This is an archive of the discontinued LLVM Phabricator instance.

Differential D36682

[test-suite] Adding miniAMR Benchmark
ClosedPublic

Authored by homerdin on Aug 14 2017, 8:12 AM.

Details

Reviewers
hfinkel
MatzeB
kristof.beyls
Commits
rOLDT311398: [test-suite] Adding the miniAMR Benchmark
rL311398: [test-suite] Adding the miniAMR Benchmark
Summary
Description:

miniAMR applies a stencil calculation on a unit cube computational domain,
which is divided into blocks. The blocks all have the same number of cells
in each direction and communicate ghost values with neighboring blocks. With
adaptive mesh refinement, the blocks can represent different levels of
refinement in the larger mesh. Neighboring blocks can be at the same level
or one level different, which means that the length of cells in neighboring
blocks can differ by only a factor of two in each direction. The calculations
on the variables in each cell is an averaging of the values in the chosen
stencil. The refinement and coarsening of the blocks is driven by objects
that are pushed through the mesh. If a block intersects with the surface
or the volume of an object, then that block can be refined. There is also
an option to uniformly refine the mesh. Each cell contains a number of
variables, each of which is evaluated indecently.

Links:

Web: https://mantevo.org/packages/
Github: https://github.com/Mantevo/miniAMR

When run on Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz:
compile_time: 33.7526 
exec_time: 0.4879 
Maximum resident set size (kbytes): 836784

Diff Detail

Repository
rL LLVM

Event Timeline

homerdin created this revision.Aug 14 2017, 8:12 AM
Herald added a subscriber: mgorny. · View Herald TranscriptAug 14 2017, 8:12 AM
hfinkel added reviewers: MatzeB, kristof.beyls.Aug 15 2017, 9:54 AM
hfinkel added inline comments.Aug 15 2017, 11:07 AM
MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/init.c
90 ↗(On Diff #110965)

This also depends on rand. See my comments on this in D36621 and D36626.

homerdin updated this revision to Diff 111234.Aug 15 2017, 1:01 PM

Substituted in the glib_compat_rand implementation to remove dependency on libc's rand.

Thanks of the feedback!
Brian

MatzeB requested changes to this revision.Aug 15 2017, 2:47 PM
  • Integration looks ok
  • The LICENSE file doesn't need an executable file mode
  • I would like to hear some comments on the license statements in the files.
MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/block.h
5–7 ↗(On Diff #111234)

See my comments in D36683

This revision now requires changes to proceed.Aug 15 2017, 2:47 PM
hfinkel edited edge metadata.Aug 15 2017, 3:56 PM
In D36682#842478, @homerdin wrote:

Substituted in the glib_compat_rand implementation to remove dependency on libc's rand.

Regarding the rand implementation, please see my comment in D36621.

Thanks of the feedback!
Brian

homerdin updated this revision to Diff 111393.Aug 16 2017, 12:00 PM
homerdin edited edge metadata.

I included the original README with an extra note at the top. I also updated the glibc_compat_rand implementation and changed the LICENSE permissions.

• dberlin added a subscriber: • dberlin.Aug 17 2017, 10:13 PM

TL;DR This statement is slightyl different, but also has no effect.
It is also a statement that is more or less required by statute/regulation.

Essentially, government contracting rules often require the government to mark what they think they have the rights/license For our purposes, it's irrelevant, the real license is the LGPL.
(IE whether the government or sandia has certain extra rights is irrelevant, the relevant rightsholders are releasing it under the LGPL)

I assigned copyright to the FSF for most of my work on GCC. In return, the FSF grants me a non-exclusive license to do whatever i want with that work.

So if i really wanted, i could write "Under the terms of my contract with the FSF, Daniel Berlin retains certain non-exclusive rights in this software" on the top of a bunch of gcc files i contribued, and it would be true.
It would be irrelevant to *anyone else*, because the license of the file (as agreed upon) is going to be GPL.

Writing that is just a statement of facts, it doesn't change the license for anyone else.

(It just so happens that these rights mean *I* do not have to respect that license, in the same way others would. But in practice, nobody else knows or cares)

• dberlin mentioned this in D36680: [test-suite] Adding Pathfinder Benchmark.Aug 17 2017, 10:13 PM
MatzeB accepted this revision.Aug 18 2017, 10:44 AM

Thanks for the clarification on the license. LGTM.

This revision is now accepted and ready to land.Aug 18 2017, 10:44 AM
Closed by commit rL311398: [test-suite] Adding the miniAMR Benchmark (authored by hfinkel). · Explain WhyAug 21 2017, 3:53 PM
This revision was automatically updated to reflect the committed changes.

Revision Contents

PathSize
test-suite/
trunk/
1 line
MultiSource/
Benchmarks/
DOE-ProxyApps-C/
2 lines
3 lines
miniAMR/
6 lines
167 lines
7 lines
296 lines
130 lines
388 lines
61 lines
609 lines
112 lines
16 lines
60 lines
152 lines
365 lines
172 lines
614 lines
61 lines
65 lines
678 lines
92 lines
303 lines
102 lines
134 lines
75 lines
55 lines

Diff 112081

test-suite/trunk/LICENSE.TXT

Show First 20 Lines • Show All 77 Lines • ▼ Show 20 Lines
SPASS: llvm-test/MultiSource/Applications/SPASS SPASS: llvm-test/MultiSource/Applications/SPASS
oggenc: llvm-test/MultiSource/Applications/oggenc oggenc: llvm-test/MultiSource/Applications/oggenc
JM: llvm-test/MultiSource/Applications/JM JM: llvm-test/MultiSource/Applications/JM
ALAC: llvm-test/MultiSource/Applications/ALAC ALAC: llvm-test/MultiSource/Applications/ALAC
ASC Sequoia: llvm-test/MultiSource/Benchmarks/ASC_Sequoia/AMGmk ASC Sequoia: llvm-test/MultiSource/Benchmarks/ASC_Sequoia/AMGmk
llvm-test/MultiSource/Benchmarks/ASC_Sequoia/IRSmk llvm-test/MultiSource/Benchmarks/ASC_Sequoia/IRSmk
llvm-test/MultiSource/Benchmarks/ASC_Sequoia/sphot llvm-test/MultiSource/Benchmarks/ASC_Sequoia/sphot
smg2000: llvm-test/MultiSource/Benchmarks/ASCI_Purple/SMG2000 smg2000: llvm-test/MultiSource/Benchmarks/ASCI_Purple/SMG2000
miniAMR: llvm-test/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR
XSBench: llvm-test/MultiSource/Benchmarks/DOE-ProxyApps-C/XSBench XSBench: llvm-test/MultiSource/Benchmarks/DOE-ProxyApps-C/XSBench
HPCCG: llvm-test/MultiSource/Benchmarks/DOE-ProxyApps-C++/HPCCG HPCCG: llvm-test/MultiSource/Benchmarks/DOE-ProxyApps-C++/HPCCG
PENNANT: llvm-test/MultiSource/Benchmarks/DOE-ProxyApps-C++/PENNANT PENNANT: llvm-test/MultiSource/Benchmarks/DOE-ProxyApps-C++/PENNANT
Fhourstones: llvm-test/MultiSource/Benchmarks/Fhourstones Fhourstones: llvm-test/MultiSource/Benchmarks/Fhourstones
Fhourstones-3.1: llvm-test/MultiSource/Benchmarks/Fhourstones-3.1 Fhourstones-3.1: llvm-test/MultiSource/Benchmarks/Fhourstones-3.1
McCat: llvm-test/MultiSource/Benchmarks/McCat McCat: llvm-test/MultiSource/Benchmarks/McCat
Olden: llvm-test/MultiSource/Benchmarks/Olden Olden: llvm-test/MultiSource/Benchmarks/Olden
OptimizerEval: llvm-test/MultiSource/Benchmarks/OptimizerEval OptimizerEval: llvm-test/MultiSource/Benchmarks/OptimizerEval
Show All 23 Lines

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/CMakeLists.txt

add_subdirectory(XSBench) add_subdirectory(XSBench)
add_subdirectory(miniAMR)

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/Makefile

# MultiSource/DOE-ProxyApps-C Makefile: Build all subdirectories automatically # MultiSource/DOE-ProxyApps-C Makefile: Build all subdirectories automatically
LEVEL = ../../.. LEVEL = ../../..
PARALLEL_DIRS = XSBench PARALLEL_DIRS = XSBench \
miniAMR
include $(LEVEL)/Makefile.programs include $(LEVEL)/Makefile.programs

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/CMakeLists.txt

set(PROG miniAMR)
list(APPEND CPPFLAGS -I.)
list(APPEND LDFLAGS -lm)
set(RUN_OPTIONS --nx 8 --ny 8 --nz 8 --num_tsteps 100)
llvm_multisource()

