SNIC Documentation - User contributions [en]

Dalton CPP-LR parallelization

2015-03-27T10:29:41Z

Chandan Basu (NSC):

{{project info
|description=Parallelization of the coupled cluster complex polarization propagator module in the Dalton program

|fields=Computational science; Computational chemistry
|financing=SNIC
|active=No
|start date=2012-12-13
|end date=
}}

This project is the result of a request for application expert support. See here for the original request ([[Media:ccp-cc_paralllelization.pdf|PDF]]).

Requestors and collaborators:
* [https://www.ifm.liu.se/theomod/compphys/panor.xml Prof. Patrick Norrman, Computational Physics, IFM - Linköping University]
* [https://www.ifm.liu.se/theomod/compphys/joaka.xml Dr. Joanna Kauczor, Computational Physics, IFM - Linköping University]

Project is established in two phases.

Phase 1 (~10 work days):
* Meet with requestors to discuss the proposal {{ck|y}}
* Profile and analyze the code to figure out the time consuming parts {{ck|y}}
* Suggest a parallelization strategy {{ck|y}}

Report of work in phase 1 ([[Media:ccp-lr.pdf|PDF]]).

Phase 2:
* Work on the parallelization with the requestors

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

EC-Earth compilation and performance analysis: Beskow

2015-03-27T10:29:05Z

Chandan Basu (NSC):

{{project info
|description=Port EC-Earth to Beskow and complete performance/scaling tests.
|fields=Climate research
|financing=SNIC
|active=No
|start date=2014-11-16
|end date=2014-12-31
}}

The project aims allow early access to the EC-Earth model on Beskow for climate researchers by:
* Creating an EC-Earth machine specific configuration for Beskow with appropriate compiler settings.
* Completing a scaling and performance analysis of EC-Earth on Beskow.
* Proposing a set of recommendations (e.g. number of ranks per sub-model) for EC-Earth production runs on Beskow.

== Report ==
The porting work with CRAY compiler is completed. EC-EARTH portal is updated with the changes.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

PRACE WP12: Network topology analysis and efficient collective design

2015-03-27T10:28:27Z

Chandan Basu (NSC):

{{project info
|description=Network topology analysis and efficient collective design and

|fields=Computational science
|financing=PRACE
|active=No
|start date=2011-09-01
|end date=2013-08-31
}}

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

NorESM code analysis

2015-03-27T10:27:46Z

Chandan Basu (NSC):

{{project info
|description=Complete performance profiling of the Norwegian Earth System Model (NorESM) on Triolith.
|fields=Climate research
|financing=SNIC
|active=No
|start date=2014-10-01
|end date=2014-12-31
}}

The project aims to understand the performance of the NorESM
code. We will:
* Review the MPI pe mapping of the different sub-models (atmosphere, land, sea ice and ocean) currently used for model simulations.
* Analyze the NorESM code to identify code sections and routines to be further evaluated for performance optimization. We will do this by running suitable test cases through analysis tools such as Scalasca and VTune.
* Propose a set of recommendations for code optimization changes based on the performance analysis.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Report ==

The final [[Media:report.pdf|report]] with guidelines for performance improvement is submitted to the users

NorESM code analysis

2015-03-27T10:27:11Z

Chandan Basu (NSC): /* Report */

{{project info
|description=Complete performance profiling of the Norwegian Earth System Model (NorESM) on Triolith.
|fields=Climate research
|financing=SNIC
|active=Yes
|start date=2014-10-01
|end date=2014-12-31
}}

The project aims to understand the performance of the NorESM
code. We will:
* Review the MPI pe mapping of the different sub-models (atmosphere, land, sea ice and ocean) currently used for model simulations.
* Analyze the NorESM code to identify code sections and routines to be further evaluated for performance optimization. We will do this by running suitable test cases through analysis tools such as Scalasca and VTune.
* Propose a set of recommendations for code optimization changes based on the performance analysis.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Report ==

The final [[Media:report.pdf|report]] with guidelines for performance improvement is submitted to the users

File:Report.pdf

2015-03-27T10:22:25Z

Chandan Basu (NSC):

