Vijay Moorthy - Academia.edu (original) (raw)

Papers by Vijay Moorthy

... Dhabaleswar K. Panda Debashis Basak Donglai Dai Ram Kesavan Rajeev Sivaram Mohammad Banikazem... more ... Dhabaleswar K. Panda Debashis Basak Donglai Dai Ram Kesavan Rajeev Sivaram Mohammad Banikazemi Vijay Moorthy ... DEBASHIS BASAK received the B.Tech degree in Computer Science and Engineering from the Indian Institute of Technology, N. Delhi, India, in ...

The IBM SP Switch-Connected NT cluster is one the newest clustering platforms available. In this ... more The IBM SP Switch-Connected NT cluster is one the newest clustering platforms available. In this paper, we discuss an experimental implementation of the Virtual Interface Architecture for this platform. We discuss different design issues involved in this implementation. In particular, we explain how the virtual-to-physical address translation can be implemented efficiently with a minimum Network Interface Card (NIC) memory requirement. We show how caching the VIA descriptors on the NIC can reduce the communication latency. We also present an efficient scheme for implementing the VIA door bells without any hardware support. A comprehensive performance evaluation study of the implementation is provided. The performance of the implemented VIA surpasses that of other existing software implementations of the VIA and is comparable to that of a hardware VIA implementation. The peak measured bandwidth for our system is observed to be 101.4 MBytes/s and the one-way latency for short messages is 18.2 microseconds. It is to be noted that the VIA implementation presented in this paper is not a part of any IBM product and no assumptions should be made regarding its availability as a product in the future

Clusters of workstations have emerged as a popular platform for parallel and distributed computin... more Clusters of workstations have emerged as a popular platform for parallel and distributed computing. Commodity high speed networks which are used to connect workstation clusters provide high bandwidth, but also have high latency. SCRAMNet is an extremely low latency replicated non-coherent shared memory network, so far used only for real-time applications. This paper reports our early experiences with using SCRAMNet for cluster computing. We have implemented a user-level zero-copy message passing protocol for SCRAMNet called the BillBoard Protocol (BBP). The one way latency for sending a 4-byte message between two nodes using the BBP is measured to be as low as 7.8 s. Since SCRAMNet supports hardware level replication of messages, it is possible to implement multicast with almost the same latency as point-to-point communication. Using the BBP, the latency for broadcasting short messages to 4 nodes is measured to be 10.1 s and the latency for a 4-node barrier is measured to be 37 s. We have also built an MPI library on top of the BBP which makes use of multicast support from the BBP. Our results demonstrate the potential of SCRAMNet as a high performance interconnect for building scalable workstation clusters supporting message passing.

tive alternative platform for high performance computing. Due to the commodity nature of workstat... more tive alternative platform for high performance computing. Due to the commodity nature of workstations and interconnects and due to the multiplicity of vendors and platforms, the NOW environments are being gradually redefined as Heterogeneous Networks of Workstations (HNOW) environments. This paper presents a new framework for implementing collective communication operations (as defined by the Message Passing Interface (MPI) standard) efficiently for the emerging HNOW environments. We first classify different types of heterogeneity in HNOW and then focus on one important characteristic: communication capabilities of workstations. Taking this characteristic into account, we propose two new approaches (Speed-Partitioned Ordered Chain (SPOC) and Fastest-Node First (FNF)) to implement collective communication operations with reduced latency. We also investigate methods for deriving optimal trees for broadcast and multicast operations. Generating such trees is shown to be computationally intensive. It is shown that the FNF approach, in spite of its simplicity, can deliver performance within 1% of the performance of the optimal trees. Finally, these new approaches are compared with the approach used in the MPICH implementation on experimental as well as on simulated testbeds. On a 24node existing HNOW environment with SGI workstations and ATM interconnection, our approaches reduce the latency of broadcast and multicast operations by a factor of up to 3:5 compared to the approach used in the existing MPICH implementation. On a 64-node simulated testbed, our approaches can reduce the latency of broadcast and multicast operations by a factor of up to 4:5. Thus, these results demonstrate that there is significant potential for our approaches to be applied towards designing scalable collective communication libraries for current and future generation HNOW environments.