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/LICENSE

GNU LESSER GENERAL PUBLIC LICENSE
Version 3, 29 June 2007
Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/>
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
This version of the GNU Lesser General Public License incorporates
the terms and conditions of version 3 of the GNU General Public
License, supplemented by the additional permissions listed below.
0. Additional Definitions.
As used herein, "this License" refers to version 3 of the GNU Lesser
General Public License, and the "GNU GPL" refers to version 3 of the GNU
General Public License.
"The Library" refers to a covered work governed by this License,
other than an Application or a Combined Work as defined below.
An "Application" is any work that makes use of an interface provided
by the Library, but which is not otherwise based on the Library.
Defining a subclass of a class defined by the Library is deemed a mode
of using an interface provided by the Library.
A "Combined Work" is a work produced by combining or linking an
Application with the Library. The particular version of the Library
with which the Combined Work was made is also called the "Linked
Version".
The "Minimal Corresponding Source" for a Combined Work means the
Corresponding Source for the Combined Work, excluding any source code
for portions of the Combined Work that, considered in isolation, are
based on the Application, and not on the Linked Version.
The "Corresponding Application Code" for a Combined Work means the
object code and/or source code for the Application, including any data
and utility programs needed for reproducing the Combined Work from the
Application, but excluding the System Libraries of the Combined Work.
1. Exception to Section 3 of the GNU GPL.
You may convey a covered work under sections 3 and 4 of this License
without being bound by section 3 of the GNU GPL.
2. Conveying Modified Versions.
If you modify a copy of the Library, and, in your modifications, a
facility refers to a function or data to be supplied by an Application
that uses the facility (other than as an argument passed when the
facility is invoked), then you may convey a copy of the modified
version:
a) under this License, provided that you make a good faith effort to
ensure that, in the event an Application does not supply the
function or data, the facility still operates, and performs
whatever part of its purpose remains meaningful, or
b) under the GNU GPL, with none of the additional permissions of
this License applicable to that copy.
3. Object Code Incorporating Material from Library Header Files.
The object code form of an Application may incorporate material from
a header file that is part of the Library. You may convey such object
code under terms of your choice, provided that, if the incorporated
material is not limited to numerical parameters, data structure
layouts and accessors, or small macros, inline functions and templates
(ten or fewer lines in length), you do both of the following:
a) Give prominent notice with each copy of the object code that the
Library is used in it and that the Library and its use are
covered by this License.
b) Accompany the object code with a copy of the GNU GPL and this license
document.
4. Combined Works.
You may convey a Combined Work under terms of your choice that,
taken together, effectively do not restrict modification of the
portions of the Library contained in the Combined Work and reverse
engineering for debugging such modifications, if you also do each of
the following:
a) Give prominent notice with each copy of the Combined Work that
the Library is used in it and that the Library and its use are
covered by this License.
b) Accompany the Combined Work with a copy of the GNU GPL and this license
document.
c) For a Combined Work that displays copyright notices during
execution, include the copyright notice for the Library among
these notices, as well as a reference directing the user to the
copies of the GNU GPL and this license document.
d) Do one of the following:
0) Convey the Minimal Corresponding Source under the terms of this
License, and the Corresponding Application Code in a form
suitable for, and under terms that permit, the user to
recombine or relink the Application with a modified version of
the Linked Version to produce a modified Combined Work, in the
manner specified by section 6 of the GNU GPL for conveying
Corresponding Source.
1) Use a suitable shared library mechanism for linking with the
Library. A suitable mechanism is one that (a) uses at run time
a copy of the Library already present on the user's computer
system, and (b) will operate properly with a modified version
of the Library that is interface-compatible with the Linked
Version.
e) Provide Installation Information, but only if you would otherwise
be required to provide such information under section 6 of the
GNU GPL, and only to the extent that such information is
necessary to install and execute a modified version of the
Combined Work produced by recombining or relinking the
Application with a modified version of the Linked Version. (If
you use option 4d0, the Installation Information must accompany
the Minimal Corresponding Source and Corresponding Application
Code. If you use option 4d1, you must provide the Installation
Information in the manner specified by section 6 of the GNU GPL
for conveying Corresponding Source.)
5. Combined Libraries.
You may place library facilities that are a work based on the
Library side by side in a single library together with other library
facilities that are not Applications and are not covered by this
License, and convey such a combined library under terms of your
choice, if you do both of the following:
a) Accompany the combined library with a copy of the same work based
on the Library, uncombined with any other library facilities,
conveyed under the terms of this License.
b) Give prominent notice with the combined library that part of it
is a work based on the Library, and explaining where to find the
accompanying uncombined form of the same work.
6. Revised Versions of the GNU Lesser General Public License.
The Free Software Foundation may publish revised and/or new versions
of the GNU Lesser General Public License from time to time. Such new
versions will be similar in spirit to the present version, but may
differ in detail to address new problems or concerns.
Each version is given a distinguishing version number. If the
Library as you received it specifies that a certain numbered version
of the GNU Lesser General Public License "or any later version"
applies to it, you have the option of following the terms and
conditions either of that published version or of any later version
published by the Free Software Foundation. If the Library as you
received it does not specify a version number of the GNU Lesser
General Public License, you may choose any version of the GNU Lesser
General Public License ever published by the Free Software Foundation.
If the Library as you received it specifies that a proxy can decide
whether future versions of the GNU Lesser General Public License shall
apply, that proxy's public statement of acceptance of any version is
permanent authorization for you to choose that version for the
Library.

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/Makefile

LEVEL = ../../../..
PROG = miniAMR
CPPFLAGS = -I.
LDFLAGS = -lm
RUN_OPTIONS = --nx 8 --ny 8 --nz 8 --num_tsteps 100
include $(LEVEL)/MultiSource/Makefile.multisrc