NorESM code analysis

2015-03-27T10:19:19Z

Chandan Basu (NSC): /* Members */

{{project info
|description=Complete performance profiling of the Norwegian Earth System Model (NorESM) on Triolith.
|fields=Climate research
|financing=SNIC
|active=Yes
|start date=2014-10-01
|end date=2014-12-31
}}

The project aims to understand the performance of the NorESM
code. We will:
* Review the MPI pe mapping of the different sub-models (atmosphere, land, sea ice and ocean) currently used for model simulations.
* Analyze the NorESM code to identify code sections and routines to be further evaluated for performance optimization. We will do this by running suitable test cases through analysis tools such as Scalasca and VTune.
* Propose a set of recommendations for code optimization changes based on the performance analysis.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Report ==

The final report with guidelines for performance improvement is submitted to the users

EC-Earth compilation and performance analysis: Beskow

2015-02-02T09:54:16Z

Chandan Basu (NSC):

{{project info
|description=Port EC-Earth to Beskow and complete performance/scaling tests.
|fields=Climate research
|financing=SNIC
|active=Yes
|start date=2014-11-16
|end date=2014-12-31
}}

The project aims allow early access to the EC-Earth model on Beskow for climate researchers by:
* Creating an EC-Earth machine specific configuration for Beskow with appropriate compiler settings.
* Completing a scaling and performance analysis of EC-Earth on Beskow.
* Proposing a set of recommendations (e.g. number of ranks per sub-model) for EC-Earth production runs on Beskow.

== Report ==
The porting work with CRAY compiler is completed. EC-EARTH portal is updated with the changes.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

User:Johan Raber (NSC)

2013-06-10T08:58:27Z

Chandan Basu (NSC): /* Projects */

{{application expert info
|first name=Johan
|last name=Raber
|centre=NSC
|fields=Computational chemistry
|fte=50
|snic ae financing=25
|other ae financing=
|financing=SNIC

|other activities=E-science coordinator at NSC.
|image=JohanR.png


}}

== Quick facts ==
* Worked at NSC as a Systems Expert, 2008-2010
* Since 2010 I work as the e-science coordinator at NSC
* Handled user relations, general support etc.
* Ph.D. in Biophysics in 2007 at Uppsala University
** Research Topic: Quantum chemical studies of an anti-cancer drug
** Tools: Gaussian, NWChem and GROMACS
* M.Sc. in Molecular Biotechnology Engineering
* Experience using [[GROMACS]] and [[Gaussian]] for QM/MD simulations

== Expertise ==
* [[expertise::Gaussian]]
* [[expertise::Gnuplot]]

== Projects ==
* [[project::NSC GPU and Accelerator Pilot]]
* [[project::PRACE WP12: Network topology analysis and efficient collective design]]
* [[project::Parallelization of a materials science code]]

User:Chandan Basu (NSC)

2013-06-10T08:57:20Z

Chandan Basu (NSC): /* Projects */

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::c]]
* [[expertise::TAU]]

== Projects ==
* [[project::Dalton CPP-LR parallelization]]
* [[project::PRACE WP12: Network topology analysis and efficient collective design]]
* [[project::Parallelization of a materials science code]]

== Old Projects ==
* [[project::Petascaling enabling and support for EC-EARTH3]]
* [[project::Improving MPI communication latency on euroben kernels]]
* [[project::Portability performance analysis and improvement of ESM]]

Parallelization of a materials science code

2013-06-10T08:55:13Z

Chandan Basu (NSC): Created page with "{{project info |description=Parallelization request for a materials science code  |fields=Computational science; Materials science |financing=SNIC |activ..."

{{project info
|description=Parallelization request for a materials science code

|fields=Computational science; Materials science
|financing=SNIC
|active=yes
|start date=2013-05-01
|end date=
}}

This project is the result of a request from a research grroup in LiU. The code is embarrassingly
parallel in nature. The parallelizing task is to send receive an array of user defined complicated
data type.

Requestors and collaborators:
* Olle Hellman @ IFM, LiU

Description

We have implemented an easy interface for MPI send / receive for derived data. The send/recv routines
"know" the data structure. These routines internally pack / unpack the structure data on a
character buffer. We have taken a data structure which is similar to what was described by
the user. We have implemented it in a template format. For the actual data structure the user
needs to modify the relevant places in pack and unpack routines. The data structure dependent
parts are placed in a separate module. In the main code the user needs to call
MPI_Send_point / MPI_Recv_point to send / recv derived data types.