In this paper, we present algorithms for efficiently implementing three collective communication ... more In this paper, we present algorithms for efficiently implementing three collective communication operations on reflective memory network clusters: Broadcast, Barrier Synchronization and All-Reduce. These algorithms have been implemented as an extension to the Bill-Board Protocol, a low-latency communication protocol for reflective memory networks that we previously developed [6,7]. Simulated performance results are presented for these algorithms. The performance of the algorithms is evaluated in detail. A comparative study with Myrinet, a popular point-to-point switched interconnect for clusters is also presented.

In this paper we present an efficient design for message passing over a reflective memory network... more In this paper we present an efficient design for message passing over a reflective memory network. First, we consider the attributes of reflective memory communication networks and the requirements to efficiently build message-passing functionality on these networks. We then introduce the Bill-Board Protocol, a lock-free protocol which provides low-latency send, receive, and multicast functionality to higher-level applications over reflective memory networks. The communication protocol and an implementation on SCRAMNet is described in detail. Lastly, the performance of this protocol is demonstrated.

Ndt & E International, 1992

Acoustic emission (AE) is potentially an ideal technique for health monitoring of large structure... more Acoustic emission (AE) is potentially an ideal technique for health monitoring of large structures due to the small number of sensors required and its high sensitivity. There has been much research conducted to characterize and provide qualitative understanding of the AE process in small specimens. Unfortunately, it is difficult to extend these results to real structures as the experimental data is dominated by geometric effects due to the small size of the specimens. The aim of this work is to provide a characterization of elastic waves emanating from fatigue cracks in plate-like structures. Fatigue crack growth was initiated in large 6082 T6 aluminium alloy plate specimens subjected to cyclic loading in the laboratory. A large specimen was used to eliminate signal reflections from the specimen edges and to enable signals from different wave modes to be separated in time. The signals were recorded using both resonant and non-resonant transducers attached to the surface of the specimens. Large numbers of AE signals were detected due to active fatigue crack propagation during the experiment. Analysis of experimental results from multiple crack growth events was used to characterize the modal and angular distributions of the radiated elastic waves. Experimental results are compared with finite element predictions to examine the mechanism of AE generation at the crack tip.

Ndt & E International, 1992

International Journal of Fatigue, 1999

A non-destructive, magnetic Barkhausen emission (MBE) technique has been used to assess various s... more A non-destructive, magnetic Barkhausen emission (MBE) technique has been used to assess various stages of low cycle fatigue (LCF) damage in 9Cr-1Mo ferritic steel. The initial decrease in the MBE peak height in the early stage of LCF cycling indicates the cyclic hardening stage, in which the formation of dislocation tangles reduces the mean free path of the domain wall movement. The increase in the MBE level again on further cycling indicates the progressive cyclic softening stage where the rearrangement of dislocation tangles into cells enhances the domain wall movement. The unaltered behaviour of MBE on continued cycling shows the saturation stage where the stabilization of dislocation substructure maintains the MBE level. Finally, a sharp increase in the MBE peak value identifies surface crack initiation and propagation, which is ascribed to the movement of additional reverse domains produced at the crack surfaces. This study establishes that the MBE technique can be used to assess the progressive degradation in the fatigue life of the ferritic steel components.

International Journal of Fatigue, 1999

International Journal of Pressure Vessels and Piping, 1995

The acoustic emission (AE) technique has been used to detect the microplastic yielding occurring ... more The acoustic emission (AE) technique has been used to detect the microplastic yielding occurring during macroscopic elastic deformation in an AISI Type 316 stainless steel. It has been observed that selection of different resonant frequency sensors is essential to detect the AE signal with maximum sensitivity at different strain levels during tensile deformation. An attempt has been made to develop a theoretical model to predict the approximate frequency range of the AE signal generated from dislocation sources operating during pre-yield and near-yield tensile deformation. The frequency of the AE signal has been calculated from the event life time of the Frank-Read and grain boundary source operations. The model for predicting the frequency of the AE signal from Frank-Read source operation during pre-yield deformation has been verified by the experiments on a nuclear grade AISI Type 316 stainless steel. This model has also been extended to predict the frequency of the AE signal from the grain boundary source operation near the macro-yield region and its validity has been verified by considering the AE results obtained on aluminium, copper and AISI Type 316 stainless steel by different investigators. This study has shown good agreement between the theoretically estimated and experimentally observed values. This study has established a simple, but reasonably accurate model which could help in selecting the resonant sensors with suitable frequency for detecting both the microplastic yielding and macroyielding with high sensitivity during proof testing of pressure vessels and pipes and other components used in various industries.