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/README

**************************************************************************
LLVM Test-suite Note:
**************************************************************************
The original source is located at https://github.com/Mantevo/miniAMR.
Beyond this paragraph is the original README contained with the source
code. The Makefile refered to within is not utilized within the
test-suite. The test-suite builds a serial version (openmp and
mpi disabled) with its own cmake and make build system.
**************************************************************************
miniAMR mini-application
--------------------------------------
Contents of this README file:
1. miniAMR overview
2. miniAMR versions
3. building miniAMR
4. running miniAMR
5. notes about the code
--------------------------------------
--------------------------------------
1. miniAMR overview
miniAMR applies a stencil calculation on a unit cube computational domain,
which is divided into blocks. The blocks all have the same number of cells
in each direction and communicate ghost values with neighboring blocks. With
adaptive mesh refinement, the blocks can represent different levels of
refinement in the larger mesh. Neighboring blocks can be at the same level
or one level different, which means that the length of cells in neighboring
blocks can differ by only a factor of two in each direction. The calculations
on the variables in each cell is an averaging of the values in the chosen
stencil. The refinement and coarsening of the blocks is driven by objects
that are pushed through the mesh. If a block intersects with the surface
or the volume of an object, then that block can be refined. There is also
an option to uniformly refine the mesh. Each cell contains a number of
variables, each of which is evaluated indepently.
--------------------------------------
2. miniAMR versions:
- miniAMR_ref:
reference version: self-contained MPI-parallel.
- miniAMR_serial
serial version of reference version
-------------------
3. Building miniAMR:
To make the code, type 'make' in the directory containing the source.
The enclosed Makefile.mpi is configured for a general MPI installation.
Other compiler or other machines will need changes in the CFLAGS
variable to correspond with the flags available for the compiler being used.
-------------------
4. Running miniAMR:
miniAMR can be run like this:
% <mpi-run-command> ./miniAMR.x
where <mpi-run-command> varies from system to system but usually looks something like 'mpirun -np 4 ' or similar.
Execution is then driven entirely by the default settings, as configured in default-settings.h. Options may be listed using
% ./miniAMR.x --help
To run the program, there are several arguments on the command line.
The list of arguments and their defaults is as follows:
--nx - block size in x
--ny - block size in y
--nz - block size in z
These control the size of the blocks in the mesh. All of these need to
be even and greater than zero. The default is 10 for each variable.
--init_x - initial blocks in x
--init_y - initial blocks in y
--init_z - initial blocks in z
These control the number of the blocks on each processor in the
initial mesh. These need to be greater than zero. The default
is 1 block in each direction per processor. The initial mesh
is a unit cube regardless of the number of blocks.
--reorder - ordering of blocks
This controls whether the blocks are ordered by the RCB algorithm
or by a natural ordering of the processors. The default is 1 which
selects the RCB ordering and the natural ordering is 0.
--npx - number of processors in the x direction
--npy - number of processors in the y direction
--npz - number of processors in the z direction
These control the number of processors is each direction. The product
of these number has to equal the number of processors being used. The
default is 1 block in each direction.
--max_blocks - maximun number of blocks per processor
The maximun number of blocks used per processor. This is the number of
blocks that will be allocated at the start of the run and the code will
fail if this number is exceeded. The default is 500 blocks.
--num_refine - number of levels of refinement
This is the number of levels of refinement that blocks which are refined
will be refined to. If it is zero then the mesh will not be refined.
the default is 5 levels of refinement.
--block_change - number of levels a block can change during refinement
This parameter controls the number of levels that a block can change
(either refining or coarsening) during a refinement step. The default
is the number of levels of refinement.
--uniform_refine - if 1, then grid is uniformly refined
This controls whether the mesh is uniformly refined. If it is 1 then the
mesh will be uniformly refined, while if it is zero, the refinement will
be controlled by objects in the mesh. The default is 1.
--refine_freq - frequency (in timesteps) of checking for refinement
This determines the frequency (in timesteps) between checking if
refinement is needed. The default is every 5 timesteps.
--target_active - target number of blocks per processor
--target_max - max number of blocks per processor
--target_min - min number of blocks per processor
These allow the user to control the number of blocks per processor.
If these are zero, then no adjustment is made. If target_active is
greater than zero than the code will adjust the number of blocks to
that target after the refinement step. If target_max is greater than
zero then the number of blocks will be reduced if it exceeds this
number. Likewise, if target_min is greater than zero, than the number
of blocks will be raised if there is less than that number after the
refinement step. The default for all of these is zero.
--inbalance - percentage inbalance to trigger inbalance
This parameter allows the user to set a percentage threshold above
which the load will be balanced amoung the processors. The value
that this is checked against is the maximum number of blocks on a
processor minus the minimum number of blocks on a processor divided
by the average. The default is zero, which means to always load
balance at each refinement step.
--lb_opt - (0, 1, 2) determine load balance strategy
If set to 0, then load balancing is not performed. The default is
set to 1 which load balances each refinement step. Setting the
parameter to 2 results in load balancing at each stage of the
refinement step. If a processor has a large number of blocks which
are refined several steps, this allows the work (and space needed)
to be shared amoung more processors.
--num_vars - number of variables (> 0)
The number of variables the will be calculated on and communicated.
The default is 40 variables.
--comm_vars - number of vars to communicate together
The number of variables that will communicated together. This will
allow shorter but more variables if it is set to something less than
the total number of variables. The default is zero which will
communicate all of the variables at once.
--num_tsteps - number of timesteps (> 0)
The number of timesteps for which the simulation will be run. The
default is 20.
--stages_per_ts - number of comm/calc stages per timestep
The number of calculate/communicate stages per timestep. The default
is 20.
--permute - (no argument) permute communication directions
If this is set, then the order of the communication directions will
be permuted through the six options available. The default is
to send messages in the x direction first, then y, and then z.
--blocking_send - (no argument) Use blocking sends in the communication
routine instead of the default nonblocking sends.
--code - change the way communication is done
The default is 0 which communicates only the ghost values that are
needed. Setting this to 1 sends all of the ghost values, and setting
this to 2 also does all of the message processing (refinement or
unrefinement) to be done on the sending side. This allows us to
more closely minic the communication behaviour of codes.
--checksum_freq - number of stages between checksums
The number of stages between calculating checksums on the variables.
The default is 5. If it is zero, no checks are performed.
--stencil - 7 or 27 point 3D stencil
The 3D stencil used for the calculations. It can be either 7 or 27
and the default is 7 since the 27 point calculation will not conserve
the sum of the variables except for the case of uniform refinement.
--error_tol - (e^{-error_tol} ; >= 0)
This determines the error tolerance for the checksums for the variables.
the tolerance is 10 to the negative power of error_tol. The default
is 8, so the default tolerance is 10^(-8).
--report_diffusion - (>= 0) none if 0
This determines if the checksums are printed when they are calculated.
The default is 0, which is no printing.
--report_perf - (0 .. 15)
This determines how the performance output is displayed. The default
is YAML output (value of 1). There are four output modes and each is
controlled by a bit in the value. The YAML output (to a file called
results.yaml) is controlled by the first bit (report_perf & 1), the
text output file (results.txt) is controlled by the second bit
(report_perf & 2), the output to standard out is controlled by the
third bit (report_perf & 4), and the output of block decomposition
at each refine step is controlled by the forth bit (report_perf & 8).
These options can be combined in any way desired and zero to four
of these options can be used in any run. Setting report_perf to 0
will result in no output.
--refine_freq - frequency (timesteps) of refinement (0 for none)
This determines how frequently (in timesteps) the mesh is checked
and refinement is done. The default is every 5 timesteps. If
uniform refinement is turned on, the setting of refine_freq does
not matter and the mesh will be refined before the first timestep.
--refine_ghosts - (no argument)
The default is to not use the ghost cells of a block to determine if
that block will be refined. Specifying this flag will allow those
ghost cells to be used.
--num_objects - (>= 0) number of objects to cause refinement
The number of objects on which refinement is based. Default is zero.
--object - type, position, movement, size, size rate of change
The object keyword has 14 arguments. The first two are integers
and the rest are floating point numbers. They are:
type - The type of object. There is 16 types of objects. They include
the surface of a rectangle (0), a solid rectangle (1),
the surface of a spheroid (2), a solid spheroid (3),
the surface of a hemispheroid (+/- with 3 cutting planes)
(4, 6, 8, 10, 12, 14),
a solid spheroid (+/- with 3 cutting planes)(5, 7, 9, 11, 13, 15),
the surface of a cylinder (20, 22, 24),
and the volume of a cylinder (21, 23, 25).
bounce - If this is 1 then an object will bounce off of the walls
when the center hits an edge of the unit cube. If it is
zero, then the object can leave the mesh.
center - Three doubles that determine the center of the object in the
x, y, and z directions.
move - Three doubles that determine the rate of movement of the center
of the object in the x, y, and z directions. The object moves
this far at each timestep.
size - The initial size of the object in the x, y, and z directions.
If any of these become negative, the object will not be used
in the calculations to determine refinement. These sizes are
from the center to the edge in the specified direction.
inc - The change in size of the object in the x, y, and z directions.
Examples of run scripts for a Cray XE6 that illustrate several of the options:
One sphere moving diagonally on 27 processors:
mpirun -np 27 -N 7 miniAMR.x --num_refine 4 --max_blocks 9000 --npx 3 --npy 3 --npz 3 --nx 8 --ny 8 --nz 8 --num_objects 1 --object 2 0 -1.71 -1.71 -1.71 0.04 0.04 0.04 1.7 1.7 1.7 0.0 0.0 0.0 --num_tsteps 100 --checksum_freq 1
An expanding sphere on 64 processors:
mpirun -np 64 miniAMR.x --num_refine 4 --max_blocks 6000 --init_x 1 --init_y 1 --init_z 1 --npx 4 --npy 4 --npz 4 --nx 8 --ny 8 --nz 8 --num_objects 1 --object 2 0 -0.01 -0.01 -0.01 0.0 0.0 0.0 0.0 0.0 0.0 0.0009 0.0009 0.0009 --num_tsteps 200 --comm_vars 2
Two moving spheres on 16 processors:
mpirun -np 16 miniAMR.x --num_refine 4 --max_blocks 4000 --init_x 1 --init_y 1 --init_z 1 --npx 4 --npy 2 --npz 2 --nx 8 --ny 8 --nz 8 --num_objects 2 --object 2 0 -1.10 -1.10 -1.10 0.030 0.030 0.030 1.5 1.5 1.5 0.0 0.0 0.0 --object 2 0 0.5 0.5 1.76 0.0 0.0 -0.025 0.75 0.75 0.75 0.0 0.0 0.0 --num_tsteps 100 --checksum_freq 4 --stages_per_ts 16
-------------------
5. The code:
block.c Routines to split and recombine blocks
check_sum.c Calculates check_sum for the arrays
comm_block.c Communicate new location for block during refine
comm.c General routine to do interblock communication
comm_parent.c Communicate refine/unrefine information to parents/children
comm_refine.c Communicate block refine/unrefine to neighbors during refine
comm_util.c Utilities to manage communication lists
driver.c Main driver
init.c Initialization routine
main.c Main routine that reads command line and launches program
move.c Routines that check overlap of objects and blocks
pack.c Pack and unpack blocks to move
plot.c Write out block information for plotting
profile.c Write out performance data
rcb.c Load balancing routines
refine.c Routines to direct refinement step
stencil.c Perform stencil calculations
target.c Add/subtract blocks to reach a target number
util.c Utility routines for timing and allocation
-- End README file.
Courtenay T. Vaughan
(ctvaugh@sandia.gov)

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/block.h

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
typedef struct {
int number;
int level;
int refine;
int new_proc;
int parent; // if original block -1,
// else if on node, number in structure
// else (-2 - parent->number)
int parent_node;
int child_number;
int nei_refine[6];
int nei_level[6]; /* 0 to 5 = W, E, S, N, D, U; use -2 for boundary */
int nei[6][2][2]; /* negative if off processor (-1 - proc) */
int cen[3];
double ****array;
} block;
block *blocks;
typedef struct {
int number;
int level;
int parent; // -1 if original block
int parent_node;
int child_number;
int refine;
int child[8]; // n if on node, number if not
// if negative, then onnode child is a parent (-1 - n)
int child_node[8];
int cen[3];
} parent;
parent *parents;
typedef struct {
int number; // number of block
int n; // position in block array
} sorted_block;
sorted_block *sorted_list;
int *sorted_index;
int my_pe;
int num_pes;
int max_num_blocks;
int target_active;
int target_max;
int target_min;
int num_refine;
int uniform_refine;
int x_block_size, y_block_size, z_block_size;
int num_vars;
int comm_vars;
int init_block_x, init_block_y, init_block_z;
int reorder;
int npx, npy, npz;
int inbalance;
int refine_freq;
int report_diffusion;
int checksum_freq;
int stages_per_ts;
int error_tol;
int num_tsteps;
int stencil;
int report_perf;
int plot_freq;
int lb_opt;
int block_change;
int code;
int permute;
int nonblocking;
int refine_ghost;
int max_num_parents;
int num_parents;
int max_active_parent;
int cur_max_level;
int *num_blocks;
int *block_start;
int num_active;
int max_active_block;
int global_active;
int x_block_half, y_block_half, z_block_half;
double tol;
double *grid_sum;
int *p8, *p2;
int mesh_size[3];
int max_mesh_size;
int *from, *to;
int msg_len[3][4];
int local_max_b;
int global_max_b;
int num_objects;
typedef struct {
int type;
int bounce;
double cen[3];
double orig_cen[3];
double move[3];
double orig_move[3];
double size[3];
double orig_size[3];
double inc[3];
} object;
object *objects;

