[Commits] [svn:einsteintoolkit] www/about/releases/ (Rev. 1021)

knarf at cct.lsu.edu knarf at cct.lsu.edu
Wed May 22 09:43:44 CDT 2013


User: knarf
Date: 2013/05/22 09:43 AM

Modified:
 /about/releases/
  ET_2013_05_announcement.php

Log:
 changelot

File Changes:

Directory: /about/releases/
===========================

File [modified]: ET_2013_05_announcement.php
Delta lines: +90 -83
===================================================================
--- about/releases/ET_2013_05_announcement.php	2013-05-22 08:32:44 UTC (rev 1020)
+++ about/releases/ET_2013_05_announcement.php	2013-05-22 14:43:44 UTC (rev 1021)
@@ -55,105 +55,112 @@
 
 The following changes lead to potential different numerical results or the need to modify optionlists:
 <ul>
- <li> Cactus: Move MPI support from flesh to a thorn; use new MPI_DIR variable to control use of MPI (see below)
- <li> McLachlan: Correct some errors in CCZ4, in particular the GammaShift parameter
- <li> McLachlan: Change the default value of apply_dissipation to "never".
+ <li> The MPI thorn, as well as several other external libraries, feature an improved
+      autodetection mechanism. Most typical installations don't have to set any options
+      to have installed libraries detected.
+ <li> The parsing code of some ccl files got re-written. Most existing files should
+      work as they did before, but there are a few corner-cases, e.g. using double
+      quotes around keywords that contain mathematical symbols.
 </ul>
 
