2nd BSC International Doctoral Symposium, Barcelona, 5th-7th May, 2015
http://hdl.handle.net/2099/16423
2024-03-19T04:31:00ZUnderstanding scientific application's performance
http://hdl.handle.net/2099/16593
Understanding scientific application's performance
Tintó Prims, Oriol
Improve Earth System model's scalability in High
Performance Computing infrastructure is crucial to achieve
efficiency in upcoming supercomputers. We study how to
increase the scalability of the widely used Oceanic Model of the
Nucleus for European Modelling of the Ocean (NEMO).
2015-06-05T10:47:11ZTintó Prims, OriolImprove Earth System model's scalability in High
Performance Computing infrastructure is crucial to achieve
efficiency in upcoming supercomputers. We study how to
increase the scalability of the widely used Oceanic Model of the
Nucleus for European Modelling of the Ocean (NEMO).Methodology to select a I/O configuration (hardware resources and stack software) in cloud platform
http://hdl.handle.net/2099/16592
Methodology to select a I/O configuration (hardware resources and stack software) in cloud platform
Gómez-Sánchez, Pilar; Rexachs, Dolores; Méndez, Sandra; Luque, Emilio
Currently, there is an increasing interest about the cloud
platform by the High Performance Computing (HPC) community,
and the Parallel Input/Output (I/O) for High Performance Systems is
not an exception. However, in cloud platform increase the number of
parameters that user can select to I/O system. For this reason, we
propose a methodology to help the user to select a configuration.
2015-06-05T10:40:09ZGómez-Sánchez, PilarRexachs, DoloresMéndez, SandraLuque, EmilioCurrently, there is an increasing interest about the cloud
platform by the High Performance Computing (HPC) community,
and the Parallel Input/Output (I/O) for High Performance Systems is
not an exception. However, in cloud platform increase the number of
parameters that user can select to I/O system. For this reason, we
propose a methodology to help the user to select a configuration.DLP acceleration on general purpose cores
http://hdl.handle.net/2099/16591
DLP acceleration on general purpose cores
Duric, Milovan; Palomar Pérez, Óscar; Unsal, Osman Sabri; Cristal Kestelman, Adrián; Valero Cortés, Mateo
High-performance and power-efficient multimedia
computing drives the design of modern and increasingly utilized
mobile devices. State-of-the-art low power processors already utilize
chip multiprocessors (CMP) that add dedicated DLP accelerators
for emerging multimedia applications and 3D games. Such
heterogeneous processors deliver desired performance and efficiency
at the cost of extra hardware specialized accelerators. In this paper,
we propose dynamically-tuned vector execution (DVX) by morphing
one or more available cores in a CMP into a DLP accelerator. DVX
improves performance and power efficiency of the CMP, without
additional costs for dedicated accelerators.
2015-06-05T10:36:32ZDuric, MilovanPalomar Pérez, ÓscarUnsal, Osman SabriCristal Kestelman, AdriánValero Cortés, MateoHigh-performance and power-efficient multimedia
computing drives the design of modern and increasingly utilized
mobile devices. State-of-the-art low power processors already utilize
chip multiprocessors (CMP) that add dedicated DLP accelerators
for emerging multimedia applications and 3D games. Such
heterogeneous processors deliver desired performance and efficiency
at the cost of extra hardware specialized accelerators. In this paper,
we propose dynamically-tuned vector execution (DVX) by morphing
one or more available cores in a CMP into a DLP accelerator. DVX
improves performance and power efficiency of the CMP, without
additional costs for dedicated accelerators.Scaling irregular array-type reductions in OmpSs
http://hdl.handle.net/2099/16590
Scaling irregular array-type reductions in OmpSs
Ciesko, Jan; Mateo Bellido, Sergi; Teruel, Xavier; Beltran Querol, Vicenç; Martorell Bofill, Xavier; Badia Sala, Rosa Maria; Labarta Mancho, Jesús José
Array-type reductions represent a frequently occurring
algorithmic pattern in many scientific applications. A special case
occurs if array elements are accessed in a non-linear, often random
manner, which makes their concurrent and scalable execution
difficult. In this work we present a new approach that consists of
language- and runtime support to facilitate programming and
delivers high scalability on modern shared-memory systems for such
irregular array-type reductions. A reference implementation in
OmpSs, a task-parallel programming model, shows promising results
with speed-ups up to 15x on the Intel Xeon processor.