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/block.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdio.h>
#include <stdlib.h>
#include "block.h"
#include "proto.h"
#include "timer.h"
// This routine splits blocks that are being refined into 8 new blocks,
// copies the data over to the new blocks, and then disconnects the
// original block from the mesh and connects the new blocks to the
// mesh. The information in old block is also transferred to a parent
// block, which also contains information to identify its children.
void split_blocks(void)
{
static int side[6][2][2] = { { {0, 2}, {4, 6} }, { {1, 3}, {5, 7} },
{ {0, 1}, {4, 5} }, { {2, 3}, {6, 7} },
{ {0, 1}, {2, 3} }, { {4, 5}, {6, 7} } };
static int off[6] = {1, -1, 2, -2, 4, -4};
int i, j, k, m, n, o, v, nl, xp, yp, zp, c, c1, other,
i1, i2, j1, j2, k1, k2, dir, fcase, pe, f, p,
level, sib[8], offset, d, half_size;
block *bp, *bp1;
parent *pp;
// have to do this by level, else could have a block at a level split and
// its offspring try to connect to a block at a lower (that will be split)
for (m = level = 0; level <= cur_max_level; level++) {
// can not use sorted list here since this routine changes the list
for (n = 0; n < max_active_block; n++)
if (blocks[n].number >= 0 && blocks[n].level == level) {
bp = &blocks[n];
if (bp->refine == 1) {
nl = bp->number - block_start[level];
zp = nl/((p2[level]*npx*init_block_x)*
(p2[level]*npy*init_block_y));
yp = (nl%((p2[level]*npx*init_block_x)*
(p2[level]*npy*init_block_y)))/
(p2[level]*npx*init_block_x);
xp = nl%(p2[level]*npx*init_block_x);
if ((num_active + 8) > max_num_blocks) {
printf("ERROR: Need more blocks %d %d on %d\n", num_active, max_num_blocks, my_pe);
exit(-1);
}
if ((num_active + 8) > local_max_b)
local_max_b = num_active + 8;
del_sorted_list(bp->number, level);
num_active += 7;
num_blocks[level]--;
num_blocks[level+1] += 8;
for (p = 0; p < max_active_parent; p++)
if (parents[p].number < 0)
break;
if (p == max_num_parents) {
printf("ERROR: Need more parents\n");
exit(-1);
}
if (p == max_active_parent)
max_active_parent++;
num_parents++;
num_refined++;
pp = &parents[p];
pp->number = bp->number;
pp->level = bp->level;
pp->parent = bp->parent;
pp->parent_node = bp->parent_node;
pp->child_number = bp->child_number;
parents[pp->parent].child[pp->child_number] = -1 - p;
pp->refine = 0;
pp->cen[0] = bp->cen[0];
pp->cen[1] = bp->cen[1];
pp->cen[2] = bp->cen[2];
// Define the 8 children
for (o = 0; o < 8; o++) {
for ( ; m < max_num_blocks; m++)
if (blocks[m].number < 0)
break;
if (m == max_num_blocks) {
printf("Error: No inactive blocks available %d %d %d\n", m, num_active, max_num_blocks);
exit(-1);
}
if ((m+1) > max_active_block)
max_active_block = m+1;
bp1 = &blocks[m];
sib[o] = m;
pp->child[o] = m;
pp->child_node[o] = my_pe;
bp1->refine = 0;
bp1->level = level + 1;
bp1->parent = p;
bp1->parent_node = my_pe;
bp1->child_number = o;
i1 = (o%2);
j1 = ((o/2)%2);
k1 = (o/4);
bp1->number = ((2*zp+k1)*(p2[level+1]*npy*init_block_y) +
(2*yp+j1))*(p2[level+1]*npx*init_block_x) +
2*xp + i1 + block_start[level+1];
add_sorted_list(m, bp1->number, (level+1));
bp1->cen[0] = bp->cen[0] +
(2*i1 - 1)*p2[num_refine - level - 1];
bp1->cen[1] = bp->cen[1] +
(2*j1 - 1)*p2[num_refine - level - 1];
bp1->cen[2] = bp->cen[2] +
(2*k1 - 1)*p2[num_refine - level - 1];
half_size = p2[num_refine - level - 1];
i1 *= x_block_half;
j1 *= y_block_half;
k1 *= z_block_half;
for (v = 0; v < num_vars; v++)
for (i2 = i = 1; i <= x_block_half; i++, i2+=2)
for (j2 = j = 1; j <= y_block_half; j++, j2+=2)
for (k2 = k = 1; k <= z_block_half; k++, k2+=2)
bp1->array[v][i2 ][j2 ][k2 ] =
bp1->array[v][i2+1][j2 ][k2 ] =
bp1->array[v][i2 ][j2+1][k2 ] =
bp1->array[v][i2+1][j2+1][k2 ] =
bp1->array[v][i2 ][j2 ][k2+1] =
bp1->array[v][i2+1][j2 ][k2+1] =
bp1->array[v][i2 ][j2+1][k2+1] =
bp1->array[v][i2+1][j2+1][k2+1] =
bp->array[v][i+i1][j+j1][k+k1]/8.0;
}
// children all defined - connect children & disconnect parent
for (c = 0; c < 6; c++) {
// deal with internal connections amoung 8 siblings
for (i = 0; i < 2; i++)
for (j = 0; j < 2; j++) {
blocks[sib[side[c][i][j]+off[c]]].nei_level[c] =
level + 1;
blocks[sib[side[c][i][j]+off[c]]].nei[c][0][0] =
sib[side[c][i][j]];
}
// deal with external connections
if (bp->nei_level[c] == -2) // external boundary
for (i = 0; i < 2; i++)
for (j = 0; j < 2; j++) {
blocks[sib[side[c][i][j]]].nei_level[c] = -2;
blocks[sib[side[c][i][j]]].nei[c][0][0] = 0;
}
else if (bp->nei_level[c] == level-1) { // level less parent
if (bp->nei[c][0][0] >= 0) { // error
printf("%d ERROR: internal misconnect block %d num %d c %d %d\n",
my_pe, n, bp->number, c, bp->nei[c][0][0]);
exit(-1);
}
} else if (bp->nei_level[c] == level) { // paretn level
if (bp->nei[c][0][0] >= 0) {
other = bp->nei[c][0][0];
c1 = (c/2)*2 + (c+1)%2;
blocks[other].nei_level[c1] = level + 1;
for (i = 0; i < 2; i++)
for (j = 0; j < 2; j++) {
bp1 = &blocks[sib[side[c][i][j]]];
bp1->nei_level[c] = level;
bp1->nei[c][0][0] = other;
blocks[other].nei[c1][i][j] = sib[side[c][i][j]];
}
}
} else if (bp->nei_level[c] == level+1) { // child level
dir = c/2;
fcase = (c%2)*10;
c1 = (c/2)*2 + (c+1)%2;
d = 2*(c%2) - 1;
for (k = fcase+6, i = 0; i < 2; i++)
for (j = 0; j < 2; j++, k++)
if (bp->nei[c][i][j] >= 0) {
other = bp->nei[c][i][j];
bp1 = &blocks[sib[side[c][i][j]]];
bp1->nei_level[c] = level+1;
bp1->nei[c][0][0] = other;
blocks[other].nei_level[c1] = level + 1;
blocks[other].nei[c1][0][0] = sib[side[c][i][j]];
}
} else {
printf("%d ERROR: misconnected b %d %d l %d nei[%d] %d\n",
my_pe, n, bp->number, level, c, bp->nei_level[c]);
exit(-1);
}
}
/* children all defined and connected - inactivate block */
bp->number = -1;
if (n < m)
m = n;
}
}
}
}
// This routine takes blocks that are to be coarsened and recombines them.
// Before this routine can be called, all of the child blocks need to be on
// the same processor as the parent. A new block is created and the parent
// and child blocks are inactivated.
void consolidate_blocks(void)
{
static int side[6][2][2] = { { {0, 2}, {4, 6} }, { {1, 3}, {5, 7} },
{ {0, 1}, {4, 5} }, { {2, 3}, {6, 7} },
{ {0, 1}, {2, 3} }, { {4, 5}, {6, 7} } };
int n, p, i, j, k, i1, j1, k1, i2, j2, k2, level, o, v, f, c, offset,
other, c1, dir, fcase, pe, nl, pos[3], d, in;
block *bp, *bp1;
parent *pp;
if (stencil == 7) // add to face case when diags are needed
f = 0;
else
f = 1;
// assume that blocks were moved back to node with parent
for (level = cur_max_level; level >= 0; level--)
for (p = 0; p < max_active_parent; p++)
if ((pp = &parents[p])->number >= 0 && pp->level == level &&
pp->refine == -1) {
for (n = 0; n < max_num_blocks; n++)
if (blocks[n].number < 0) // found inactive block
break;
if (n == max_num_blocks) {
printf("Out of free blocks in consolidate_blocks %d\n", my_pe);
exit(-1);
} else
bp = &blocks[n];
if ((n+1) > max_active_block)
max_active_block = n+1;
if ((num_active + 1) > local_max_b)
local_max_b = num_active + 1;
num_active -= 7;
num_reformed++;
num_blocks[level]++;
num_blocks[level+1] -= 8;
bp->number = pp->number;
pp->number = -1;
bp->level = pp->level;
bp->parent = pp->parent;
bp->parent_node = pp->parent_node;
bp->child_number = pp->child_number;
parents[bp->parent].child[bp->child_number] = n;
add_sorted_list(n, bp->number, level);
bp->refine = 0;
bp->cen[0] = pp->cen[0];
bp->cen[1] = pp->cen[1];
bp->cen[2] = pp->cen[2];
// Copy child arrays back to new block.
for (o = 0; o < 8; o++) {
