Users

users@lists.einsteintoolkit.org

3 participants
3984 discussions

cactus performance
by Jose Fiestas Iquira 29 Mar '12

29 Mar '12

Hello, Is there documentation about performance of Cactus ETK in large machines. I have some questions regarding best performance according to initial conditions, calculation time required, etc. If there are performance plots like Flops vs. Number of nodes would help me as well. Other main question I have is if Cactus scales in large machines. I am using McLachlan, but any other application would give me an idea of what I should expect for my runs in big machines. Thanks, Jose

6 13

ETK VM
by Dennis Castleberry 29 Mar '12

29 Mar '12

Dear Frank, and group, The current version of the Einstein Toolkit VM (along with instructions and bug notes) is located at <http://cct.lsu.edu/~dcastl2>. It's LZMA-compressed (my pipelines to the file server are unfortunately short-lived). Regards, Dennis Castleberry dcastl2(a)tigers.lsu.edu

3 8

shrinking of computational domain at recovery
by Kentaro Takami 29 Mar '12

29 Mar '12

Hi, I want to shrink the computational domain at recovery from checkpoint. For example, I want to change the domain from initial domain size: coordbase::xmin = 0.0 coordbase::xmax = 50.00 coordbase::ymin = 0.0 coordbase::ymax = 50.0 coordbase::zmin = 0.0 coordbase::zmax = 50.00 to coordbase::xmin = 0.0 coordbase::xmax = 30.00 coordbase::ymin = 0.0 coordbase::ymax = 30.0 coordbase::zmin = 0.0 coordbase::zmax = 30.00 Is it possible? Is there already such a mechanism? Thanks. Kentaro TAKAMI

2 2

working nautilus config (options) file! (now doing "large" runs)
by Scott Hawley 27 Mar '12

27 Mar '12

# feel free to tweak this and include in known architectures. # nautilus.nics.xsede.org CC = gcc CXX = g++ F77 = gfortran F90 = gfortran CPPFLAGS = -LMPICH_IGNORE_CXX_SEEK -I/opt/sgi/mpt/mpt-2.04/include CFLAGS = -std=gnu99 -O2 F77FLAGS = -O2 F90FLAGS = -O2 CXXFLAGS = -O2 # Note: -mp=nonuma switches off --exceptions OPENMP = yes CPP_OPENMP_FLAGS = -fopenmp FPP_OPENMP_FLAGS = -fopenmp C_OPENMP_FLAGS = -fopenmp CXX_OPENMP_FLAGS = -fopenmp F77_OPENMP_FLAGS = -fopenmp F90_OPENMP_FLAGS = -fopenmp #prepend-path PATH /opt/sgi/mpt/mpt-2.04/bin #prepend-path CPATH /opt/sgi/mpt/mpt-2.04/include #prepend-path FPATH /opt/sgi/mpt/mpt-2.04/include #prepend-path LD_LIBRARY_PATH /opt/sgi/mpt/mpt-2.04/lib # The default is already -Minfo=warn WARN = yes CPP_WARN_FLAGS = FPP_WARN_FLAGS = C_WARN_FLAGS = CXX_WARN_FLAGS = F90_WARN_FLAGS = F77_WARN_FLAGS = # The option -trapuv leads to "cccccccc" segmentation faults -- not sure # whether this is an error in Cactus or Intel. # -check all prints run-time warnings each time a temporary array is created # -fmudflap leads to internal compiler errors # Try: output_conversion, power, uninit, args DEBUG = no OTHER = -fp-stack-check -ftrapuv CPP_DEBUG_FLAGS = -g -DCARPET_DEBUG FPP_DEBUG_FLAGS = -g -DCARPET_DEBUG C_DEBUG_FLAGS = -g -O0 CXX_DEBUG_FLAGS = -g -O0 F77_DEBUG_FLAGS = -g -O0 F90_DEBUG_FLAGS = -g -O0 LIBS = gfortran LIB_DIR = /opt/sgi/mpt/mpt-2.04/lib BLAS_DIR = BUILD BLAS_LIBS = blas FFTW_DIR = BUILD GSL_DIR = BUILD HDF5_DIR = BUILD LIBSZ_DIR = BUILD IPM_DIR = /nics/b/home/eschnett/ipm_0.982 LAPACK = yes LAPACK_DIR = BUILD LAPACK_LIBS = lapack MPI = NATIVE NATIVE_MPI_INC_DIRS = /opt/mpt/3.4.2/xt/mpich2-gnu/include NATIVE_MPI_LIB_DIRS = /opt/mpt/3.4.2/xt/mpich2-gnu/lib /opt/xt-pe/2.2.41/lib /opt/mpt/3.4.2/xt/mpich2-pgi/lib /opt/xt-libsci/10.3.9/pgi/lib /opt/mpt/3.4.2/xt/sma/lib /opt/mpt/3.4.2/xt/util/lib /opt/mpt/3.4.2/xt/pmi/lib NATIVE_MPI_LIBS = mpi++ mpi PTHREADS = yes SSL_DIR = BUILD TEXMEX_DIR = /nics/c/home/hawleys/Cactus/arrangements/texmex

1 0

GetComponent requests "clean checkout"
by Luca Baiotti 26 Mar '12

26 Mar '12

Hello, I have a comment/request on this error message from GetComponents: > ./bin/GetComponents --update --verbose [...] Error: The URL for KrancNumericalTools/GenericFD has changed, please perform a clean checkout. After showing the below error message, GetComponents stopped. But maybe a user doesn't want to perform a clean checkout every time such a mismatch occurs. I suggest to change the error message so that it instructs the users to solve the problem (by manually checking out the new repository) and then rerun GetComponents. Even better, GetComponents could do something like renaming the old checkout "thorn" to "thorn.old" and checking out the new repository. In any case, it would be useful if this error could be made non-fatal, namely if one could continue to update of the other thorns even without the failed thorn. What do you think? Cheers, Luca