The routine can be used as a stand alone library to be linked at run time. The interface can be
made more user friendly by "overloading" the MPI_Send_point / MPI_Recv_point routines as
MPI_Send / MPI_recv respectively.

[https://snicdocs.nsc.liu.se/w/images/e/ef/Mpi_derived.tgz The code]

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

File:Mpi derived.tgz

2013-06-10T08:54:46Z

Chandan Basu (NSC):

User:Chandan Basu (NSC)

2013-04-23T08:17:41Z

Chandan Basu (NSC):

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::c]]
* [[expertise::TAU]]

== Projects ==
* [[project::Dalton CPP-LR parallelization]]
* [[project::PRACE WP12: Network topology analysis and efficient collective design]]

== Old Projects ==
* [[project::Petascaling enabling and support for EC-EARTH3]]
* [[project::Improving MPI communication latency on euroben kernels]]
* [[project::Portability performance analysis and improvement of ESM]]

User:Chandan Basu (NSC)

2013-04-23T08:03:15Z

Chandan Basu (NSC): /* Projects */

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::c]]
* [[expertise::TAU]]

== Projects ==
* [[project::Petascaling enabling and support for EC-EARTH3]]
* [[project::Improving MPI communication latency on euroben kernels]]
* [[project::Portability performance analysis and improvement of ESM]]
* [[project::Dalton CPP-LR parallelization]]
* [[project::PRACE WP12: Network topology analysis and efficient collective design]]

PRACE WP12: Network topology analysis and efficient collective design

2013-04-23T08:02:31Z

Chandan Basu (NSC): Created page with "{{project info |description=Network topology analysis and efficient collective design and  |fields=Computational science |financing=PRACE |active=yes |s..."

{{project info
|description=Network topology analysis and efficient collective design and

|fields=Computational science
|financing=PRACE
|active=yes
|start date=2011-09-01
|end date=2013-08-31
}}

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

User:Chandan Basu (NSC)

2013-04-23T08:01:37Z

Chandan Basu (NSC): /* Projects */

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::c]]
* [[expertise::TAU]]

== Projects ==
* [[project::Petascaling enabling and support for EC-EARTH3]]
* [[project::Improving MPI communication latency on euroben kernels]]
* [[project::Portability performance analysis and improvement of ESM]]
* [[project::Dalton CPP-LR parallelization]]
* [[project::PRACE WP12: Efficient collective design and network topology analysis]]
* [[project::PRACE WP12: Network topology analysis and efficient collective design]]

User:Chandan Basu (NSC)

2013-04-23T07:40:59Z

Chandan Basu (NSC): /* Projects */

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::c]]
* [[expertise::TAU]]

== Projects ==
* [[project::Petascaling enabling and support for EC-EARTH3]]
* [[project::Improving MPI communication latency on euroben kernels]]
* [[project::Portability performance analysis and improvement of ESM]]
* [[project::Dalton CPP-LR parallelization]]
* [[project::PRACE WP12: Efficient collective design and network topology analysis]]

User:Chandan Basu (NSC)

2013-04-23T07:35:24Z

Chandan Basu (NSC): /* Expertise */

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::c]]
* [[expertise::TAU]]

== Projects ==
* [[project::Petascaling enabling and support for EC-EARTH3]]
* [[project::Improving MPI communication latency on euroben kernels]]
* [[project::Portability performance analysis and improvement of ESM]]
* [[project::Dalton CPP-LR parallelization]]

TAU

2012-09-24T08:25:41Z

Chandan Basu (NSC):

{{software info
|description=profiling and tracing tool-kit for performance analysis of parallel programs
|license=free
|fields=Performance optimisation; Parallel programming; Computational science
}}

{{PAGENAME}} Performance System is a portable {{#show: {{PAGENAME}} |?description}} written in Fortran, C, C++, Java, Python.

== Availability ==
{{list resources for software}}

== License ==
{{show license}}

== Presentations/Reviews ==
Overview [[Media:Tau.pdf|presentation]] on the capabilities of TAU, given by [[User:Chandan_Basu_(NSC)|Chandan Basu]] from [[NSC|Nsc]] to the SNIC [[Application_experts|application expert]] group meeting, 19 - 20 Sept, C3SE, Göteborg. Click [[Media:Tau.pdf|here]] to see the slides.