bp1 = &blocks[pp->child[o]];
del_sorted_list(bp1->number, (level+1));
bp1->number = -1;
i1 = (o%2)*x_block_half;
j1 = ((o/2)%2)*y_block_half;
k1 = (o/4)*z_block_half;
for (v = 0; v < num_vars; v++)
for (i2 = i = 1; i <= x_block_half; i++, i2+=2)
for (j2 = j = 1; j <= y_block_half; j++, j2+=2)
for (k2 = k = 1; k <= z_block_half; k++, k2+=2)
bp->array[v][i+i1][j+j1][k+k1] =
bp1->array[v][i2 ][j2 ][k2 ] +
bp1->array[v][i2+1][j2 ][k2 ] +
bp1->array[v][i2 ][j2+1][k2 ] +
bp1->array[v][i2+1][j2+1][k2 ] +
bp1->array[v][i2 ][j2 ][k2+1] +
bp1->array[v][i2+1][j2 ][k2+1] +
bp1->array[v][i2 ][j2+1][k2+1] +
bp1->array[v][i2+1][j2+1][k2+1];
}
// now figure out communication
for (c = 0; c < 6; c++) {
other = pp->child[side[c][0][0]]; // first child on this side
// four options - boundary, level of parent, level of children,
// and level of children + 1 (that are offnode and unrefining)
if (blocks[other].nei_level[c] == -2) {
// external boundary (only need to check one child)
bp->nei_level[c] = -2;
bp->nei_refine[c] = 0;
} else if (blocks[other].nei_level[c] == level) {
// same level as parent
if (blocks[other].nei[c][0][0] >= 0) {
// on node - if it gets consolidated later, it will fix
// the connections at that point
c1 = (c/2)*2 + (c+1)%2;
bp->nei[c][0][0] = blocks[other].nei[c][0][0];
bp->nei_level[c] = level;
bp->nei_refine[c] = 0;
blocks[blocks[other].nei[c][0][0]].nei[c1][0][0] = n;
blocks[blocks[other].nei[c][0][0]].nei_level[c1] = level;
blocks[blocks[other].nei[c][0][0]].nei_refine[c1] = 0;
}
} else {
dir = c/2;
fcase = (c%2)*10;
offset = p2[num_refine - level - 1];
for (k = fcase+6, i = 0; i < 2; i++)
for (j = 0; j < 2; j++, k++) {
other = pp->child[side[c][i][j]];
if (blocks[other].nei[c][0][0] >= 0) {
if (blocks[other].nei_level[c] == level+2) {
printf("%d ERROR: %d con %d block %d c %d wrong level %d\n",
my_pe, p, n, other, c, level);
exit(-1);
}
c1 = (c/2)*2 + (c+1)%2;
bp->nei[c][i][j] = blocks[other].nei[c][0][0];
bp->nei_level[c] = level + 1;
bp->nei_refine[c] = 0;
blocks[blocks[other].nei[c][0][0]].nei[c1][0][0] =
n;
blocks[blocks[other].nei[c][0][0]].nei_level[c1] =
level;
blocks[blocks[other].nei[c][0][0]].nei_refine[c1] =
0;
}
}
}
}
}
}
void add_sorted_list(int n, int number, int level)
{
int i, j;
for (i = sorted_index[level]; i < sorted_index[level+1]; i++)
if (number > sorted_list[i].number)
break;
for (j = sorted_index[num_refine+1]; j > i; j--) {
sorted_list[j].number = sorted_list[j-1].number;
sorted_list[j].n = sorted_list[j-1].n;
}
sorted_list[i].number = number;
sorted_list[i].n = n;
for (i = level+1; i <= (num_refine+1); i++)
sorted_index[i]++;
}
void del_sorted_list(int number, int level)
{
int i, j;
for (i = sorted_index[level]; i < sorted_index[level+1]; i++)
if (number == sorted_list[i].number)
break;
if (number != sorted_list[i].number) {
printf("ERROR: del_sorted_list on %d - number %d not found\n",
my_pe, number);
exit(-1);
}
for (j = level+1; j <= (num_refine+1); j++)
sorted_index[j]--;
for (j = i; j < sorted_index[num_refine+1]; j++) {
sorted_list[j].number = sorted_list[j+1].number;
sorted_list[j].n = sorted_list[j+1].n;
}
}
int find_sorted_list(int number, int level)
{
int i;
for (i = sorted_index[level]; i < sorted_index[level+1]; i++)
if (number == sorted_list[i].number)
return sorted_list[i].n;
printf("ERROR: find_sorted_list on %d - number %d not found\n",
my_pe, number);
exit(-1);
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/check_sum.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <math.h>
#include "block.h"
#include "timer.h"
#include "proto.h"
// Generate check sum for a variable over all active blocks.
double check_sum(int var)
{
int n, in, i, j, k;
double sum, block_sum, t1, t2;
block *bp;
t1 = timer();
sum = 0.0;
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
bp = &blocks[n];
if (bp->number >= 0) {
block_sum = 0.0;
for (i = 1; i <= x_block_size; i++)
for (j = 1; j <= y_block_size; j++)
for (k = 1; k <= z_block_size; k++)
block_sum += bp->array[var][i][j][k];
sum += block_sum;
}
}
t2 = timer();
timer_cs_calc += t2 - t1;
total_red++;
return sum;
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/comm.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdio.h>
#include <stdlib.h>
#include "block.h"
#include "timer.h"
#include "proto.h"
// The routines in this file are used in the communication of ghost values
// between blocks, both on processor and off processor.
// Main communication routine that sends and recieves ghost values between
// blocks on different processors and directos blocks on the same processor
// to exchange their ghost values.
void comm(int start, int num_comm, int stage)
{
int i, j, k, l, m, n, dir, o, in, which, offset, type;
int permutations[6][3] = { {0, 1, 2}, {1, 2, 0}, {2, 0, 1},
{0, 2, 1}, {1, 0, 2}, {2, 1, 0} };
double t1, t2, t3, t4;
block *bp;
for (o = 0; o < 3; o++) {
if (permute)
dir = permutations[stage%6][o];
else
dir = o;
type = dir;
t1 = timer();
// While values are being sent over the mesh, go through and direct
// blocks to exchange ghost values with other blocks that are on
// processor. Also apply boundary conditions for boundary of domain.
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
bp = &blocks[n];
if (bp->number >= 0)
for (l = dir*2; l < (dir*2 + 2); l++) {
if (bp->nei_level[l] == bp->level) {
t2 = timer();
if ((m = bp->nei[l][0][0]) > n) {
on_proc_comm(n, m, l, start, num_comm);
counter_same[dir] += 2;
}
timer_comm_same[dir] += timer() - t2;
} else if (bp->nei_level[l] == (bp->level+1)) {
t2 = timer();
for (i = 0; i < 2; i++)
for (j = 0; j < 2; j++)
if ((m = bp->nei[l][i][j]) > n) {
on_proc_comm_diff(n, m, l, i, j, start, num_comm);
counter_diff[dir] += 2;
}
timer_comm_diff[dir] += timer() - t2;
} else if (bp->nei_level[l] == (bp->level-1)) {
t2 = timer();
if ((m = bp->nei[l][0][0]) > n) {
k = dir*2 + 1 - l%2;
for (i = 0; i < 2; i++)
for (j = 0; j < 2; j++)
if (blocks[m].nei[k][i][j] == n) {
on_proc_comm_diff(m, n, k, i, j, start, num_comm);
counter_diff[dir] += 2;
}
}
timer_comm_diff[dir] += timer() - t2;
} else if (bp->nei_level[l] == -2) {
t2 = timer();
apply_bc(l, bp, start, num_comm);
counter_bc[dir]++;
timer_comm_bc[dir] += timer() - t2;
} else {
printf("ERROR: misconnected block\n");
exit(-1);
}
}
}
timer_comm_dir[dir] += timer() - t1;
}
}
// Routine that does on processor communication between two blocks that
// are at the same level of refinement.
void on_proc_comm(int n, int n1, int l, int start, int num_comm)
{
int i, j, k, m;
int is, ie, js, je;
block *bp, *bp1;
/* Determine direction and then exchange data across the face
*/
if (!code) {
if ((l/2) == 0) { /* West, East */
if ((l%2) == 0) { /* West */
bp = &blocks[n];
bp1 = &blocks[n1];
} else { /* East */
bp1 = &blocks[n];
bp = &blocks[n1];
}
for (m = start; m < start+num_comm; m++)
for (j = 1; j <= y_block_size; j++)
for (k = 1; k <= z_block_size; k++) {
bp1->array[m][x_block_size+1][j][k] = bp->array[m][1][j][k];
bp->array[m][0][j][k] = bp1->array[m][x_block_size][j][k];
}
} else if ((l/2) == 1) { /* South, North */
if ((l%2) == 0) { /* South */
bp = &blocks[n];
bp1 = &blocks[n1];
} else { /* North */
bp1 = &blocks[n];
bp = &blocks[n1];
}
if (stencil == 7) {
is = 1;
ie = x_block_size;
} else {
is = 0;
ie = x_block_size + 1;
}
for (m = start; m < start+num_comm; m++)
for (i = is; i <= ie; i++)
for (k = 1; k <= z_block_size; k++) {
bp1->array[m][i][y_block_size+1][k] = bp->array[m][i][1][k];
bp->array[m][i][0][k] = bp1->array[m][i][y_block_size][k];