 <!-- TODO -->
-<p><ul><li><b>Changes since last release</b></li></ul></p>
+<p><ul><li><b>Larger changes since last release</b></li></ul></p>
 <ul>
- <li> GetComponents:
- <ul>
-  <li> anonymous checkout is now used by default. Use --no-anonymous for authenticated checkouts
-  <li> no longer use "shallow git clones" by default
- </ul>
+ <li> New components:
+  <ul>
+   <li> EinsteinAnalysis/Outflow
+   <li> Carpet/CycleClock
+   <li> LSUThorns/PeriodicCarpet
+   <li> CactusTest/TestFortranCrayPointers
+   <li> CactusDoc/CoreDoc
+  </ul>
+ </li>
  <li> simfactory
  <ul>
   <li> updates of various machine configurations
-  <li> Replace large files by hard links during cleanup
-  <li> Basic configuration file for Ubuntu, Debian and Fedora operating systems.
+  <li> remove existing default allocations; use local configuration to use these
  </ul>
- <li> utils: syntax highlight for vim and emacs
  <li> Cactus flesh:
  <ul>
-  <li> grid functions may be padded, i.e. the driver may allocate more grid points than requested. The allocated size of grid functions can be accessed using cctk_ash. Note however, that this can be different from cctk_lsh. The latter is that part (size) of cctk_ash that is used within a simulation, i.e. that was set as simulation size via parameters. This is necessary for vectorization where the size of arrays needs to be a multiple of some number, but the simulation domain might not be.  There is no padding by default.
-  <li> correctly detect circular schedule items
-  <li> replace CCTK_RESTRICT by restrict, which is now defined by the flesh
-  <li> Change per thorn -DTHORN_IS_xxx to a per thorn -I bindings/include/xxx
-  <li> Correct error in determining dependencies
-  <li> have -reconfig target execute -cleandeps after reconfiguration
-  <li> add parameters to reverse un-enforced Cactus schedule ordering
-  <li> Move MPI support from flesh to a thorn
-  <li> Support various new Darwin minor versions
-  <li> increase default value of RELTOL to 1e-12
-  <li> Enforce that arrangement names are legal C identifiers
-  <li> Update Fortran API for CCTK_LOOP macros
-  <li> Add autoconf check whether C99 is supported (now required)
-  <li> New option -P (--exit-after-param-check) to exit after param check
-  <li> allow ${parfile} in parameter files to allow ${parfile}/subdir
+  <li> Do not buffer output from thorn configuration scripts
+  <li> Introduce OPTIONAL_IFACTIVE for thorn dependencies
+  <li> Automatically activate required thorns
+  <li> Support IBM Blue Gene compilers
+  <li> Change the parameter parser to Piraha
+  <li> Implement CCTK_Error and friends (marked as non-returning for the compiler)
+  <li> Replace Cactus complex number type with C/C++ complex numbers
+  <li> Workaround for compiler-bug with Intel compiler and 'restrict'.
+       In particular, for the current and recent Intel compilers (version 13),
+       the 'restrict' keyword is disabled in Cactus code by default, because
+       the compiler tends to produce wrong code otherwise.
  </ul>
  <li> Carpet
  <ul>
-  <li> Now lives in git again
-  <li> A parameter CarpetLib::interpolate_from_buffer_zones was introduced which, in order to get the correct behaviour (not to use the buffer zones for interpolations), should be set to "no". However, since this changes the way things had been done in the past and this change isn't well tested yet, the default is still "yes" (as like before the parameter was introduced). The default will very likely be changed for the next Einstein Toolkit release, after more testing.
-  <li> Added "balanced" domain decomposition
-  <li> Update setting CPU affinity
-  <li> change defaults of periodic_[xyz] to "yes" to match PUGH
-  <li> CarpetIOHDF5: add option checkpoint_every_divisor
-  <li> CarpetIOHDF5: teach the checkpointer to write index files
-  <li> CarpetIOHDF5: add support for index files to file reader
-  <li> LoopControl: Change printstats default to "no"
-  <li> CarpetInterp2: New ENO2 interpolator.
-  <li> Support more than one hierarchy of timers
-  <li> Add timers for refinement levels
-  <li> CarpetLib: Support higher-then 2nd order restriction operator for cell-centered meshes for Langrange prolongated grid functions.
+  <li> Also here: too many interesting fixes to mention all, some:
+  <li> CarpetIOHDF5: Sliced HDF5 output index file was not stored correctly
+  <li> Carpet: Correct load-balancing algorithm to really prefer splitting in the z direction
+  <li> CarpetRegrid2: Allow pseudo-2D domains
  </ul>
- <li> EinsteinExact: New arrangement, to eventually replace the Exact thorn. The thorns in this arrangement are generated by Kranc
+ <li> MoL: Allow skipping MoL's initial copy from past to current timelevel
+ <li> RotatingSymmetry90: Correct index calculation for "DD" tensors
+ <li> Formaline: add off-line tarball extraction utility
+ <li> ADMConstraints: support "new" interface to TmunuBase
+ <li> AHFinderDirect
+  <ul>
+   <li>Dont use recovered origin if AH was never found
+   <li>Dont use initial guess origins from parameters when tracking from gridscalar
+  </ul>
+ <li> Hydro_Analysis
+  <ul>
+   <li> add option to average the location of multiple identical maxima 