== Experts ==
{{list experts}}

== Links ==
* [http://www.cs.uoregon.edu/research/tau/home.php Official website]
* [http://www.cs.uoregon.edu/research/tau/tau-usersguide.pdf Manual]
* [http://tau.uoregon.edu/tau.ppt Tutorial]

TAU

2012-09-24T08:25:17Z

Chandan Basu (NSC):

{{software info
|description=profiling and tracing tool-kit for performance analysis of parallel programs
|license=free
|fields=Performance optimisation; Parallel programming; Computational science
}}

{{PAGENAME}} Performance System is a portable {{#show: {{PAGENAME}} |?description}} written in Fortran, C, C++, Java, Python.

== Availability ==
{{list resources for software}}

== License ==
{{show license}}

== Presentations/Reviews ==
Overview [[Media:Tau.pdf|presentation]] on the capabilities of TAU, given by [[User:Chandan_Basu_(NSC)|Chandan Basu]] from [[NSC|Nsc]] to the SNIC [[Application_experts|application expert]] group meeting, 19 - 20 Sept, C2SE, Göteborg. Click [[Media:Tau.pdf|here]] to see the slides.

== Experts ==
{{list experts}}

== Links ==
* [http://www.cs.uoregon.edu/research/tau/home.php Official website]
* [http://www.cs.uoregon.edu/research/tau/tau-usersguide.pdf Manual]
* [http://tau.uoregon.edu/tau.ppt Tutorial]

TAU

2012-09-24T08:22:53Z

Chandan Basu (NSC):

{{software info
|description=profiling and tracing tool-kit for performance analysis of parallel programs
|license=free
|fields=Performance optimisation; Parallel programming; Computational science
}}

{{PAGENAME}} Performance System is a portable {{#show: {{PAGENAME}} |?description}} written in Fortran, C, C++, Java, Python.

== Availability ==
{{list resources for software}}

== License ==
{{show license}}

== Presentations/Reviews ==
Overview [[Media:Tau.pdf|presentation]] on the capabilities of TAU, given by [[User:Chandan_Basu_(NSC)|Chandan Basu]] from [[NSC|Nsc]] to the SNIC [[Application_experts|application expert]] group, when they met in Göteborg. Click [[Media:Tau.pdf|here]] to see the slides.

== Experts ==
{{list experts}}

== Links ==
* [http://www.cs.uoregon.edu/research/tau/home.php Official website]
* [http://www.cs.uoregon.edu/research/tau/tau-usersguide.pdf Manual]
* [http://tau.uoregon.edu/tau.ppt Tutorial]

File:Tau.pdf

2012-09-24T08:21:17Z

Chandan Basu (NSC):

TAU

2012-09-24T08:13:56Z

Chandan Basu (NSC):

{{software info
|description=profiling and tracing tool-kit for performance analysis of parallel programs
|license=free
|fields=Performance optimisation; Parallel programming; Computational science
}}

{{PAGENAME}} Performance System is a portable {{#show: {{PAGENAME}} |?description}} written in Fortran, C, C++, Java, Python.

== Availability ==
{{list resources for software}}

== License ==
{{show license}}

== Presentations/Reviews ==
Overview [[Media:tau.pdf|presentation]] on the capabilities of TAU, given by [[User:Chandan_Basu_(NSC)|Chandan Basu]] from [[NSC|Nsc]] to the SNIC [[Application_experts|application expert]] group, when they met in Göteborg. Click [[Media:tau.pdf|here]] to see the slides.