}
} else if ((l/2) == 2) { /* Down, Up */
if ((l%2) == 0) { /* Down */
bp = &blocks[n];
bp1 = &blocks[n1];
} else { /* Up */
bp1 = &blocks[n];
bp = &blocks[n1];
}
if (stencil == 7) {
is = 1;
ie = x_block_size;
js = 1;
je = y_block_size;
} else {
is = 0;
ie = x_block_size + 1;
js = 0;
je = y_block_size + 1;
}
for (m = start; m < start+num_comm; m++)
for (i = is; i <= ie; i++)
for (j = js; j <= je; j++) {
bp1->array[m][i][j][z_block_size+1] = bp->array[m][i][j][1];
bp->array[m][i][j][0] = bp1->array[m][i][j][z_block_size];
}
}
} else { /* set all ghosts */
if ((l/2) == 0) { /* West, East */
if ((l%2) == 0) { /* West */
bp = &blocks[n];
bp1 = &blocks[n1];
} else { /* East */
bp1 = &blocks[n];
bp = &blocks[n1];
}
for (m = start; m < start+num_comm; m++)
for (j = 0; j <= y_block_size+1; j++)
for (k = 0; k <= z_block_size+1; k++) {
bp1->array[m][x_block_size+1][j][k] = bp->array[m][1][j][k];
bp->array[m][0][j][k] = bp1->array[m][x_block_size][j][k];
}
} else if ((l/2) == 1) { /* South, North */
if ((l%2) == 0) { /* South */
bp = &blocks[n];
bp1 = &blocks[n1];
} else { /* North */
bp1 = &blocks[n];
bp = &blocks[n1];
}
for (m = start; m < start+num_comm; m++)
for (i = 0; i <= x_block_size+1; i++)
for (k = 0; k <= z_block_size+1; k++) {
bp1->array[m][i][y_block_size+1][k] = bp->array[m][i][1][k];
bp->array[m][i][0][k] = bp1->array[m][i][y_block_size][k];
}
} else if ((l/2) == 2) { /* Down, Up */
if ((l%2) == 0) { /* Down */
bp = &blocks[n];
bp1 = &blocks[n1];
} else { /* Up */
bp1 = &blocks[n];
bp = &blocks[n1];
}
for (m = start; m < start+num_comm; m++)
for (i = 0; i <= x_block_size+1; i++)
for (j = 0; j <= y_block_size+1; j++) {
bp1->array[m][i][j][z_block_size+1] = bp->array[m][i][j][1];
bp->array[m][i][j][0] = bp1->array[m][i][j][z_block_size];
}
}
}
}
// Routine that does on processor communication between two blocks that are
// at different levels of refinement. The order of the blocks that are
// being input determine which block is at a higher level of refinement.
void on_proc_comm_diff(int n, int n1, int l, int iq, int jq,
int start, int num_comm)
{
int i, j, k, m;
int i0, i1, i2, i3, j0, j1, j2, j3, k0, k1, k2, k3;
block *bp, *bp1;
bp = &blocks[n];
bp1 = &blocks[n1];
/* (iq, jq) quarter face on block n to whole face on block n1
*/
if (!code) {
/* only have to communicate ghost values - bp is level, bp1 is level+1 -
* in 2 to 1 case get 0..block_half from one proc and
* block_half+1..block_size+1 from another
* in 1 to 2 case get 0..block_size+1 from 0..block_half+1 or
* block_half..block_size+1 with averages
*/
if ((l/2) == 0) {
if (l == 0) { /* West */
i0 = 0;
i1 = 1;
i2 = x_block_size + 1;
i3 = x_block_size;
} else { /* East */
i0 = x_block_size + 1;
i1 = x_block_size;
i2 = 0;
i3 = 1;
}
j1 = jq*y_block_half;
k1 = iq*z_block_half;
for (m = start; m < start+num_comm; m++)
for (j = 1; j <= y_block_half; j++)
for (k = 1; k <= z_block_half; k++) {
bp1->array[m][i2][2*j-1][2*k-1] =
bp1->array[m][i2][2*j-1][2*k ] =
bp1->array[m][i2][2*j ][2*k-1] =
bp1->array[m][i2][2*j ][2*k ] =
bp->array[m][i1][j+j1][k+k1]/4.0;
bp->array[m][i0][j+j1][k+k1] =
bp1->array[m][i3][2*j-1][2*k-1] +
bp1->array[m][i3][2*j-1][2*k ] +
bp1->array[m][i3][2*j ][2*k-1] +
bp1->array[m][i3][2*j ][2*k ];
}
} else if ((l/2) == 1) {
if (l == 2) { /* South */
j0 = 0;
j1 = 1;
j2 = y_block_size + 1;
j3 = y_block_size;
} else { /* North */
j0 = y_block_size + 1;
j1 = y_block_size;
j2 = 0;
j3 = 1;
}
i1 = jq*x_block_half;
k1 = iq*z_block_half;
for (m = start; m < start+num_comm; m++)
for (i = 1; i <= x_block_half; i++)
for (k = 1; k <= z_block_half; k++) {
bp1->array[m][2*i-1][j2][2*k-1] =
bp1->array[m][2*i-1][j2][2*k ] =
bp1->array[m][2*i ][j2][2*k-1] =
bp1->array[m][2*i ][j2][2*k ] =
bp->array[m][i+i1][j1][k+k1]/4.0;
bp->array[m][i+i1][j0][k+k1] =
bp1->array[m][2*i-1][j3][2*k-1] +
bp1->array[m][2*i-1][j3][2*k ] +
bp1->array[m][2*i ][j3][2*k-1] +
bp1->array[m][2*i ][j3][2*k ];
}
} else if ((l/2) == 2) {
if (l == 4) { /* Down */
k0 = 0;
k1 = 1;
k2 = z_block_size + 1;
k3 = z_block_size;
} else { /* Up */
k0 = z_block_size + 1;
k1 = z_block_size;
k2 = 0;
k3 = 1;
}
i1 = jq*x_block_half;
j1 = iq*y_block_half;
for (m = start; m < start+num_comm; m++)
for (i = 1; i <= x_block_half; i++)
for (j = 1; j <= y_block_half; j++) {
bp1->array[m][2*i-1][2*j-1][k2] =
bp1->array[m][2*i-1][2*j ][k2] =
bp1->array[m][2*i ][2*j-1][k2] =
bp1->array[m][2*i ][2*j ][k2] =
bp->array[m][i+i1][j+j1][k1]/4.0;
bp->array[m][i+i1][j+j1][k0] =
bp1->array[m][2*i-1][2*j-1][k3] +
bp1->array[m][2*i-1][2*j ][k3] +
bp1->array[m][2*i ][2*j-1][k3] +
bp1->array[m][2*i ][2*j ][k3];
}
}
} else { /* transfer ghosts */
if ((l/2) == 0) {
if (l == 0) { /* West */
i0 = 0;
i1 = 1;
i2 = x_block_size + 1;
i3 = x_block_size;
} else { /* East */
i0 = x_block_size + 1;
i1 = x_block_size;
i2 = 0;
i3 = 1;
}
j1 = jq*y_block_half;
k1 = iq*z_block_half;
j2 = y_block_size + 1;
j3 = y_block_half + 1;
k2 = z_block_size + 1;
k3 = z_block_half + 1;
for (m = start; m < start+num_comm; m++) {
bp1->array[m][i2][0][0] = bp->array[m][i1][j1][k1]/4.0;
for (k = 1; k <= z_block_half; k++)
bp1->array[m][i2][0][2*k-1] =
bp1->array[m][i2][0][2*k ] = bp->array[m][i1][j1][k+k1]/4.0;
bp1->array[m][i2][0][k2] = bp->array[m][i1][j1][k3+k1]/4.0;
if (jq == 0) {
if (iq == 0)
bp->array[m][i0][0][0 ] = bp1->array[m][i3][0][0 ];
else
bp->array[m][i0][0][k2] = bp1->array[m][i3][0][k2];
for (k = 1; k <= z_block_half; k++)
bp->array[m][i0][0][k+k1] = (bp1->array[m][i3][0][2*k-1] +
bp1->array[m][i3][0][2*k ]);
}
for (j = 1; j <= y_block_half; j++) {
bp1->array[m][i2][2*j-1][0] =
bp1->array[m][i2][2*j ][0] = bp->array[m][i1][j+j1][k1]/4.0;
if (iq == 0)
bp->array[m][i0][j+j1][0 ] = (bp1->array[m][i3][2*j-1][0 ] +
bp1->array[m][i3][2*j ][0 ]);
else
bp->array[m][i0][j+j1][k2] = (bp1->array[m][i3][2*j-1][k2] +
bp1->array[m][i3][2*j ][k2]);
for (k = 1; k <= z_block_half; k++) {
bp1->array[m][i2][2*j-1][2*k-1] =
bp1->array[m][i2][2*j-1][2*k ] =
bp1->array[m][i2][2*j ][2*k-1] =
bp1->array[m][i2][2*j ][2*k ] =
bp->array[m][i1][j+j1][k+k1]/4.0;
bp->array[m][i0][j+j1][k+k1] =
bp1->array[m][i3][2*j-1][2*k-1] +
bp1->array[m][i3][2*j-1][2*k ] +
bp1->array[m][i3][2*j ][2*k-1] +
bp1->array[m][i3][2*j ][2*k ];
}
bp1->array[m][i2][2*j-1][k2] =
bp1->array[m][i2][2*j ][k2] = bp->array[m][i1][j+j1][k3+k1]/4.0;
}
bp1->array[m][i2][j2][0] = bp->array[m][i1][j3+j1][k1]/4.0;
for (k = 1; k <= z_block_half; k++)
bp1->array[m][i2][j2][2*k-1] =
bp1->array[m][i2][j2][2*k ] = bp->array[m][i1][j3+j1][k+k1]/4.0;
bp1->array[m][i2][j2][k2] = bp->array[m][i1][j3+j1][k3+k1]/4.0;
if (jq == 1) {
if (iq == 0)
bp->array[m][i0][j2][0 ] = bp1->array[m][i3][j2][0 ];
else
bp->array[m][i0][j2][k2] = bp1->array[m][i3][j2][k2];
for (k = 1; k <= z_block_half; k++)
bp->array[m][i0][j2][k+k1] = (bp1->array[m][i3][j2][2*k-1] +
bp1->array[m][i3][j2][2*k ]);
}
}
} else if ((l/2) == 1) {
if (l == 2) { /* South */
j0 = 0;
j1 = 1;
j2 = y_block_size + 1;
j3 = y_block_size;
} else { /* North */
j0 = y_block_size + 1;
j1 = y_block_size;
j2 = 0;
j3 = 1;
}
i1 = jq*x_block_half;
k1 = iq*z_block_half;
i2 = x_block_size + 1;
i3 = x_block_half + 1;
k2 = z_block_size + 1;
k3 = z_block_half + 1;
for (m = start; m < start+num_comm; m++) {
bp1->array[m][0][j2][0 ] = bp->array[m][i1][j1][k1]/4.0;
for (k = 1; k <= z_block_half; k++)
bp1->array[m][0][j2][2*k-1] =
bp1->array[m][0][j2][2*k ] = bp->array[m][i1][j1][k+k1]/4.0;
bp1->array[m][0][j2][k2] = bp->array[m][i1][j1][k3+k1]/4.0;
