Nowadays, there are several open-source solutions for building private, public and even hybrid clouds such as Eucalyptus, Apache Cloud Stack and Open Stack. KVM is one of the supported hypervisors for these cloud platforms. Different KVM configurations are being supplied by these platforms and, in some cases, a subset of CPU features are being presented to guest systems, providing a basic abstraction of the underlying CPU. One of the reasons for limiting the features of the Virtual CPU is to guarantee the guest compatibility with different hardware in heterogeneous environments. However, in a large number of situations, the cloud is deployed on an homogeneous set of hosts. In these cases, this limitation can affect the performance of applications being executed in guest systems. In this paper, we have analyzed the architecture, the KVM setup, and the performance of the Virtual Machines deployed by three popular cloud management platforms: Eucalyptus, Apache Cloud Stack and Open Stack, employing a representative set of applications.

An efficient implementation of the density-gradient (DG) approach for the finite element and finite difference methods and its application in drift-diffusion (D-D) simulations is described in detail. The new, second-order differential (SOD) scheme is compatible with relatively coarse grids even for large density variations thus applicable to device simulations with complex 3-D geometries. Test simulations of a 1-D metal-oxide semiconductor diode demonstrate that the DG approach discretized using our SOD scheme can be accurately calibrated against Schrödinger-Poisson calculations exhibiting lower discretization error than the previous schemes when using coarse grids and the same results for very fine meshes. 3-D test D-D simulations using the finite element method are performed on two devices: a 10 nm gate length double gate metal-oxide-semiconductor field-effect transistor (MOSFET) and a 40 nm gate length Tri-Gate fin field-effect transistor (FinFET). In 3-D D-D simulations, the SOD scheme is able to converge to physical solutions at high voltages even if the previous schemes fail when using the same mesh and equivalent conditions. The quantum corrected D-D simulations using the SOD scheme also converge with an atomistic mesh used for the 10 nm double gate MOSFET saving computational resources and can be accurately calibrated against the results from non-equilibrium Green's functions approach. Finally, the simulated ID-VGcharacteristics for the 40 nm gate length Tri-Gate are in an excellent agreement with experimental data.

The next substitution of the Enabling Grids for E-sciencE project (EGEE) in 2010 by the European Grid Initiative (EGI), where grid infrastructure of each country will be run by National Grid Initiatives (NGI), is giving a boost to the NGI development. In this context, the Spanish National Grid Initiative (es-NGI) is being developed by the e-science Spanish network. The es-NGI is developing virtual organizations where common area applications are associated. In this context, the MOSFET virtual organisation (VO-MOSFET) was born in 2009 to perform semiconductor device simulations using the es-NGI infrastructure. This virtual organization is supported by the es-NGI resource centres and it is developing a job submission and monitoring system independent of the grid middleware. In this paper a general description of the VO-MOSFET and some application level utilities of the job submission system are presented. Furthermore, a gridification example of a nanoelectronic simulation is presented proving the grid benefits for the field of nanoelectronic simulation.