be stretched. Experimental results show that the proposed algorithm outperforms the traditional schedule and stretch method in the various evaluations of target real applications.

A memory-efficient parallel string matching for intrusion detection systems

As the variety of hazardous packet payload contents increases, the intrusion detection system (IDS) should be able to detect numerous patterns in real time. For this reason, this paper proposes an Aho-Corasick algorithm based parallel string matching. In order to balance memory usage between homogeneous finite-state machine (FSM) tiles for each string matcher, an optimal set of bit position groups is determined. Target patterns are sorted by binary-reflected gray code (BRGC), which reduces bit transitions in patterns mapped onto a string matcher. In the evaluations of Snort rules, the proposed string matching outperforms the existing bit-split string matching.

An efficient IP address lookup algorithm based on a small balanced tree using entry reduction

Due to a tremendous increase in internet traffic, backbone routers must have the capability to forward massive incoming packets at several gigabits per second. IP address lookup is one of the most challenging tasks for high-speed packet forwarding. Some high-end routers have been implemented with hardware parallelism using ternary content addressable memory (TCAM). However, TCAM is much more expensive in terms of circuit complexity as well as power consumption. Therefore, efficient algorithmic solutions are essentially required to be implemented using network processors as low cost solutions. Among the state-of-the-art algorithms for IP address lookup, a binary search based on a balanced tree is effective in providing a low-cost solution. In order to construct a balanced search tree, the prefixes with the nesting relationship should be converted into completely disjointed prefixes. A leaf-pushing technique is very useful to eliminate the nesting relationship among prefixes [V. Srinivasan, G. Varghese, Fast address lookups using controlled prefix expansion, ACM Transactions on Computer Systems 17 (1) (1999) 1-40]. However, it creates duplicate prefixes, thus expanding the search tree. This paper proposes an efficient IP address lookup algorithm based on a small balanced tree using entry reduction. The leaf-pushing technique is used for creating the completely disjointed entries. In the leaf-pushed prefixes, there are numerous pairs of adjacent prefixes with similarities in prefix strings and output ports. The number of entries can be significantly reduced by the use of a new entry reduction method which merges pairs with these similar prefixes. After sorting the reduced disjointed entries, a small balanced tree is constructed with a very small node size. Based on this small balanced tree, a native binary search can be effectively used in address lookup issue. In addition, we propose a new multi-way search algorithm to improve a binary search for IPv4 address lookup. As a result, the proposed algorithms offer excellent lookup performance along with reduced memory requirements. Besides, these provide good scalability for large amounts of routing data and for the address migration toward IPv6. Using both various IPv4 and IPv6 routing data, the performance evaluation results demonstrate that the proposed algorithms have better performance in terms of lookup speed, memory requirement and scalability for the growth of entries and IPv6, as compared with other algorithms based on a binary search.

Lifetime Reliability Enhancement of Microprocessors: Mitigating the Impact of Negative Bias Temperature Instability

Ensuring lifetime reliability of microprocessors has become more critical. Continuous scaling and increasing temperatures due to growing power density are threatening lifetime reliability. Negative bias temperature instability (NBTI) has been known for decades, but its impact has been insignificant compared to other factors. Aggressive scaling, however, makes NBTI the most serious threat to chip lifetime reliability in todays and future process technologies. The delay of microprocessors gradually increases as time goes by, due to stress and recovery phases. The delay eventually becomes higher than the value required to meet design constraints, which results in failed systems. In this article, the mechanism of NBTI and its effects on lifetime reliability are presented, then various techniques to mitigate NBTI degradation on microprocessors are introduced. The mitigation can be addressed at either the circuit level or architectural level. Circuit-level techniques include design-time techniques such as transistor sizing and NBTI-aware synthesis. Forward body biasing, and adaptive voltage scaling are adaptive techniques that can mitigate NBTI degradation at the circuit level by controlling the threshold voltage or supply voltage to hide the lengthened delay caused by NBTI degradation. Reliability has been regarded as something to be addressed by chip manufacturers. However, there are recent attempts to bring lifetime reliability problems to the architectural level. Architectural techniques can reduce the cost added by circuit-level techniques, which are based on the worst-case degradation estimation. Traditional low-power and thermal management techniques can be successfully extended to deal with reliability problems since aging is dependent on power consumption and temperature. Self-repair is another option to enhance the lifetime of microprocessors using either core-level or lower-level redundancy. With a growing thermal crisis and constant scaling, lifetime reliability requires more intensive research in conjunction with other design issues.

A memory-efficient heterogeneous parallel pattern matching scheme in deep packet inspection

This paper presents for hardware-based parallel pattern matching scheme that adopts heterogeneous bit-split string matchers for deep packet inspection (DPI) devices. Considering the pattern lengths, a set of target patterns is partitioned into two subsets for short and long patterns. By adopting the appropriate bit-split string matcher types for the two subsets, the memory requirements can be optimized for the bit-split parallel pattern matching engine. Experimental results show that the total memory requirements decrease by 39.40% and 20.52%, in comparison with the existing bit-split pattern matching approaches.

Thermal-aware dynamic voltage frequency scaling for many-core processors under process variations

Process variations yield the asymmetry on core performance in many-core processors. Adaptive voltage scaling can hide the variations, but that results in the different thermal characteristics of cores. By using the thermal characteristics, the efficiency of energy optimization and temperature management can be improved. Experiments showed that the proposed dynamic voltage frequency scaling consumes up to 25.2% less energy than the existing thermal management technique while remaining the ratio of peak temperature violations under 1%.

Process variation-aware floorplanning for 3D many-core processors

Thermal management is one of the critical issues in 3D many-core processors design. 3D many-core floorplanning has so far focused on only the configuration of cores and memories across layers. However, 3D floorplanning should also take die stack ordering into account because the characteristics of dies may vary due to growing process variations. A new 3D floorplanning approach which covers die stack ordering is proposed. The evaluation shows that peak steady state temperature is reduced by about 2 K without any overhead in manufacturing process.

Recovery-enhancing task scheduling for multicore processors under NBTI impact

Task scheduling for multicore processors typically has aimed at higher throughput and lower power consumption, but the lifetime reliability may be harmed if negative bias temperature instability is not considered together. In this letter, a task scheduling which extends lifetime is proposed. Cores are given the chance to recover as long as performance is not degraded. The degradation simulator shows that the proposed scheduling improves the lifetime of a multicore processor by up to 36% over the existing task scheduling.

Dynamic thermal management for 3D multicore processors under process variations

Stacking core layers is emerging as an alternative for future high performance computing, but thermal problems have to be tackled first. When adaptive voltage scaling is adopted to hide the growing variation in the performance of cores, as a result, heat generation of each core varies. By exploiting the static thermal characteristics, the efficiency of dynamic thermal management can be improved. The proposed thermal management reduces the energy consumption by up to 30.02% compared with existing techniques, while keeping the ratio of temperature violations around 1%.