if (jq == 0) {
if (iq == 0)
bp->array[m][0][j0][0 ] = bp1->array[m][0][j3][0 ];
else
bp->array[m][0][j0][k2] = bp1->array[m][0][j3][k2];
for (k = 1; k <= z_block_half; k++)
bp->array[m][0][j0][k+k1] = (bp1->array[m][0][j3][2*k-1] +
bp1->array[m][0][j3][2*k ]);
}
for (i = 1; i <= x_block_half; i++) {
bp1->array[m][2*i-1][j2][0] =
bp1->array[m][2*i ][j2][0] = bp->array[m][i+i1][j1][k1]/4.0;
if (iq == 0)
bp->array[m][i+i1][j0][0 ] = (bp1->array[m][2*i-1][j3][0 ] +
bp1->array[m][2*i ][j3][0 ]);
else
bp->array[m][i+i1][j0][k2] = (bp1->array[m][2*i-1][j3][k2] +
bp1->array[m][2*i ][j3][k2]);
for (k = 1; k <= z_block_half; k++) {
bp1->array[m][2*i-1][j2][2*k-1] =
bp1->array[m][2*i-1][j2][2*k ] =
bp1->array[m][2*i ][j2][2*k-1] =
bp1->array[m][2*i ][j2][2*k ] =
bp->array[m][i+i1][j1][k+k1]/4.0;
bp->array[m][i+i1][j0][k+k1] =
bp1->array[m][2*i-1][j3][2*k-1] +
bp1->array[m][2*i-1][j3][2*k ] +
bp1->array[m][2*i ][j3][2*k-1] +
bp1->array[m][2*i ][j3][2*k ];
}
bp1->array[m][2*i-1][j2][k2] =
bp1->array[m][2*i ][j2][k2] = bp->array[m][i+i1][j1][k3+k1]/4.0;
}
bp1->array[m][i2][j2][0 ] = bp->array[m][i3+i1][j1][k1]/4.0;
for (k = 1; k <= z_block_half; k++)
bp1->array[m][i2][j2][2*k-1] =
bp1->array[m][i2][j2][2*k ] = bp->array[m][i3+i1][j1][k+k1]/4.0;
bp1->array[m][i2][j2][k2] = bp->array[m][i3+i1][j1][k3+k1]/4.0;
if (jq == 1) {
if (iq == 0)
bp->array[m][i2][j0][0 ] = bp1->array[m][i2][j3][0 ];
else
bp->array[m][i2][j0][k2] = bp1->array[m][i2][j3][k2];
for (k = 1; k <= z_block_half; k++)
bp->array[m][i2][j0][k+k1] = (bp1->array[m][i2][j3][2*k-1] +
bp1->array[m][i2][j3][2*k ]);
}
}
} else if ((l/2) == 2) {
if (l == 4) { /* Down */
k0 = 0;
k1 = 1;
k2 = z_block_size + 1;
k3 = z_block_size;
} else { /* Up */
k0 = z_block_size + 1;
k1 = z_block_size;
k2 = 0;
k3 = 1;
}
i1 = jq*x_block_half;
j1 = iq*y_block_half;
i2 = x_block_size + 1;
i3 = x_block_half + 1;
j2 = y_block_size + 1;
j3 = y_block_half + 1;
for (m = start; m < start+num_comm; m++) {
bp1->array[m][0][0 ][k2] = bp->array[m][i1][j1][k1]/4.0;
for (j = 1; j <= y_block_half; j++)
bp1->array[m][0][2*j-1][k2] =
bp1->array[m][0][2*j ][k2] = bp->array[m][i1][j+j1][k1]/4.0;
bp1->array[m][0][j2][k2] = bp->array[m][i1][j3+j1][k1]/4.0;
if (jq == 0) {
if (iq == 0)
bp->array[m][0][0 ][k0] = bp1->array[m][0][0 ][k3];
else
bp->array[m][0][j2][k0] = bp1->array[m][0][j2][k3];
for (j = 1; j <= y_block_half; j++)
bp->array[m][0][j+j1][k0] = (bp1->array[m][0][2*j-1][k3] +
bp1->array[m][0][2*j ][k3]);
}
for (i = 1; i <= x_block_half; i++) {
bp1->array[m][2*i-1][0][k2] =
bp1->array[m][2*i ][0][k2] = bp->array[m][i+i1][j1][k1]/4.0;
if (iq == 0)
bp->array[m][i+i1][0][k0] = (bp1->array[m][2*i-1][0][k3] +
bp1->array[m][2*i ][0][k3]);
else
bp->array[m][i+i1][j2][k0] = (bp1->array[m][2*i-1][j2][k3] +
bp1->array[m][2*i ][j2][k3]);
for (j = 1; j <= y_block_half; j++) {
bp1->array[m][2*i-1][2*j-1][k2] =
bp1->array[m][2*i-1][2*j ][k2] =
bp1->array[m][2*i ][2*j-1][k2] =
bp1->array[m][2*i ][2*j ][k2] =
bp->array[m][i+i1][j+j1][k1]/4.0;
bp->array[m][i+i1][j+j1][k0] =
bp1->array[m][2*i-1][2*j-1][k3] +
bp1->array[m][2*i-1][2*j ][k3] +
bp1->array[m][2*i ][2*j-1][k3] +
bp1->array[m][2*i ][2*j ][k3];
}
bp1->array[m][2*i-1][j2][k2] =
bp1->array[m][2*i ][j2][k2] = bp->array[m][i+i1][j3+j1][k1]/4.0;
}
bp1->array[m][i2][0 ][k2] = bp->array[m][i3+i1][j1][k1]/4.0;
for (j = 1; j <= y_block_half; j++)
bp1->array[m][i2][2*j-1][k2] =
bp1->array[m][i2][2*j ][k2] = bp->array[m][i3+i1][j+j1][k1]/4.0;
bp1->array[m][i2][j2][k2] = bp->array[m][i3+i1][j3+j1][k1]/4.0;
if (jq == 1) {
if (iq == 0)
bp->array[m][i2][0 ][k0] = bp1->array[m][i2][0 ][k3];
else
bp->array[m][i2][j2][k0] = bp1->array[m][i2][j2][k3];
for (j = 1; j <= y_block_half; j++)
bp->array[m][i2][j+j1][k0] = (bp1->array[m][i2][2*j-1][k3] +
bp1->array[m][i2][2*j ][k3]);
}
}
}
}
}
// Apply reflective boundary conditions to a face of a block.
void apply_bc(int l, block *bp, int start, int num_comm)
{
int var, i, j, k, f, t;
t = 0;
f = 1;
if (!code && stencil == 7)
switch (l) {
case 1: t = x_block_size + 1;
f = x_block_size;
case 0: for (var = start; var < start+num_comm; var++)
for (j = 1; j <= y_block_size; j++)
for (k = 1; k <= z_block_size; k++)
bp->array[var][t][j][k] = bp->array[var][f][j][k];
break;
case 3: t = y_block_size + 1;
f = y_block_size;
case 2: for (var = start; var < start+num_comm; var++)
for (i = 1; i <= x_block_size; i++)
for (k = 1; k <= z_block_size; k++)
bp->array[var][i][t][k] = bp->array[var][i][f][k];
break;
case 5: t = z_block_size + 1;
f = z_block_size;
case 4: for (var = start; var < start+num_comm; var++)
for (i = 1; i <= x_block_size; i++)
for (j = 1; j <= y_block_size; j++)
bp->array[var][i][j][t] = bp->array[var][i][j][f];
break;
}
else
switch (l) {
case 1: t = x_block_size + 1;
f = x_block_size;
case 0: for (var = start; var < start+num_comm; var++)
for (j = 0; j <= y_block_size+1; j++)
for (k = 0; k <= z_block_size+1; k++)
bp->array[var][t][j][k] = bp->array[var][f][j][k];
break;
case 3: t = y_block_size + 1;
f = y_block_size;
case 2: for (var = start; var < start+num_comm; var++)
for (i = 0; i <= x_block_size+1; i++)
for (k = 0; k <= z_block_size+1; k++)
bp->array[var][i][t][k] = bp->array[var][i][f][k];
break;
case 5: t = z_block_size + 1;
f = z_block_size;
case 4: for (var = start; var < start+num_comm; var++)
for (i = 0; i <= x_block_size+1; i++)
for (j = 0; j <= y_block_size+1; j++)
bp->array[var][i][j][t] = bp->array[var][i][j][f];
break;
}
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/driver.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdio.h>
#include <math.h>
#include "block.h"
#include "timer.h"
#include "proto.h"
// Main driver for program.
void driver(void)
{
int ts, var, start, number, stage, comm_stage;
double t1, t2, t3, t4;
double sum;
init();
init_profile();
counter_malloc_init = counter_malloc;
size_malloc_init = size_malloc;
t1 = timer();
if (num_refine || uniform_refine) refine(0);
t2 = timer();
timer_refine_all += t2 - t1;
if (plot_freq)
plot(0);
t3 = timer();
timer_plot += t3 - t2;
nb_min = nb_max = global_active;
for (comm_stage = 0, ts = 1; ts <= num_tsteps; ts++) {
for (stage = 0; stage < stages_per_ts; stage++, comm_stage++) {
total_blocks += global_active;
if (global_active < nb_min)
nb_min = global_active;
if (global_active > nb_max)
nb_max = global_active;
for (start = 0; start < num_vars; start += comm_vars) {
if (start+comm_vars > num_vars)
number = num_vars - start;
else
number = comm_vars;
t3 = timer();
comm(start, number, comm_stage);
t4 = timer();
timer_comm_all += t4 - t3;
for (var = start; var < (start+number); var++) {
stencil_calc(var);
t3 = timer();
timer_calc_all += t3 - t4;
if (checksum_freq && !(stage%checksum_freq)) {
sum = check_sum(var);
if (report_diffusion && !my_pe)
printf("%d var %d sum %lf old %lf diff %lf tol %lf\n",
ts, var, sum, grid_sum[var],
(fabs(sum - grid_sum[var])/grid_sum[var]), tol);
if (fabs(sum - grid_sum[var])/grid_sum[var] > tol) {
if (!my_pe)
printf("Time step %d sum %lf (old %lf) variable %d difference too large\n", ts, sum, grid_sum[var], var);
return;
}
grid_sum[var] = sum;
}
t4 = timer();
timer_cs_all += t4 - t3;
}
}
}
if (num_refine && !uniform_refine) {
move();
if (!(ts%refine_freq))
refine(ts);
}
t2 = timer();
timer_refine_all += t2 - t4;
t3 = timer();
if (plot_freq && !(ts%plot_freq))
plot(ts);
timer_plot += timer() - t3;
}
timer_all = timer() - t1;
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/glibc_compat_rand.h

