Energy Optimization of Memory Intensive Parallel Workloads

Energy consumption is an important concern in modern multicore processors. The energy consumed by a multicore processor during the execution of an application can be minimized by tuning the hardware state utilizing knobs such as frequency, voltage etc. The existing theoretical work on energy minimization using Global DVFS (Dynamic Voltage and Frequency Scaling), despite being thorough, ignores the time and the energy consumed by the CPU on memory accesses and the dynamic energy consumed by the idle cores. This article presents an analytical energy-performance model for parallel workloads that accounts for the time and the energy consumed by the CPU chip on memory accesses in addition to the time and energy consumed by the CPU on CPU instructions. In addition, the model we present also accounts for the dynamic energy consumed by the idle cores. The existing work on global DVFS for parallel workloads shows that using a single frequency for the entire duration of a parallel application is not energy optimal and that varying the frequency according to the changes in the parallelism of the workload can save energy. We present an analytical framework around our energy-performance model to predict the operating frequencies (that depend upon the amount of parallelism) for global DVFS that minimize the overall CPU energy consumption. We show how the optimal frequencies in our model differ from the optimal frequencies in a model that does not account for memory accesses. We further show how the memory intensity of an application affects the optimal frequencies.

Abeta oligomer toxicity inhibitor protects memory in models of synaptic toxicity

BACKGROUND AND PURPOSE:

Amyloid-ß (Aß) aggregation into synaptotoxic, prefibrillar oligomers is a major pathogenic event underlying the neuropathology of Alzheimer's disease (AD). The pharmacological and neuroprotective properties of a novel Aß aggregation inhibitor, SEN1269, were investigated on aggregation and cell viability and in test systems relevant to synaptic function and memory, using both synthetic Aß(1-42) and cell-derived Aß oligomers.
EXPERIMENTAL APPROACH:

Surface plasmon resonance studies measured binding of SEN1269 to Aß(1-42) . Thioflavin-T fluorescence and MTT assays were used to measure its ability to block Aß(1-42) -induced aggregation and reduction in cell viability. In vitro and in vivo long-term potentiation (LTP) experiments measured the effect of SEN1269 on deficits induced by synthetic Aß(1-42) and cell-derived Aß oligomers. Following i.c.v. administration of the latter, a complex (alternating-lever cyclic ratio) schedule of operant responding measured effects on memory in freely moving rats.
KEY RESULTS:

SEN1269 demonstrated direct binding to monomeric Aß(1-42) , produced a concentration-related blockade of Aß(1-42) aggregation and protected neuronal cell lines exposed to Aß(1-42) . In vitro, SEN1269 alleviated deficits in hippocampal LTP induced by Aß(1-42) and cell-derived Aß oligomers. In vivo, SEN1269 reduced the deficits in LTP and memory induced by i.c.v. administration of cell-derived Aß oligomers.
CONCLUSIONS AND IMPLICATIONS:

SEN1269 protected cells exposed to Aß(1-42) , displayed central activity with respect to reducing Aß-induced neurotoxicity and was neuroprotective in electrophysiological and behavioural models of memory relevant to Aß-induced neurodegeneration. It represents a promising lead for designing inhibitors of Aß-mediated synaptic toxicity as potential neuroprotective agents for treating AD.

Roles for the subiculum in spatial information processing, memory, motivation and the temporal control of behaviour

The subiculum is in a pivotal position governing the output of the hippocampal formation. Despite this, it is a rather under-explored and sometimes ignored structure. Here, we discuss recent data indicating that the subiculum participates in a wide range of neurocognitive functions and processes. Some of the functions of subiculum are relatively well-known-these include providing a relatively coarse representation of space and participating in, and supporting certain aspects of, memory (particularly in the dynamic bridging of temporal intervals). The subiculum also participates in a wide variety of other neurocognitive functions too. however. Much less well-known are roles for the subiculum, and particularly the ventral subiculum, in the response to fear, stress and anxiety, and in the generation of motivated behaviour (particularly the behaviour that underlies drug addiction and the response to reward). There is an emerging suggestion that the subiculum participates in the temporal control of behaviour. It is notable that these latter findings have emerged from a consideration of instrumental behaviour using operant techniques; it may well be the case that the use of the watermaze or similar spatial tasks to assess subicular function (on the presumption that its functions are very similar to the hippocampus proper) has obscured rather than revealed neurocognitive functions of subiculum. The anatomy of subiculum suggests it participates in a rather subtle fashion in a very broad range of functions, rather than in a relatively more isolated fashion in a narrower range of functions, as might be the case for