== Experts ==
{{list experts}}

== Links ==
* [http://www.cs.uoregon.edu/research/tau/home.php Official website]
* [http://www.cs.uoregon.edu/research/tau/tau-usersguide.pdf Manual]
* [http://tau.uoregon.edu/tau.ppt Tutorial]

File:7.5.D Guidelines for D7.5.pdf

2012-03-12T13:45:11Z

Chandan Basu (NSC): uploaded a new version of "File:7.5.D Guidelines for D7.5.pdf"

File:7.5.D Guidelines for D7.5.pdf

2012-03-12T13:40:25Z

Chandan Basu (NSC): uploaded a new version of "File:7.5.D Guidelines for D7.5.pdf"

Improving MPI communication latency on euroben kernels

2012-03-09T13:12:41Z

Chandan Basu (NSC):

{{project info
|description=Improving the MPI collective performance by network aware communication
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
|border=1px solid black
}}

The goal of this project is to improve the performance of some euroben kernels. This project is part of PRACE-1ip, WP7.5. The euroben kernel is a set of synthetic benchmark codes which is developed by PRACE. This kernel is used for benchmarking PRACE systems. Some of these benchmarks, e.g., mod2f uses MPI_Alltoall based collective for doing fast Fourier transform. Generally the MPI_Alltoall scaling on many nodes is not good. The performance of alltoall operation can be improved by introducing network awareness in alltoall operation. This can be done by writing a modified alltoall and then calling that from the application code. We tested our modified alltoall on mod2f benchmark with good results

Status

* Started code modification. The basic algorithm will remain same. Will map the communication such that there is more communication within node than outside. [30/03/2011]

* Some initial results are obtained and some improvements are achieved [25/05/2011]

* The tests are run on Curie. The standard MPI_Alltoallv calls are repalced by topology aware MPI_Alltoallv_tuned calls. [16/01/2012]

* Contributed to deliverable D7.5: [[media:7.5.D_Guidelines_for_D7.5.pdf|report]] [28/02/2012]
* Writing final report

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

File:7.5.D Guidelines for D7.5.pdf

2012-03-08T14:39:45Z

Chandan Basu (NSC): uploaded a new version of "File:7.5.D Guidelines for D7.5.pdf"

Improving MPI communication latency on euroben kernels

2012-03-08T14:16:35Z

Chandan Basu (NSC):

{{project info
|description=Improving the MPI collective performance by network aware communication
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
|border=1px solid black
}}

The goal of this project is to improve the performance of some euroben kernels. This project is part of PRACE-1ip, WP7.5. The euroben kernel is a set of synthetic benchmark codes which is developed by PRACE. This kernel is used for benchmarking PRACE systems. Some of these benchmarks, e.g., mod2f uses MPI_Alltoall based collective for doing fast Fourier transform. Generally the MPI_Alltoall scaling on many nodes is not good. The performance of alltoall operation can be improved by introducing network awareness in alltoall operation. This can be done by writing a modified alltoall and then calling that from the application code. We tested our modified alltoall on mod2f benchmark with good results

Status

* Started code modification. The basic algorithm will remain same. Will map the communication such that there is more communication within node than outside. [30/03/2011]

* Some initial results are obtained and some improvements are achieved [25/05/2011]

* The tests are run on Curie. The standard MPI_Alltoallv calls are repalced by topology aware MPI_Alltoallv_tuned calls. [16/01/2012]

* Contributed to deliverable D7.5 [28/02/2012] [[File:7.5.D_Guidelines_for_D7.5.pdf]]
* Writing final report

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

Improving MPI communication latency on euroben kernels

2012-03-08T14:12:18Z

Chandan Basu (NSC):

{{project info
|description=Improving the MPI collective performance by network aware communication
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
|border=1px solid black
}}

The goal of this project is to improve the performance of some euroben kernels. This project is part of PRACE-1ip, WP7.5. The euroben kernel is a set of synthetic benchmark codes which is developed by PRACE. This kernel is used for benchmarking PRACE systems. Some of these benchmarks, e.g., mod2f uses MPI_Alltoall based collective for doing fast Fourier transform. Generally the MPI_Alltoall scaling on many nodes is not good. The performance of alltoall operation can be improved by introducing network awareness in alltoall operation. This can be done by writing a modified alltoall and then calling that from the application code. We tested our modified alltoall on mod2f benchmark with good results

Status

* Started code modification. The basic algorithm will remain same. Will map the communication such that there is more communication within node than outside. [30/03/2011]

* Some initial results are obtained and some improvements are achieved [25/05/2011]

* The tests are run on Curie. The standard MPI_Alltoallv calls are repalced by topology aware MPI_Alltoallv_tuned calls. [16/01/2012]

* Contributed to deliverable D7.5 [28/02/2012] {{pdf|7.5.D_Guidelines_for_D7.5.pdf‎|Contributed to deliverable D7.5}}

