The Experts below are selected from a list of 118752 Experts worldwide ranked by ideXlab platform
Robert Reiz - One of the best experts on this subject based on the ideXlab platform.
-
manoelcampos/cloudsim-plus: Vertical VM Scaling, Initial Host Fault Injection, Automatic generation of IDs ...
2017Co-Authors: Manoel Campos Da Silva Filho, Raysa Oliveira, Robert ReizAbstract:Added Vertical VM Scaling Mechanism (#7) for up and down scaling of VM resources such as Ram, Bandwidth and PEs (CPUs). double getUtilization() method in the UtilizationModel class to get the utilization percentage of a given resource at the Current Simulation Time. Methods getUtilizationOfRam(), getUtilizationOfBw(), getUtilizationOfCpu() added to Cloudlet in order to get utilization percentage of RAM, BW and CPU, respectively, for the Current Simulation Time. UtilizationModel.Unit enum to define the measuring unit in which a Cloudlet resource, to which a UtilizationModel is associated to, will be used. The enum values can be PERCENTAGE or ABSOLUTE, that respectively defines that the Cloudlet resource usage will be in percentage or absolute values. The existing UtilizationModels continue to define the value in percentage, as described in their documentation. The UtilizationModelDynamic (previously called UtilizationModelArithmeticProgression) allows setting a different unit for such an UtilizationModel (#62). UtilizationModelDynamic now allows defining the resource usage increment behavior using a Lambda Expression, enabling the developer to give a function that performs the increment in an arithmetic, geometric, exponential or any other kind of progression he/she needs (#64). Updated the DatacenterBroker interface and implementing classes, including the methods setVmComparator and setCloudletComparator to enable a developer to set a Comparator object (which can be given as a Lambda Expression) to sort VMs and Cloudlets before they are actually requested to be created in some Datacenter. This enables defining priorities to request the creation of such objects. If no Comparator is defined, the objects creation request follows the order in which they were submitted. Host Fault Injection Mechanism (under development) to enable injection of random failures into Hosts PEs: it injects failures into Host PEs and reallocates working PEs to running VMs. When all PEs from a Host fail, it starts clones of failed VMs to recovery from failure. This way, it is simulated the instantiation of VM snapshots into different Hosts (#81). Added the method Host.getWorkingPesList(). Poisson Distribution implementation enabling the Simulation of inter-arrival Times of events such as Host failures. Added the method void submitCloudletList(List
Manoel Campos Da Silva Filho - One of the best experts on this subject based on the ideXlab platform.
-
manoelcampos/cloudsim-plus: Vertical VM Scaling, Initial Host Fault Injection, Automatic generation of IDs ...
2017Co-Authors: Manoel Campos Da Silva Filho, Raysa Oliveira, Robert ReizAbstract:Added Vertical VM Scaling Mechanism (#7) for up and down scaling of VM resources such as Ram, Bandwidth and PEs (CPUs). double getUtilization() method in the UtilizationModel class to get the utilization percentage of a given resource at the Current Simulation Time. Methods getUtilizationOfRam(), getUtilizationOfBw(), getUtilizationOfCpu() added to Cloudlet in order to get utilization percentage of RAM, BW and CPU, respectively, for the Current Simulation Time. UtilizationModel.Unit enum to define the measuring unit in which a Cloudlet resource, to which a UtilizationModel is associated to, will be used. The enum values can be PERCENTAGE or ABSOLUTE, that respectively defines that the Cloudlet resource usage will be in percentage or absolute values. The existing UtilizationModels continue to define the value in percentage, as described in their documentation. The UtilizationModelDynamic (previously called UtilizationModelArithmeticProgression) allows setting a different unit for such an UtilizationModel (#62). UtilizationModelDynamic now allows defining the resource usage increment behavior using a Lambda Expression, enabling the developer to give a function that performs the increment in an arithmetic, geometric, exponential or any other kind of progression he/she needs (#64). Updated the DatacenterBroker interface and implementing classes, including the methods setVmComparator and setCloudletComparator to enable a developer to set a Comparator object (which can be given as a Lambda Expression) to sort VMs and Cloudlets before they are actually requested to be created in some Datacenter. This enables defining priorities to request the creation of such objects. If no Comparator is defined, the objects creation request follows the order in which they were submitted. Host Fault Injection Mechanism (under development) to enable injection of random failures into Hosts PEs: it injects failures into Host PEs and reallocates working PEs to running VMs. When all PEs from a Host fail, it starts clones of failed VMs to recovery from failure. This way, it is simulated the instantiation of VM snapshots into different Hosts (#81). Added the method Host.getWorkingPesList(). Poisson Distribution implementation enabling the Simulation of inter-arrival Times of events such as Host failures. Added the method void submitCloudletList(List
Raysa Oliveira - One of the best experts on this subject based on the ideXlab platform.
-
manoelcampos/cloudsim-plus: Vertical VM Scaling, Initial Host Fault Injection, Automatic generation of IDs ...