Valved Dataflow For FPGA Memory Hierarchy Synthesis

For modern FPGA, implementation of memory intensive processing applications such as high end image and video processing systems necessitates manual design of complex multilevel memory hierarchies incorporating off-chip DDR and onchip BRAM and LUT RAM. In fact, automated synthesis of multi-level memory hierarchies is an open problem facing high level synthesis technologies for FPGA devices. In this paper we describe the first automated solution to this problem.
By exploiting a novel dataflow application modelling dialect, known as Valved Dataflow, we show for the first time how, not only can such architectures be automatically derived, but also that the resulting implementations support real-time processing for current image processing application standards such as H.264. We demonstrate the viability of this approach by reporting the performance and cost of hierarchies automatically generated for Motion Estimation, Matrix Multiplication and Sobel Edge Detection applications on Virtex-5 FPGA.

On The Energy-Efficiency of Byte-Addressable Non-Volatile Memory

Non-Volatile Memory (NVM) technology holds promise to replace SRAM and DRAM at various levels of the memory hierarchy. The interest in NVM is motivated by the difficulty faced in scaling DRAM beyond 22 nm and, long-term, lower cost per bit. While offering higher density and negligible static power (leakage and refresh), NVM suffers increased latency and energy per memory access. This paper develops energy and performance models of memory systems and applies them to understand the energy-efficiency of replacing or complementing DRAM with NVM. Our analysis focusses on the application of NVM in main memory. We demonstrate that NVM such as STT-RAM and RRAM is energy-efficient for memory sizes commonly employed in servers and high-end workstations, but PCM is not. Furthermore, the model is well suited to quickly evaluate the impact of changes to the model parameters, which may be achieved through optimization of the memory architecture, and to determine the key parameters that impact system-level energy and performance.

The Multidirectional Memory of Charlie Hebdo

This article will discuss notions and concepts of remembering in the aftermath of the Charlie Hebdo attacks. Much has been written about the immediate response to the attacks, both commending the collective spirit of unity that defined the ‘marche républicaine’ of 11 January 2015, and criticising the alleged hypocrisy and cynicism of, most notably, the political figures that took to the streets that day, hand in hand. I will consider a selection of the memory practices that have emerged since then, notably on the anniversary of the event. This demonstration of memory provides key insights into the form and manner of remembering within a particular cultural group, but also reflects how the present moment is integral to our understanding of memory. The purpose of this article is to consider how official and non-official remembering of Charlie Hebdo can intertwine as well as pull in separate directions. A focus on the politics, the language, the aesthetics and the geography of commemorative activities in this article will enable an appreciation of the multidirectional character of remembering Charlie Hebdo.

Errorless learning and memory performance in schizophrenia

There is evidence that patients with schizophrenia have impaired explicit memory and intact implicit memory. The present study sought to replicate and extend that of O'Carroll et al. [O'Carroll, R.E., Russell, H.H., Lawrie, S.M. and Johnstone, E.C., 1999. Errorless learning and the cognitive rehabilitation of memory-impaired schizophrenic patients. Psychological Medicine 29, 105-112.] which reported that for memory-impaired patients with schizophrenia performance on a (cued) word recall task is enhanced using errorless learning techniques (in which errors are prevented during learning) compared to errorful learning (the traditional trial-and-error approach). Thirty patients with a DSM-IV diagnosis of schizophrenia and fifteen healthy controls (HC) participated. The Rivermead Behavioural Memory Test was administered and from their scores, the schizophrenic patients were classified as either memory-impaired (MIS), or memory-unimpaired (MUS). During the training phase two lists of words were learned separately, one using the errorless learning approach and the other using an errorful approach. Subjects were then tested for their recall of the words using cued recall. After errorful learning training, performance on word recall for the MIS group was impaired compared to the MUS and HC groups. However, after errorless learning training, no significant differences in performance were found between the three groups. Errorless learning may play an important role in remediation of cognitive deficits for patients with schizophrenia. (c) 2007 Elsevier Ireland Ltd. All rights reserved.

Patterns of neurocognitive impairment in first-episode bipolar disorder and schizophrenia.

BACKGROUND:
Researching psychotic disorders in unison rather than as separate diagnostic groups is widely advocated, but the viability of such an approach requires careful consideration from a neurocognitive perspective.
AIMS:
To describe cognition in people with bipolar disorder and schizophrenia and to examine how known causes of variability in individual's performance contribute to any observed diagnostic differences.
METHOD:
Neurocognitive functioning in people with bipolar disorder (n = 32), schizophrenia (n = 46) and healthy controls (n = 67) was compared using analysis of covariance on data from the Northern Ireland First Episode Psychosis Study.
RESULTS:
The bipolar disorder and schizophrenia groups were most impaired on tests of memory, executive functioning and language. The bipolar group performed significantly better on tests of response inhibition, verbal fluency and callosal functioning. Between-group differences could be explained by the greater proclivity of individuals with schizophrenia to experience global cognitive impairment and negative symptoms.
CONCLUSIONS:
Particular impairments are common to people with psychosis and may prove useful as endophenotypic markers. Considering the degree of individuals' global cognitive impairment is critical when attempting to understand patterns of selective impairment both within and between these diagnostic groups.