* Writing final report

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

File:7.5.D Guidelines for D7.5.pdf

2012-03-08T14:06:30Z

Chandan Basu (NSC):

Improving MPI communication latency on euroben kernels

2012-03-08T13:48:02Z

Chandan Basu (NSC):

{{project info
|description=Improving the MPI collective performance by network aware communication
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
|border=1px solid black
}}

The goal of this project is to improve the performance of some euroben kernels. This project is part of PRACE-1ip, WP7.5. The euroben kernel is a set of synthetic benchmark codes which is developed by PRACE. This kernel is used for benchmarking PRACE systems. Some of these benchmarks, e.g., mod2f uses MPI_Alltoall based collective for doing fast Fourier transform. Generally the MPI_Alltoall scaling on many nodes is not good. The performance of alltoall operation can be improved by introducing network awareness in alltoall operation. This can be done by writing a modified alltoall and then calling that from the application code. We tested our modified alltoall on mod2f benchmark with good results

Status

* Started code modification. The basic algorithm will remain same. Will map the communication such that there is more communication within node than outside. [30/03/2011]

* Some initial results are obtained and some improvements are achieved [25/05/2011]

* The tests are run on Curie. The standard MPI_Alltoallv calls are repalced by topology aware MPI_Alltoallv_tuned calls. [16/01/2012]
* Writing final report. [28/02/2012]

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

Improving MPI communication latency on euroben kernels

2012-03-06T15:28:43Z

Chandan Basu (NSC):

{{project info
|description=Improving the MPI collective performance by network aware communication
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
|border=1px solid black
}}

The goal of this project is to improve the performance of some euroben kernels. This project is part of PRACE-1ip, WP7.5. The euroben kernel is a set of synthetic benchmark codes which is developed by PRACE. This kernel is used for benchmarking PRACE systems. Some of these benchmarks, e.g., mod2f uses MPI_Alltoall based collective for doing fast Fourier transform. Generally the MPI_Alltoall scaling on many nodes is not good. The performance of alltoall operation can be improved by introducing network awareness in alltoall operation. This can be done by writing a modified alltoall and then calling that from the application code. We tested our modified alltoall on mod2f benchmark with good results

status

* Started code modification. The basic algorithm will remain same. Will map the communication such that there is more communication within node than outside. [30/03/2011]

* Some initial results are obtained and some improvements are achieved [25/05/2011]

* The tests are run on Curie. The standard MPI_Alltoallv calls are repalced by topology aware MPI_Alltoallv_tuned calls. [16/01/2012]
* Writing final report. [28/02/2012]

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

Portability performance analysis and improvement of ESM

2012-02-08T14:21:05Z

Chandan Basu (NSC):

{{project info
|description=Porting and performance analysis of earth system models (ESM) on different architectures
|financing=Other
|active=Yes
|start date=2009-09-01
|end date=2013-02-29
}}

This is a EU funded project called IS-ENES. The current prject is part of IS-ENES WP8 Task 2. This project is a collaboration between NSC and SMHI. The goal of this project is porting and performance analysis of different ESMs on different platforms. We have done porting and analysis of NEMO and EC-EARTH3 on different systems,e.g., Ekman and Lindgren at PDC and PRACE tier 0 system Curie.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

Petascaling enabling and support for EC-EARTH3

2012-02-08T14:20:19Z

Chandan Basu (NSC):

{{project info
|description=Petascaling of high resolution EC-EARTH on PRACE Tier-0 Curie System
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

The goal of this project is performance analysis and scaling improvement of the high-resolution configuration of the coupled EC-EARTH3. This project is part of PRACE-1ip WP7.2 and WP7.5. The high resolution version of the EC-EARTH is ported on Curie. The scaling is achieved upto 3500 cores.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

User:Chandan Basu (NSC)

2012-02-08T12:46:48Z

Chandan Basu (NSC):

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::TAU]]


{{#set:project=[[Petascaling enabling and support for EC-EARTH3]]}}
{{#set:project=[[Improving MPI communication latency on euroben kernels]]}}
{{#set:project=[[Portability performance analysis and improvement of ESM]]}}

Portability performance analysis and improvement of ESM

2012-02-08T12:45:03Z

Chandan Basu (NSC):

{{project info
|description=Porting and performance analysis of earth system models (ESM) on different architectures
|financing=Other
|active=Yes
|start date=2009-09-01
|end date=2013-02-29
}}

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Description ==
This is a EU funded project called IS-ENES. The current prject is part of IS-ENES WP8 Task 2. This project is a collaboration between NSC and SMHI. The goal of this project is porting and performance analysis of different ESMs on different platforms. We have done porting and analysis of NEMO and EC-EARTH3 on different systems,e.g., Ekman and Lindgren at PDC and PRACE tier 0 system Curie.