2 1

Einstein Toolkit Meeting
by Frank Loeffler 25 Mar '12

25 Mar '12

Hi Please consider joining the weekly Einstein Toolkit phone call at 10 am US central time on Mondays. As usual, you can find instructions how to join on the following web site: http://einsteintoolkit.org/community/support/ I short: the number is (+1) 225-578-4942 or (+1) 866-573-0359 and the conference id is 118682#. This week's meeting will be fully devoted to the spring ET workshop [1]. Everybody should already know what they are supposed to be doing, but I would like to hear how preparations are coming along. I expect the virtualbox environment to be ready by tomorrow as well. There will be no meeting on April 2nd due to a large number of the maintainers away for the April APS meeting and preparing the workshop. April 9th is Easter Monday, and while not observed in most parts of the US it is in a lot of other countries, and most of us will have seen each other anyway just a couple of days before that, so we will resume the usual schedule on Monday April 16th. Of course questions are as always welcome on the mailing list; so don't hesitate to report problems or request help. Also note for Monday that Europe (or at least large parts I guess) now also changed to daylight saving time, as did the US two weeks ago already. For most of us it should now be the 'usual' time difference again. [1] https://docs.einsteintoolkit.org/et-docs/ET_Workshop_Spring_2012 Frank Loeffler

1 0

FFTW build error
by Scott Hawley 25 Mar '12

25 Mar '12

Hi. I'm unable to build the latest Maxwell release anymore. Following the download instructions verbatim, I hit a snag when compiling. There seems to be a problem with FFTW3. I have it set in my options file to "BUILD", and what eventually happens is.... (below) Any suggestions? Thanks, Scott this is on Nautilus. libtool: compile: gcc -DHAVE_CONFIG_H -I. -I.. -I../simd -LMPICH_IGNORE_CXX_SEEK -I/opt/sgi/mpt/mpt- 2.04/include -std=c99 -MT timer.lo -MD -MP -MF .deps/timer.Tpo -c timer.c -o timer.o In file included from timer.c:29: cycle.h: In function ‘getticks’: cycle.h:226: error: ‘asm’ undeclared (first use in this function) cycle.h:226: error: (Each undeclared identifier is reported only once cycle.h:226: error: for each function it appears in.) cycle.h:226: error: expected ‘;’ before ‘volatile’ gmake[4]: *** [timer.lo] Error 1 gmake[3]: *** [all-recursive] Error 1 gmake[2]: *** [all] Error 2 + (( 2 )) + echo 'BEGIN ERROR' + echo 'Error while building FFTW3. Aborting.' + echo 'END ERROR' + exit 1 FFTW3 selected, but FFTW3_DIR not set. Checking some places... FFTW3 not found Building FFTW3... FFTW3 install directory, FFTW3_INSTALL_DIR, not set. Installing in the default configuration location . CST error 1: -> Configuration script for thorn FFTW3 returned exit code 1 Error message: 'Error while building FFTW3. Aborting.' -- Scott H. Hawley, Ph.D. Assistant Professor of Physics Belmont University. 615-460-6206. Sent from my iPad

2 1

ML and ADMConstraints
by Luca Baiotti 23 Mar '12

23 Mar '12

Hello, McLachlan has its own code to compute the violation of the constraints, but is thorn ADMConstraints also supposed to work with ML? I found that ML_ADMConstraints and ADMConstraints give results that differ of orders of magnitude. Best, Luca

4 5

Room share desired ET Workshop 3.-6. April
by Alex Vañó Viñuales 21 Mar '12

21 Mar '12

Hi all, first of all I apologize for using Users for this purpose, but I really need a quick answer regarding this issue. For those who are attending the Spring Einstein Toolkit Workshop 2012 ( https://docs.einsteintoolkit.org/et-docs/ET_Workshop_Spring_2012) and have not arranged their lodgings yet: I am looking for someone to share the room with. Given that the special rate at the Hampton Inn Atlanta finished yesterday, I propose to share a room in the Hotel Midtown Atlanta for the days 3.-6. April. (In reality I arrive on the 31. March, but don't think that anybody will be there that early). In case somebody was interested, please send me an email as soon as possible and in any case not later than Friday 23. March at 8:30 EDT (-0400). Thank you, Alex Vano-Vinuales -- -------------------------------------------------- Alex Vano Vinuales Departament de Fisica Edifici Mateu Orfila i Rotger Universitat de les Illes Balears Cra. de Valldemossa km. 7.5 07122 Palma de Mallorca Illes Balears, Spain Phone: +34-971-17-2536 --------------------------------------------------

1 0

Meeting Minutes
by Roland Haas 20 Mar '12

20 Mar '12

Present were: Roland, Frank, Erik, Tanja, Bruno, Peter, Steve, Eloisa, Philipp, Ian * Steve's suggestion to split Cactus included files that contain per-thorn information into individual files was discussed. Will go ahead and apply patch. Expect some glitches until everything is sorted out. ET paper: * deadline for changes is tomorrow * provide all used parameter files in EinsteinExamples repository (already in ET) * switched to EVO for author discussion of outstanding todo items * please work on your assigned items today Yours, Roland

2 1

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

Users