Memory retrieval as temporal discrimination

Temporal distinctiveness models of memory retrieval claim that memories are organised partly in terms of their positions along a temporal dimension, and suggest that memory retrieval involves temporal discrimination. According to such models the retrievability of memories should be related to the discriminability of their temporal distances at the time of retrieval. This prediction is tested directly in three pairs of experiments that examine (a) memory retrieval and (b) identification of temporal durations that correspond to the temporal distances of the memories. Qualitative similarities between memory retrieval and temporal discrimination are found in probed serial recall (Experiments 1 and 2), immediate and delayed free recall (Experiments 3 and 4) and probed serial recall of grouped lists (Experiments 5 and 6). The results are interpreted as consistent with the suggestion that memory retrieval is indeed akin to temporal discrimination. (C) 2008 Elsevier Inc. All rights reserved.

Memory-Centric Hardware Synthesis from Dataflow Models

Generation of hardware architectures directly from dataflow representations is increasingly being considered as research moves toward system level design methodologies. Creation of networks of IP cores to implement actor functionality is a common approach to the problem, but often the memory sub-systems produced using these techniques are inefficiently utilised. This paper explores some of the issues in terms of memory organisation and accesses when developing systems from these high level representations. Using a template matching design study, challenges such as modelling memory reuse and minimising buffer requirements are examined, yielding results with significantly less memory requirements and costly off-chip memory accesses.

SoC memory hierarchy derivation from dataflow graphs

Hardware synthesis from dataflow graphs of signal processing systems is a growing research area as focus shifts to high level design methodologies. For data intensive systems, dataflow based synthesis can lead to an inefficient usage of memory due to the restrictive nature of synchronous dataflow and its inability to easily model data reuse. This paper explores how dataflow graph changes can be used to drive both the on-chip and off-chip memory organisation and how these memory architectures can be mapped to a hardware implementation. By exploiting the data reuse inherent to many image processing algorithms and by creating memory hierarchies, off-chip memory bandwidth can be reduced by a factor of a thousand from the original dataflow graph level specification of a motion estimation algorithm, with a minimal increase in memory size. This analysis is verified using results gathered from implementation of the motion estimation algorithm on a Xilinx Virtex-4 FPGA, where the delay between the memories and processing elements drops from 14.2 ns down to 1.878 ns through the refinement of the memory architecture. Care must be taken when modeling these algorithms however, as inefficiencies in these models can be easily translated into overuse of hardware resources.

Figuring Abjection: The Slave Mother in the Early Creole Novel

While twentieth century Caribbean literature in French (particularly post-Césaire) has generated a large body of criticism, writing from the nineteenth century has been largely neglected. This article begins by contextualising the Creole novel of the early nineteenth century in cultural and historical terms, before proceeding to an analysis of two novels published in 1835 by Martinican authors: Outre-mer by Louis de Maynard, and Les Créoles by Jules Levilloux. In the few studies that exist, these texts have been read in opposition to each other in terms of their portrayal of the (male) mulatto; Levilloux has generally been considered the more progressive writer in this regard. However they are in fact in striking harmony in their depiction of the black mother, a figure (in both senses, as her physiognomy is central in her portrayal) who has until now been overlooked. For both writers, the elderly black mother is an abject and wretched creature. She has necessarily to be shown to be repulsive, filthy and morally hideous in old age in order to counteract the fascination she provokes, and to embody a phantasised repellent to the desires of the white male.

History, an Exit Strategy: The Retrofuture Fabulations of kara lynch

A cartographer constructs a map of an individual creative history, that of the American artist kara lynch, as it emerges in connection to a collective history of African American cultural expression. Positioning history as complex, dynamic systems of interwoven memory networks, the map follows lynch’s traversals through various “zones of cultural haunting”: places where collective memories made invisible through systematic processes of cultural erasure may be recovered and revived. Through these traversals, which are inspired by lynch’s “forever project” Invisible, the map covers such terrains as haunted narratives, mechanisms of abstraction and coding within African American media production, water as an informational technology, the distribution of memory in blood, the dialectics of materiality and immateriality that frame considerations of black subjectivity, and the possibility that place of music might not be the site of sound but instead the social production of memory.