+   <li> add parameters to control which interpolator is used in GRHydro_Separation
+  </ul>
+ <li> HydroBase
+  <ul>
+   <li> Split temperature and entropy evolution methods
+   <li> Add gridfunctions and boilerplate for vector potential
+  </ul>
+ <li> TmunuBase: change the default of "support_old_CalcTmunu_mechanism" to "no"
+ <li> EOS_Omni
+  <ul>
+   <li> improve solution for density based on pressure, temperature, Y_e (but slow)
+   <li> remove poly_gamma_ini
+   <li> add support for cold tabulated EOS (P=P(rho)) that can be supplemented with a thermal gamma law
+   <li> add possibility to stich on a polytrope to the lower end of a tabulated cold EOS
+   <li> extrapolate to negative temperatures
+</ul>
+ <li> GRHydro
+  <ul>
+   <li> remove poly_gamma_ini
+   <li> make more paramters steerable
+   <li> remove any traces of non-functional Fisheye support
+   <li> Introduce variable atmopshere level / tolerance as function of radius. Off by default
+   <li> Implemented WENO5
+   <li> Added MP5 reconstruction method
+   <li> implemented H viscosity for HLLE solver to eliminate the carbuncle instability
+   <li> a lot of smaller fixes/extensions (too many to list here)
+  </ul>
+ <li> GRHydro_InitData
+  <ul>
+   <li> add "magnetized Bondi" to hydrobase::initial_bvec
+   <li> As with GRHydro: a lot of smaller fixes, too many to list here
+  </ul>
+ <li> SetMask_SphericalSurface: fix setup of multiple spheres
+ <li> HDF5: update to 1.8.10-patch1
+ <li> MPI: improved auto-detection, update to OpenMPI 1.6.4
+ <li> OpenSSL: update  to 1.0.1e
+ <li> hwloc: library dealing with low-level hardware configurations, version 1.7
+ <li> zlib: update to 1.2.8
+ <li> Vectors: major update
+ <li> EinsteinExact: Various, smaller updates (noteworthy, but too many)
+ <li> Kranc: Various, smaller updates (noteworthy, but too many)
  <li> McLachlan
- <ul>
-  <li> Correct some errors in CCZ4, in particular the GammaShift parameter
-  <li> Change the default value of apply_dissipation to "never".
- </ul>
- <li> ExternalLibraries: MPI is now treated like any other library on which flesh and thorns may depend, such as e.g. the GSL. It is thus configured via thorn ExternalLibraries/MPI, and not via the flesh any more. In particular, the option "MPI = ..." does not work any more and must not be used; instead, use e.g. "MPI_DIR = <path>" to point to an existing installation. You may also say "MPI_DIR = BUILD" to have Cactus build OpenMPI automatically, which is a convenient way of using MPI on a system where it may not be installed.
- <li> LocalInterp: The interpolation loop uses OpenMP
- <li> MoL: add multirate capabilities and three new multirate rk schemes
- <li> SphericalSurface: make maxnphi and maxntheta parameters steerable
- <li> NaNChecker: Add options to use restriction mask to ignore unused points, output more information about the failed points, use OpenMP during NaN check
- <li> Multipole: Implement Driscoll&Healy integration
- <li> HydroBase: schedule hydrobase_InitExcisionmask global-early loop-local
- <li> TmunuBase: support_old_CalcTmunu_mechanism is now deprecated and is likely to be removed soon
- <li> EOS_Omni: make more parameters steerable, add option read_table_on_single_process to have a named processor read the full table and MPI_Bcast it to the other processors
- <li> GRHydro
- <ul>
-  <li> C2P: resort to bisection if Newton-Raphson fails. This is only activated if user requested this
-  <li> introduce parameters GRHydro_Y_e_min and GRHydro_Y_e_max to fix issues with EOS tables that don't extend sufficiently far
-  <li> Fixes to atmosphere mask and the way we sync
-  <li> various fixes to ePPM
-  <li> Option to use the slow sector of multirate RK methods
-  <li> fixed issues with hot EOS treatment
-  <li> Option to reconstruct W*vel instead of vel
-  <li> use an early atmosphere reset to ensure past timelevels are initialized correctly
-  <li> various MHD fixes
- </ul>
- <li> Exact: openmp fixes
- <li> GRHydro_InitData: new poloidal magnetic field routine, Added option to smooth metric inside horizon, Alfven wave test
- <li> SummationByParts: improve OpenMP parallelization
- <li> PITTNull*: respect Cactus' IO::truncate parameter
- <li> AEIThorns/Trigger now part of the Einstein Toolkit
+  <ul>
+   <li> McLachlan_BSSN: fix the Hamiltonian constraint for CCZ4
+   <li> Fix the sign for 2nd and 6th order dissipation operators
+  </ul>
+ <li> 
 
-External libraries:
- <li> MPI: new - use this instead of the flesh mechanism (see above)
- <li> BLAS: update to 3.4.2
- <li> FFTW3: update to 3.3.2
- <li> HDF5: will likely not build fortran and C++ api for next release by default (but still does now)
- <li> LAPACK: update to 3.4.2
- <li> LORENE: don't use OpenMP unless it's enabled within Cactus
- <li> OpenSSL: improve search for existing library
- <li> zlib: update to 1.2.7
- <li> hwloc: new thorn for libhwloc
-</ul>
-
 <p><ul><li><b>How to upgrade from Ørsted (ET_2012_11)</b></li></ul></p>
 
 <p>To upgrade from the previous release, use GetComponents with the new thornlist to check out the new version.</p>



More information about the Commits mailing list