Portability performance analysis and improvement of ESM

2012-02-08T12:39:09Z

Chandan Basu (NSC): Created page with "{{project info |description=Porting and performance analysis of earth system models (ESM) on different architectures |financing=Other |active=Yes |start date=2009-09-01 |end dat..."

{{project info
|description=Porting and performance analysis of earth system models (ESM) on different architectures
|financing=Other
|active=Yes
|start date=2009-09-01
|end date=2013-02-29
}}

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Description ==
The goal of this project is porting and performance analysis of different ESMs on different platforms. This project is part of IS-ENES WP8 Task 2. We have done porting and analysis of NEMO and EC-EARTH3 on different systems,e.g., Ekman and Lindgren at PDC and PRACE tier 0 system Curie.

Improving MPI communication latency on euroben kernels

2012-02-08T12:20:25Z

Chandan Basu (NSC): /* Description */

{{project info
|description=Improving the MPI collective performance by network aware communication
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Description ==
The goal of this project is to improve the performance of some euroben kernels. This project is part of PRACE-1ip, WP7.5. The euroben kernel is a set of synthetic benchmark codes which is developed by PRACE. This kernel is used for benchmarking PRACE systems. Some of these benchmarks, e.g., mod2f uses MPI_Alltoall based collective for doing fast Fourier transform. Generally the MPI_Alltoall scaling on many nodes is not good. The performance of alltoall operation can be improved by introducing network awareness in alltoall operation. This can be done by writing a modified alltoall and then calling that from the application code. We tested our modified alltoall on mod2f benchmark with good results

Improving MPI communication latency on euroben kernels

2012-02-08T12:19:01Z

Chandan Basu (NSC): /* Description */

{{project info
|description=Improving the MPI collective performance by network aware communication
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Description ==
The goal of this project is to improve the performance of some euroben kernels. This project is part of PRACE-1ip, WP7.5. The euroben kernel is a set of synthetic benchmark codes which is developed by PRACE. This kernel is used for benchmarking PRACE systems. Some of these benchmarks, e.g., mod2f uses MPI_Alltoall based collective for doing fast Fourier transform. Generally the MPI_Alltoall scaling on large nodes is not good. The performance of alltoall operation can by introducing network awareness in alltoall operation. This can be done by writing a modified alltoall and then calling that from the application code. We tested our modified alltoall on mod2f benchmark with good results

User:Chandan Basu (NSC)

2012-02-08T12:06:03Z

Chandan Basu (NSC):

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::TAU]]


{{#set:project=[[Petascaling enabling and support for EC-EARTH3]]}}
{{#set:project=[[Improving MPI communication latency on euroben kernels]]}}

Improving MPI communication latency on euroben kernels

2012-02-08T12:04:38Z

Chandan Basu (NSC): Created page with "{{project info |description=Improving the MPI collective performance by network aware communication |financing=PRACE |active=Yes |start date=2010-07-01 |end date=2012-06-30 }} ..."

Petascaling enabling and support for EC-EARTH3

2012-02-08T11:17:02Z

Chandan Basu (NSC):

{{project info
|description=Petascaling of high resolution EC-EARTH on PRACE Tier-0 Curie System
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

== Description ==
The goal of this project is performance analysis and scaling improvement of the high-resolution configuration of the coupled EC-EARTH3. This project is part of PRACE-1ip WP7.2 and WP7.5. The high resolution version of the EC-EARTH is ported on Curie. The scaling is achieved upto 3500 cores.