Ndt & E International, 2003

The effect of tempering on the magnetic Barkhausen noise (MBN) signal profile was studied in case... more The effect of tempering on the magnetic Barkhausen noise (MBN) signal profile was studied in case-carburised EN36 steel using a range of magnetic excitation frequencies and a number of frequency ranges for analysis of the MBN signal. The MBN level generally increases with tempering due to coarsening of the microstructure. With higher values of excitation frequency, f EX , the MBN profile exhibits a single peak, but with low values of f EX , double peaks are observed. The MBN peak obtained with higher f EX was found to correlate well with hardness changes in a region, down to 100 mm below the surface. The analysis of the MBN signal produced with low f EX , in narrow frequency ranges selected by software frequency filtering, showed variations in the extent of changes in the relative height of the two MBN peaks in the profile. After taking into account the skin depth-frequency relation for the MBN signal, variations in the values of the two MBN peaks in different analysing frequency ranges were found to correlate well with hardness variations at different depths down to 425 mm below the surface. An empirical relationship has been established between the hardness-depth profile and the MBN measurements. q

International Journal of Pressure Vessels and Piping, 1995

Ndt & E International, 2003

Ndt & E International, 2005

The effect of bending fatigue at different maximum stress levels on the magnetic Barkhausen emiss... more The effect of bending fatigue at different maximum stress levels on the magnetic Barkhausen emission (MBE) has been studied in case-carburised En36 steel specimens. The low frequency MBE profile has been measured after unloading the specimen at different number of fatigue cycles. It has been found that, beyond 1000 MPa, the MBE peak height decreases just after few thousand cycles and the percentage reduction in the MBE peak increases with maximum bending stress level. The reduction in MBE peak at lower stresses (<1400 MPa) is attributed mainly to the effect of residual stresses becoming more compressive below the surface due to the application prior tensile stress. At higher stresses (1500 MPa), the variation in the MBE peak also indicates the effect of cyclic hardening and softening with progressive fatigue cycles. The MBE profiles measured after monotonic loading and unloading with different maximum stress levels also show similar reduction in the MBE peak with increase in pre-stress level. However, at higher stresses (>1400 MPa), the cyclic loading shows larger reduction in the MBE peak than the monotonic loading. This is attributed to the effect of cyclic microplasticity induced enhancement of dislocation density in addition to the residual stress modification. This study clearly shows the MBE technique can be used to detect the maximum stress level seen by the specimen beyond 1000 MPa. Any overstressing of this case-carburised steel beyond the fatigue limit of 1150 MPa can be easily detected from the percentage reduction in the MBE peak. Since the crack propagation stage is insignificant in these hard steels, the detection of any overstressing using the MBE technique would be very useful in assessing and preventing the impending catastrophic failure.

Ndt & E International, 2005

Nondestructive Testing and Evaluation, 2008

The Magnetic Barkhausen Emission (MBE) technique is considered as a potential Non-Destructive Eva... more The Magnetic Barkhausen Emission (MBE) technique is considered as a potential Non-Destructive Evaluation (NDE) method for assessing material properties in case-hardened gears. For gear manufacturing industries, the evaluation of material properties such as tempering induced changes in hardness level and the associated microstuctural changes, variations in case-depth of hardened layer, grinding process induced alteration in microstructure, hardness and residual stress distributions are of primary importance in order to ensure the quality of gears during service. This paper presents an overview of different research and developmental studies carried out at Design Unit, Newcastle University on the MBE technique suitable for different applications in the evaluation of quality of gears. This paper gives an overview of different studies carried out at Design Unit, Newcastle University, to assess the feasibility of applying the MBE technique towards quality evaluation during different stages of gear manufacturing. The capabilities of the MBE

Ndt & E International, 1992

International Journal of Fatigue, 1999

International Journal of Pressure Vessels and Piping, 1995

Ndt & E International, 2003

International Journal of Pressure Vessels and Piping, 1995

Ndt & E International, 2003

Ndt & E International, 2005

Nondestructive Testing and Evaluation, 2008