Approximate computing has recently emerged as a promising approach to energy-efficient design of digital systems. Approximate computing relies on the ability of many systems and applications to tolerate some loss of quality or optimality in the computed result. By relaxing the need for fully precise or completely deterministic operations, approximate computing techniques allow substantially improved energy efficiency. This paper reviews recent progress in the area, including design of approximate arithmetic blocks, pertinent error and quality measures, and algorithm-level techniques for approximate computing.

/pdf/approximate-computing-an-emerging-paradigm-for-energy-z7hetfdl3q.pdf

Approximate computing: An emerging paradigm for energy-efficient design

Recent technological advances in the development of flash-memory based devices have consolidated their leadership position as the preferred storage media in the embedded systems market and opened new vistas for deployment in enterprise-scale storage systems. Unlike hard disks, flash devices are free from any mechanical moving parts, have no seek or rotational delays and consume lower power. However, the internal idiosyncrasies of flash technology make its performance highly dependent on workload characteristics. The poor performance of random writes has been a cause of major concern, which needs to be addressed to better utilize the potential of flash in enterprise-scale environments. We examine one of the important causes of this poor performance: the design of the Flash Translation Layer (FTL), which performs the virtual-to-physical address translations and hides the erase-before-write characteristics of flash. We propose a complete paradigm shift in the design of the core FTL engine from the existing techniques with our Demand-based Flash Translation Layer (DFTL), which selectively caches page-level address mappings. We develop a flash simulation framework called FlashSim. Our experimental evaluation with realistic enterprise-scale workloads endorses the utility of DFTL in enterprise-scale storage systems by demonstrating: (i) improved performance, (ii) reduced garbage collection overhead and (iii) better overload behavior compared to state-of-the-art FTL schemes. For example, a predominantly random-write dominant I/O trace from an OLTP application running at a large financial institution shows a 78% improvement in average response time (due to a 3-fold reduction in operations of the garbage collector), compared to a state-of-the-art FTL scheme. Even for the well-known read-dominant TPC-H benchmark, for which DFTL introduces additional overheads, we improve system response time by 56%.

/pdf/dftl-a-flash-translation-layer-employing-demand-based-1jqzm0wlg2.pdf

DFTL: a flash translation layer employing demand-based selective caching of page-level address mappings

Data sets in large applications are often too massive to fit completely inside the computers internal memory. The resulting input/output communication (or I/O) between fast internal memory and slower external memory (such as disks) can be a major performance bottleneck. In this article we survey the state of the art in the design and analysis of external memory (or EM) algorithms and data structures, where the goal is to exploit locality in order to reduce the I/O costs. We consider a variety of EM paradigms for solving batched and online problems efficiently in external memory. For the batched problem of sorting and related problems such as permuting and fast Fourier transform, the key paradigms include distribution and merging. The paradigm of disk striping offers an elegant way to use multiple disks in parallel. For sorting, however, disk striping can be nonoptimal with respect to I/O, so to gain further improvements we discuss distribution and merging techniques for using the disks independently. We also consider useful techniques for batched EM problems involving matrices (such as matrix multiplication and transposition), geometric data (such as finding intersections and constructing convex hulls), and graphs (such as list ranking, connected components, topological sorting, and shortest paths). In the online domain, canonical EM applications include dictionary lookup and range searching. The two important classes of indexed data structures are based upon extendible hashing and B-trees. The paradigms of filtering and bootstrapping provide a convenient means in online data structures to make effective use of the data accessed from disk. We also reexamine some of the above EM problems in slightly different settings, such as when the data items are moving, when the data items are variable-length (e.g., text strings), or when the allocated amount of internal memory can change dynamically. Programming tools and environments are available for simplifying the EM programming task. During the course of the survey, we report on some experiments in the domain of spatial databases using the TPIE system (transparent parallel I/O programming environment). The newly developed EM algorithms and data structures that incorporate the paradigms we discuss are significantly faster than methods currently used in practice.

https://users.cs.duke.edu/~reif/courses/alglectures/vitter.papers/Vit.IO_survey.pdf

External memory algorithms and data structures: dealing with massive data

2004아테네 올림픽대회출전 한국레슬링 대표선수의 무산소성 운동능력 및 등속성 근력

Virtualized data centers enable sharing of resources among hosted applications. However, it is difficult to satisfy service-level objectives(SLOs) of applications on shared infrastructure, as application workloads and resource consumption patterns change over time. In this paper, we present AutoControl, a resource control system that automatically adapts to dynamic workload changes to achieve application SLOs. AutoControl is a combination of an online model estimator and a novel multi-input, multi-output (MIMO) resource controller. The model estimator captures the complex relationship between application performance and resource allocations, while the MIMO controller allocates the right amount of multiple virtualized resources to achieve application SLOs. Our experimental evaluation with RUBiS and TPC-W benchmarks along with production-trace-driven workloads indicates that AutoControl can detect and mitigate CPU and disk I/O bottlenecks that occur over time and across multiple nodes by allocating each resource accordingly. We also show that AutoControl can be used to provide service differentiation according to the application priorities during resource contention.