2017Co-Authors: Manoel Campos Da Silva Filho, Raysa Oliveira, Robert ReizAbstract:Added Vertical VM Scaling Mechanism (#7) for up and down scaling of VM resources such as Ram, Bandwidth and PEs (CPUs). double getUtilization() method in the UtilizationModel class to get the utilization percentage of a given resource at the Current Simulation Time. Methods getUtilizationOfRam(), getUtilizationOfBw(), getUtilizationOfCpu() added to Cloudlet in order to get utilization percentage of RAM, BW and CPU, respectively, for the Current Simulation Time. UtilizationModel.Unit enum to define the measuring unit in which a Cloudlet resource, to which a UtilizationModel is associated to, will be used. The enum values can be PERCENTAGE or ABSOLUTE, that respectively defines that the Cloudlet resource usage will be in percentage or absolute values. The existing UtilizationModels continue to define the value in percentage, as described in their documentation. The UtilizationModelDynamic (previously called UtilizationModelArithmeticProgression) allows setting a different unit for such an UtilizationModel (#62). UtilizationModelDynamic now allows defining the resource usage increment behavior using a Lambda Expression, enabling the developer to give a function that performs the increment in an arithmetic, geometric, exponential or any other kind of progression he/she needs (#64). Updated the DatacenterBroker interface and implementing classes, including the methods setVmComparator and setCloudletComparator to enable a developer to set a Comparator object (which can be given as a Lambda Expression) to sort VMs and Cloudlets before they are actually requested to be created in some Datacenter. This enables defining priorities to request the creation of such objects. If no Comparator is defined, the objects creation request follows the order in which they were submitted. Host Fault Injection Mechanism (under development) to enable injection of random failures into Hosts PEs: it injects failures into Host PEs and reallocates working PEs to running VMs. When all PEs from a Host fail, it starts clones of failed VMs to recovery from failure. This way, it is simulated the instantiation of VM snapshots into different Hosts (#81). Added the method Host.getWorkingPesList(). Poisson Distribution implementation enabling the Simulation of inter-arrival Times of events such as Host failures. Added the method void submitCloudletList(List
Jean Rouat - One of the best experts on this subject based on the ideXlab platform.
-
a parallel supercomputer implementation of a biological inspired neural network and its use for pattern recognition
Journal of Physics: Conference Series, 2012Co-Authors: Vincent De Ladurantaye, Jean Lavoie, Jocelyn Bergeron, Maxime Parenteau, Ramin Pichevar, Jean RouatAbstract:A parallel implementation of a large spiking neural network is proposed and evaluated. The neural network implements the binding by synchrony process using the Oscillatory Dynamic Link Matcher (ODLM). Scalability, speed and performance are compared for 2 implementations: Message Passing Interface (MPI) and Compute Unified Device Architecture (CUDA) running on clusters of multicore supercomputers and NVIDIA graphical processing units respectively. A global spiking list that represents at each instant the state of the neural network is described. This list indexes each neuron that fires during the Current Simulation Time so that the influence of their spikes are simultaneously processed on all computing units. Our implementation shows a good scalability for very large networks. A complex and large spiking neural network has been implemented in parallel with success, thus paving the road towards real-life applications based on networks of spiking neurons. MPI offers a better scalability than CUDA, while the CUDA implementation on a GeForce GTX 285 gives the best cost to performance ratio. When running the neural network on the GTX 285, the processing speed is comparable to the MPI implementation on RQCHP's Mammouth parallel with 64 notes (128 cores).
Vincent De Ladurantaye - One of the best experts on this subject based on the ideXlab platform.
-
a parallel supercomputer implementation of a biological inspired neural network and its use for pattern recognition
Journal of Physics: Conference Series, 2012Co-Authors: Vincent De Ladurantaye, Jean Lavoie, Jocelyn Bergeron, Maxime Parenteau, Ramin Pichevar, Jean RouatAbstract:A parallel implementation of a large spiking neural network is proposed and evaluated. The neural network implements the binding by synchrony process using the Oscillatory Dynamic Link Matcher (ODLM). Scalability, speed and performance are compared for 2 implementations: Message Passing Interface (MPI) and Compute Unified Device Architecture (CUDA) running on clusters of multicore supercomputers and NVIDIA graphical processing units respectively. A global spiking list that represents at each instant the state of the neural network is described. This list indexes each neuron that fires during the Current Simulation Time so that the influence of their spikes are simultaneously processed on all computing units. Our implementation shows a good scalability for very large networks. A complex and large spiking neural network has been implemented in parallel with success, thus paving the road towards real-life applications based on networks of spiking neurons. MPI offers a better scalability than CUDA, while the CUDA implementation on a GeForce GTX 285 gives the best cost to performance ratio. When running the neural network on the GTX 285, the processing speed is comparable to the MPI implementation on RQCHP's Mammouth parallel with 64 notes (128 cores).