User:Chandan Basu (NSC)

2012-02-08T11:09:18Z

Chandan Basu (NSC):

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|snic ae financing=25
|other ae financing=IS-ENES and PRACE
|financing=EU (75%), SNIC (25%)
|fields=Computational science
|general activities=Working on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::TAU]]


{{#set:project=[[Petascaling enabling and support for EC-EARTH3]]}}

Petascaling enabling and support for EC-EARTH3

2012-02-08T11:00:02Z

Chandan Basu (NSC): Created page with "{{project info |description=Petascaling of high resolution EC-EARTH on PRACE Tier-0 Curie System |financing=PRACE |active=Yes |start date=2010-07-01 |end date=2012-06-30 }} == ..."

Synthetic Benchmark on Curie

2012-02-07T13:15:40Z

Chandan Basu (NSC):

{{project info
|description=Synthetic Benchmark for PRACE Tier-0 Curie System
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

Synthetic Benchmark for [https://snicdocs.nsc.liu.se/wiki/PRACE PRACE] Tier-0 Curie System is a task at [https://snicdocs.nsc.liu.se/wiki/PRACE PRACE] 1IP project. It aims to benchmark the performance of a cluster-type [http://www-hpc.cea.fr/en/complexe/tgcc-curie.htm Curie] system which is one of largest HPC systems serviced by [https://snicdocs.nsc.liu.se/wiki/PRACE PRACE].

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

Synthetic Benchmark on Curie

2012-02-07T13:11:50Z

Chandan Basu (NSC):

{{project info
|description=Synthetic Benchmark for [https://snicdocs.nsc.liu.se/wiki/PRACE PRACE] Tier-0 Curie System
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

Synthetic Benchmark for PRACE Tier-0 Curie System is a task at PRACE 1IP project. It aims to benchmark the performance of a cluster-type [http://www-hpc.cea.fr/en/complexe/tgcc-curie.htm Curie] system which is one of largest HPC systems serviced by PRACE.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

Synthetic Benchmark on Curie

2012-02-07T13:08:38Z

Chandan Basu (NSC):

{{project info
|description=Synthetic Benchmark for PRACE Tier-0 Curie System
|financing=PRACE
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

Synthetic Benchmark for PRACE Tier-0 Curie System is a task at PRACE 1IP project. It aims to benchmark the performance of a cluster-type [http://www-hpc.cea.fr/en/complexe/tgcc-curie.htm Curie] system which is one of largest HPC systems serviced by PRACE.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

Synthetic Benchmark on Curie

2012-02-07T13:08:13Z

Chandan Basu (NSC):

{{project info
|description=Synthetic Benchmark for PRACE Tier-0 Curie System
|financing=PRACE Project
|active=Yes
|start date=2010-07-01
|end date=2012-06-30
}}

Synthetic Benchmark for PRACE Tier-0 Curie System is a task at PRACE 1IP project. It aims to benchmark the performance of a cluster-type [http://www-hpc.cea.fr/en/complexe/tgcc-curie.htm Curie] system which is one of largest HPC systems serviced by PRACE.

== Members ==
{{#ask: [[Category:Person]] [[project::{{PAGENAME}}]]
|?centre
|?role
|?field
}}

User:Chandan Basu (NSC)

2012-01-25T10:38:11Z

Chandan Basu (NSC):

User:Chandan Basu (NSC)

2011-10-06T08:03:18Z

Chandan Basu (NSC):

{{application expert info
|first name=Chandan
|last name=Basu
|centre=NSC
|fte=100
|financing=75% EU, 25% SNIC
|fields=Computational science
|general activities=Work on climate and weather codes;EU projects IS-ENES and PRACE.
|other activities=Collaborative work with SMHI and MISU; Benchmarking.
|image=chandan.jpg
|office=Room 179, House G; Campus Valla;Linköping University
|phone=013 282579 (office); 0736569357 (mobile)
}}

== Quick facts ==
* Ph.D. in Computational physics
* Working at NSC since 2009
* Working on benchmarking, optimization, and scaling issues
* Involved in [https://is.enes.org/ IS-ENES] and [http://www.prace-ri.eu/ PRACE]
* Involved in optimization of meteo-climatology codes EC-EARTH and CESM

== Expertise ==
* [[expertise::MPI]]
* [[expertise::OpenMP]]
* [[expertise::Fortran]]
* [[expertise::TAU]]