2015-06-05T10:33:16ZCiesko, JanMateo Bellido, SergiTeruel, XavierBeltran Querol, VicençMartorell Bofill, XavierBadia Sala, Rosa MariaLabarta Mancho, Jesús JoséArray-type reductions represent a frequently occurring
algorithmic pattern in many scientific applications. A special case
occurs if array elements are accessed in a non-linear, often random
manner, which makes their concurrent and scalable execution
difficult. In this work we present a new approach that consists of
language- and runtime support to facilitate programming and
delivers high scalability on modern shared-memory systems for such
irregular array-type reductions. A reference implementation in
OmpSs, a task-parallel programming model, shows promising results
with speed-ups up to 15x on the Intel Xeon processor.Hecuba: NoSql made easy
http://hdl.handle.net/2099/16589
Hecuba: NoSql made easy
Alomar, Guillem; Becerra Fontal, Yolanda; Torres Viñals, Jordi
Non-relational databases are nowadays a common
solution when dealing with huge data set and massive query work
load. These systems have been redesigned from scratch in order to
achieve scalability and availability at the cost of providing only a
reduce set of low-level functionality, thus forcing the client
application to take care of complex logics. As a solution, our
research group developed Hecuba, a set of tools and interfaces,
which aims to facilitate programmers with an efficient and easy
interaction with non-relational technologies.
2015-06-05T10:05:31ZAlomar, GuillemBecerra Fontal, YolandaTorres Viñals, JordiNon-relational databases are nowadays a common
solution when dealing with huge data set and massive query work
load. These systems have been redesigned from scratch in order to
achieve scalability and availability at the cost of providing only a
reduce set of low-level functionality, thus forcing the client
application to take care of complex logics. As a solution, our
research group developed Hecuba, a set of tools and interfaces,
which aims to facilitate programmers with an efficient and easy
interaction with non-relational technologies.Influence of temperature on the topological features of inner cavities in cytoglobin
http://hdl.handle.net/2099/16584
Influence of temperature on the topological features of inner cavities in cytoglobin
Seira, Constantí; Luque, F. Javier; Bidon-Chanal, Axel
Cytoglobin (Cygb) is a novel member of the globin family in man, but
there is no clear evidence about its biological function. Cygb exhibits
a highly complex ligand rebinding kinetics, which agrees with the
structural plasticity of the inner cavities and tunnels found in the
protein matrix. In this work we have examined the effect of
temperature on the topological features of Cygb. To this end, the
structural and dynamical properties of human Cygb are compared
with those determined for the Antarctic fish Chaenocephalus
aceratus. The results support a distinct temperature-dependence of
the topological features in the two proteins, suggesting different
adaptations to cold and warm environments.
2015-06-04T14:02:10ZSeira, ConstantíLuque, F. JavierBidon-Chanal, AxelCytoglobin (Cygb) is a novel member of the globin family in man, but
there is no clear evidence about its biological function. Cygb exhibits
a highly complex ligand rebinding kinetics, which agrees with the
structural plasticity of the inner cavities and tunnels found in the
protein matrix. In this work we have examined the effect of
temperature on the topological features of Cygb. To this end, the
structural and dynamical properties of human Cygb are compared
with those determined for the Antarctic fish Chaenocephalus
aceratus. The results support a distinct temperature-dependence of
the topological features in the two proteins, suggesting different
adaptations to cold and warm environments.Theoretical study on the activation mechanism of AMP-kinase by means of Molecular Dynamics Simulations
http://hdl.handle.net/2099/16583
Theoretical study on the activation mechanism of AMP-kinase by means of Molecular Dynamics Simulations
Estarellas, C.; Leal Moraes, G.; Castro Villalba, Antonio; Lameira, J.; Bidon-Chanal, A.; Luque, F. J.
Mammalian AMP-activated protein kinase (AMPK) is a
Ser108/Thr132 heterotrimeric enzyme complex (one catalytic subunit
a and two regulatory subunits b and g) with a key role as sensor in
the cellular energy homeostasis. This function confers AMPK a major
role in numerous metabolic disorders, such as type 2 diabetes,
obesity and cancer, and explains the progressive interest as a
therapeutic target. AMPK is regulated by several mechanisms
including indirect and direct activators, which show clear specificity
by a particular subunit. We have carried out a series of molecular
dynamic simulations of the apo and holo forms of AMPK to gain
insight into the mechanism of AMPK activation.
