A Novel Many Core Simulator for very Large Clusters running Multiple Applications

نویسندگان

  • Venkateswaran Nagarajan
  • Ram Srivatsa Kannan
  • Prashanth Thinakaran
  • Rajagopal Hariharan
  • Vinesh Srinivasan
  • Bharanidharan Vasudevan
  • Prashanth Manohar
  • AswinKumar Sridharan
  • Karthikeyan Palavedu Saravanan
  • Vignesh Adhinarayanan
  • Vignesh Veppur Sankaranarayanan
چکیده

Advent of multi-core designs has paved way towards breaking the Exa-scale performance barrier and hence it is been a major focus for modern computer architecture research. It is evident that future generation super computers would have hundreds of multi-core processors at the node level and proportionately a higher number at the cluster level too. Simulation has become a primary technique for evaluating the performance of new design proposals in computer architecture. These node architectures which are generally simulated for its correctness are scaled up to the cluster level. So, there arises a necessity to scale node level simulators to evaluate the metrics of cluster with the help of large scale simulation. Hence, a proactive paradigm shift from the conventional simulation techniques towards effective simulation of cluster level design and addressing its needs becomes inevitable. This research trend accentuates the need for developing a cycle accurate multi-core processor simulator that is scalable to cluster level and therefore we present a scalable simulator that couples a cycle-accurate node simulator with a generic supercomputer network model. At this context, there occurs a necessity towards simulating the node architecture consisting of functional units, network topology, memory hierarchy etc. with cycle accuracy since the Architectural simulators without cycle accuracy which examines the impact of low level system changes on application performance have not historically scaled well [1] [2]. For example these coarse-grained simulators skew dramatically of the order 100x when these changes are scaled up over hundreds of systems at the cluster level.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Robust Controller Design Based-on Aerodynamic Load Simulator Identification Driven by PMSM for Hardware-in-the-Loop Simulations

Aerodynamic load simulators generate the required time varying load to test the actuator’s performance in the laboratory. Electric Load Simulator (ELS) as one of variety of the dynamic load simulators should follows the rotation of the Under Test Actuator (UTA) and applies the desired torque to UTA’s rotor at the same time. In such a situation, a very large torque is imposed to the ELS from the...

متن کامل

Truncated Hepatitis B virus like nanoparticles: A novel drug delivery platform for cancer therapy

Nowadays, Nano-sized drug delivery systems have been studied extensively for theirpotential in cancer therapy. Various drug nanocarriers are being developed including liposomes, micelles, and Virus like nanoparticles (VLNPs). VLNPs offer many advantages for developing smart drug delivery systems due to their precise and repeated structures and relatively large cargo capacities. Truncated ...

متن کامل

Synthesis and Characterization of a Novel Fe3O4-SiO2@Gold Core-Shell Biocompatible Magnetic Nanoparticles for Biological and Medical Applications

Objectives: The study of core-shell magnetic nanoparticles has a wide range of applications because of the unique combination of the nanoscale magnetic core and the functional shell. Characterization and application of one important class of core-shell magnetic nanoparticles (MNPs), i.e., iron oxide core (Fe3O4/γ-Fe2O3) with a silica shell and outer of gold (Fe3O4-SiO2@Gold (FSG)) in Boron Neut...

متن کامل

Multi-connection and Multi-core Aware All-gather on Infiniband Clusters

MPI_Allgather is a collective communication operation that is intensively used in many scientific applications. Due to high data exchange volume in MPI_Allgather, efficient and scalable implementation of this operation is critical to the performance of scientific applications running on emerging multi-core clusters. Mellanox ConnectX is a modern InfiniBand host channel adapter that is able to s...

متن کامل

Parallel Branch Prediction on GPU Platform

Branch Prediction is a common function in nowadays microprocessor. Branch predictor is duplicated into multiple copies in each core of a multicore and many-core processor and makes prediction for multiple concurrent running programs respectively. To evaluate the parallel branch prediction in many-core processor, existed schemes generally use a parallel simulator running in CPU which does not ha...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2012