/pdf/automated-control-of-multiple-virtualized-resources-4y2y4mu57k.pdf

Automated control of multiple virtualized resources

Virtualized servers run a diverse set of virtual machines (VMs), ranging from interactive desktops to test and development environments and even batch workloads. Hypervisors are responsible for multiplexing the underlying hardware resources among VMs while providing them the desired degree of isolation using resource management controls. Existing methods provide many knobs for allocating CPU and memory to VMs, but support for control of IO resource allocation has been quite limited. IO resource management in a hypervisor introduces significant new challenges and needs more extensive controls than in commodity operating systems.This paper introduces a novel algorithm for IO resource allocation in a hypervisor. Our algorithm, mClock, supports proportional-share fairness subject to minimum reservations and maximum limits on the IO allocations for VMs. We present the design of mClock and a prototype implementation inside the VMware ESX server hypervisor. Our results indicate that these rich QoS controls are quite effective in isolating VM performance and providing better application latency. We also show an adaptation of mClock (called dmClock) for a distributed storage environment, where storage is jointly provided by multiple nodes.

/pdf/mclock-handling-throughput-variability-for-hypervisor-io-5ch1bnor97.pdf

mClock: handling throughput variability for hypervisor IO scheduling

Storage consolidation is becoming an attractive paradigm for data organization because of the economies of sharing and the ease of centralized management. However, sharing of resources is viable only if applications can be isolated from each other. This work targets the problem of providing performance guarantees to an application irrespective of the behavior of other workloads. Application requirements are represented in terms of the average throughput, latency and maximum burst size. Most earlier schemes only do weighted bandwidth allocation; schemes that provide control of latency either cannot handle bursts or penalize applications for their own prior behavior, such as using spare capacity. Our algorithm pClock is based on arrival curves that intuitively capture the bandwidth and burst requirements of applications. We show analytically that an application following its arrival curve never misses its deadline. We have implemented pClock both in DiskSim and as a module in the Linux kernel 2.6. Our evaluation shows three important features of pClock: (1) benefits over existing algorithms; (2) efficient performance isolation and burst handling; and (3) the ability to allocate spare capacity to either speed up some applications or to a background utility, such as backup. pClock can be efficiently implemented in a system without much overhead.

pClock: an arrival curve based approach for QoS guarantees in shared storage systems

Speculative adders have attracted strong interest for reducing critical path delays to sub-logarithmic delays by exploiting the trade-offs between reliability and performance. Speculative adders also find use in the design of reliable variable latency adders, which combine speculation with error correction to achieve high performance for low area overhead over traditional adders. This paper describes speculative carry select addition (SCSA), a novel function speculation technique for the design of low error-rate speculative adders and low overhead, high performance, reliable variable latency adders. We develop an analytical model for the error rate of SCSA to facilitate both design exploration and convergence. We show that for an error rate of 0.01% (0.25%), SCSA-based speculative addition is 10% faster than the DesignWare adder with up to 43% (56%) area reduction. Further, on average, variable latency addition using SCSA-based speculative adders is 10% faster than the DesignWare adder with area requirements of -19% to 16% (-17% to 29%) for unsigned random (signed Gaussian) inputs.

High performance reliable variable latency carry select addition

An efficient multiversion access structure for a transaction-time database is presented. Our method requires optimal storage and query times for several important queries and logarithmic update times. Three version operations-inserts, updates, and deletes-are allowed on the current database, while queries are allowed on any version, present or past. The following query operations are performed in optimal query time: key range search, key history search, and time range view. The key-range query retrieves all records having keys in a specified key range at a specified time; the key history query retrieves all records with a given key in a specified time range; and the time range view query retrieves all records that were current during a specified time interval. Special cases of these queries include the key search query, which retrieves a particular version of a record, and the snapshot query which reconstructs the database at some past time. To the best of our knowledge no previous multiversion access structure simultaneously supports all these query and version operations within these time and space bounds. The bounds on query operations are worst case per operation, while those for storage space and version operations are (worst-case) amortized over a sequence of version operations. Simulation results show that good storage utilization and query performance is obtained.

/pdf/an-efficient-multiversion-access-structure-223ekhze2c.pdf

An efficient multiversion access structure

In-memory computing is gaining popularity as a means of sidestepping the performance bottlenecks of block storage operations. However, the volatile nature of DRAM makes these systems vulnerable to system crashes, while the need to continuously refresh massive amounts of passive memoryresident data increases power consumption. Emerging storage-class memory (SCM) technologies combine fast DRAM-like cache-line access granularity with the persistence of storage devices like disks or SSDs, resulting in potential 10x-100x performance gains, and low passive power consumption. This unification of storage and memory into a single directly-accessible persistent tier raises significant reliability and pro-grammability challenges. In this paper, we present SoftWrAP, an open-source framework for Software based Write-Aside Persistence. SoftWrAP provides lightweight atomicity and durability for SCM storage transactions, while ensuring fast paths to data in processor caches, DRAM, and persistent memory tiers. We use our framework to evaluate both handcrafted SCM-based microbenchmarks as well as existing applications, specifically the STX B+Tree library and SQLite database, backed by emulated SCM. Our results show significant benefits of SoftWrAP over existing methods such as undo logging and shadow copying, and can match non-atomic durable writes to SCM, thereby gaining atomic consistency almost for free.

/pdf/softwrap-a-lightweight-framework-for-transactional-support-j2o718s449.pdf

Peter Varman

Papers

mClock: handling throughput variability for hypervisor IO scheduling

pClock: an arrival curve based approach for QoS guarantees in shared storage systems

High performance reliable variable latency carry select addition

An efficient multiversion access structure

SoftWrAP: A lightweight framework for transactional support of storage class memory