2015-06-04T13:53:57ZEstarellas, C.Leal Moraes, G.Castro Villalba, AntonioLameira, J.Bidon-Chanal, A.Luque, F. J.Mammalian AMP-activated protein kinase (AMPK) is a
Ser108/Thr132 heterotrimeric enzyme complex (one catalytic subunit
a and two regulatory subunits b and g) with a key role as sensor in
the cellular energy homeostasis. This function confers AMPK a major
role in numerous metabolic disorders, such as type 2 diabetes,
obesity and cancer, and explains the progressive interest as a
therapeutic target. AMPK is regulated by several mechanisms
including indirect and direct activators, which show clear specificity
by a particular subunit. We have carried out a series of molecular
dynamic simulations of the apo and holo forms of AMPK to gain
insight into the mechanism of AMPK activation.Mathematical representation of the Hardware Round-Robin Scheduler analytical model for single-ISA heterogeneous architectures
http://hdl.handle.net/2099/16582
Mathematical representation of the Hardware Round-Robin Scheduler analytical model for single-ISA heterogeneous architectures
Nemirovsky, Daniel; Markovic, Nikola; Unsal, Osman Sabri; Valero Cortés, Mateo; Cristal Kestelman, Adrián
2015-06-04T12:54:34ZNemirovsky, DanielMarkovic, NikolaUnsal, Osman SabriValero Cortés, MateoCristal Kestelman, AdriánA framework for multidimensional indexes on distributed and highly-available data stores
http://hdl.handle.net/2099/16581
A framework for multidimensional indexes on distributed and highly-available data stores
Cugnasco, Cesare; Becerra Fontal, Yolanda
No-relational databases are nowadays a common solution
when dealing with a huge data set and massive query workload.
These systems have been redesigned from scratch in order to achieve
scalability and availability at the cost of providing only a reduce set
of low-level functionality, thus forcing the client application to
implement complex logic. As a solution, our research group
developed Hecuba, a set of tools and interfaces, which aims to
facilitate developers with an efficient and painless interaction with
non-relational technologies.
This paper presents a part of Hecuba related to a particular missing
feature: multidimensional indexing. Our work focuses on the design
of architectures and the algorithms for providing multidimensional
indexing on a distributed database without compromising scalability
and availability.
2015-06-04T12:19:37ZCugnasco, CesareBecerra Fontal, YolandaNo-relational databases are nowadays a common solution
when dealing with a huge data set and massive query workload.
These systems have been redesigned from scratch in order to achieve
scalability and availability at the cost of providing only a reduce set
of low-level functionality, thus forcing the client application to
implement complex logic. As a solution, our research group
developed Hecuba, a set of tools and interfaces, which aims to
facilitate developers with an efficient and painless interaction with
non-relational technologies.
This paper presents a part of Hecuba related to a particular missing
feature: multidimensional indexing. Our work focuses on the design
of architectures and the algorithms for providing multidimensional
indexing on a distributed database without compromising scalability
and availability.Resource management for software defined data centers for heterogeneous infrastructures
http://hdl.handle.net/2099/16580
Resource management for software defined data centers for heterogeneous infrastructures
Carneiro do Amaral, Marcelo; Carrera Pérez, David
Software Defined Data Center (SDDC) provides more
resource management flexibility since everything is defined as a
software, including the network as Software Defined Network
(SDN).Typically, cloud providers overlook the network, which is
configured in static way. SDN can help to meet applications goals
with dynamic network configuration and provide best-efforts for QoS.
Additionally, SDDC might benefit by instead of be composed by
heavy Virtual Machines, use light-weight OS Containers. Despite the
advantages of SDDC and OS Containers, it brings more complexity
for resource provisioning. The goal of this project is to optimize the
management of container based workloads deployed on Software defined
Data Centers enabled with heterogeneous network fabrics
through the use of network-aware placement algorithms that are
driven by performance models.
2015-06-04T12:11:42ZCarneiro do Amaral, MarceloCarrera Pérez, DavidSoftware Defined Data Center (SDDC) provides more
resource management flexibility since everything is defined as a
software, including the network as Software Defined Network
(SDN).Typically, cloud providers overlook the network, which is
configured in static way. SDN can help to meet applications goals
with dynamic network configuration and provide best-efforts for QoS.
Additionally, SDDC might benefit by instead of be composed by
heavy Virtual Machines, use light-weight OS Containers. Despite the
advantages of SDDC and OS Containers, it brings more complexity
for resource provisioning. The goal of this project is to optimize the
management of container based workloads deployed on Software defined
Data Centers enabled with heterogeneous network fabrics
through the use of network-aware placement algorithms that are
driven by performance models.