/*===------------- glibc_compat_rand.h- glibc rand emulation --------------===*\
* |*
* |* The LLVM Compiler Infrastructure
* |*
* |* This file is distributed under the University of Illinois Open Source
* |* License. See LICENSE.TXT for details.
* |*
* \*===----------------------------------------------------------------------===*/
#ifndef GLIBC_COMPAT_RAND_H
#define GLIBC_COMPAT_RAND_H
int glibc_compat_rand(void);
void glibc_compat_srand(unsigned int seed);
#endif /* GLIBC_COMPAT_RAND_H */

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/glibc_compat_rand.c

/*===------------ glibc_compat_rand.c - glibc rand emulation --------------===*\
* *
* * The LLVM Compiler Infrastructure
* *
* * This file is distributed under the University of Illinois Open Source
* * License. See LICENSE.TXT for details.
* *
* \*===----------------------------------------------------------------------===*/
#include "glibc_compat_rand.h"
/**
* * This rand implementation is designed to emulate the implementation of
* * rand/srand in recent versions of glibc. This is used for programs which
* * require this specific rand implementation in order to pass verification
* * tests.
* *
* * For more information, see: http://www.mathstat.dal.ca/~selinger/random/
* **/
#define TABLE_SIZE 344
static unsigned int table[TABLE_SIZE];
static int next;
int glibc_compat_rand(void) {
/* Calculate the indices i-3 and i-31 in the circular vector. */
int i3 = (next < 3) ? (TABLE_SIZE + next - 3) : (next - 3);
int i31 = (next < 31) ? (TABLE_SIZE + next - 31) : (next - 31);
table[next] = table[i3] + table[i31];
unsigned int r = table[next] >> 1;
++next;
if (next >= TABLE_SIZE)
next = 0;
return r;
}
void glibc_compat_srand(unsigned int seed) {
if (seed == 0)
seed = 1;
table[0] = seed;
for (int i = 1; i < 31; i++) {
int r = (16807ll * table[i - 1]) % 2147483647;
if (r < 0)
r += 2147483647;
table[i] = r;
}
for (int i = 31; i < 34; i++)
table[i] = table[i - 31];
for (int i = 34; i < TABLE_SIZE; i++)
table[i] = table[i - 31] + table[i - 3];
next = 0;
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/init.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdlib.h>
#include <math.h>
#include "block.h"
#include "proto.h"
#ifndef NO_GLIBC_COMPAT_RAND
#include "glibc_compat_rand.h"
#define rand glibc_compat_rand
#define srand glibc_compat_srand
#endif
// Initialize the problem and setup initial blocks.
void init(void)
{
int n, var, i, j, k, l, m, o, size, dir, i1, i2, j1, j2, k1, k2, ib, jb, kb;
int start[num_pes], pos[3][num_pes], pos1[npx][npy][npz], set,
num, npx1, npy1, npz1, pes, fact, fac[25], nfac, f;
block *bp;
tol = pow(10.0, ((double) -error_tol));
p2[0] = p8[0] = 1;
for (i = 0; i < (num_refine+1); i++) {
p8[i+1] = p8[i]*8;
p2[i+1] = p2[i]*2;
sorted_index[i] = 0;
}
sorted_index[num_refine+1] = 0;
block_start[0] = 0;
local_max_b = global_max_b = init_block_x*init_block_y*init_block_z;
num = num_pes*global_max_b;
for (i = 1; i <= num_refine; i++) {
block_start[i] = block_start[i-1] + num;
num *= 8;
num_blocks[i] = 0;
}
x_block_half = x_block_size/2;
y_block_half = y_block_size/2;
z_block_half = z_block_size/2;
max_active_block = init_block_x*init_block_y*init_block_z;
num_active = max_active_block;
global_active = num_active*num_pes;
num_parents = max_active_parent = 0;
size = p2[num_refine+1]; /* block size is p2[num_refine+1-level]
* smallest block is size p2[1], so can find
* its center */
mesh_size[0] = npx*init_block_x*size;
mesh_size[1] = npy*init_block_y*size;
mesh_size[2] = npz*init_block_z*size;
for (o = n = k = 0; k < init_block_z; k++)
for (j = 0; j < init_block_y; j++)
for (i = 0; i < init_block_x; i++, n++) {
bp = &blocks[o];
bp->level = 0;
bp->number = n;
bp->parent = -1;
bp->cen[0] = i*size + size/2;
bp->cen[1] = j*size + size/2;
bp->cen[2] = k*size + size/2;
add_sorted_list(o, n, 0);
for (var = 0; var < num_vars; var++)
for (ib = 1; ib <= x_block_size; ib++)
for (jb = 1; jb <= y_block_size; jb++)
for (kb = 1; kb <= z_block_size; kb++)
bp->array[var][ib][jb][kb] =
((double) rand())/((double) RAND_MAX);
if (i == 0) { /* 0 boundary */
bp->nei_level[0] = -2;
bp->nei[0][0][0] = 0;
} else { /* neighbor on core */
bp->nei_level[0] = 0;
bp->nei[0][0][0] = o - 1;
}
bp->nei_refine[0] = 0;
if (i == (init_block_x - 1)) {
bp->nei_level[1] = -2;
bp->nei[1][0][0] = 0;
} else { /* neighbor on core */
bp->nei_level[1] = 0;
bp->nei[1][0][0] = o + 1;
}
bp->nei_refine[1] = 0;
if (j == 0) {
bp->nei_level[2] = -2;
bp->nei[2][0][0] = 0;
} else { /* neighbor on core */
bp->nei_level[2] = 0;
bp->nei[2][0][0] = o - init_block_x;
}
bp->nei_refine[2] = 0;
if (j == (init_block_y - 1)) {
bp->nei_level[3] = -2;
bp->nei[3][0][0] = 0;
} else { /* neighbor on core */
bp->nei_level[3] = 0;
bp->nei[3][0][0] = o + init_block_x;
}
bp->nei_refine[3] = 0;
if (k == 0) {
bp->nei_level[4] = -2;
bp->nei[4][0][0] = 0;
} else { /* neighbor on core */
bp->nei_level[4] = 0;
bp->nei[4][0][0] = o - init_block_x*init_block_y;
}
bp->nei_refine[4] = 0;
if (k == (init_block_z - 1)) {
bp->nei_level[5] = -2;
bp->nei[5][0][0] = 0;
} else { /* neighbor on core */
bp->nei_level[5] = 0;
bp->nei[5][0][0] = o + init_block_x*init_block_y;
}
bp->nei_refine[5] = 0;
o++;
}
for (var = 0; var < num_vars; var++)
grid_sum[var] = check_sum(var);
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/main.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include "block.h"
#include "timer.h"
#include "proto.h"
int main(int argc, char** argv)
{
int i, ierr, object_num;
int params[35];
double *objs;
#include "param.h"
my_pe = 0;
num_pes = 1;
counter_malloc = 0;
size_malloc = 0.0;
/* set initial values */
for (i = 1; i < argc; i++)
if (!strcmp(argv[i], "--max_blocks"))
max_num_blocks = atoi(argv[++i]);
else if (!strcmp(argv[i], "--target_active"))
target_active = atoi(argv[++i]);
else if (!strcmp(argv[i], "--target_max"))
target_max = atoi(argv[++i]);
else if (!strcmp(argv[i], "--target_min"))
target_min = atoi(argv[++i]);
else if (!strcmp(argv[i], "--num_refine"))
num_refine = atoi(argv[++i]);
else if (!strcmp(argv[i], "--block_change"))
block_change = atoi(argv[++i]);
else if (!strcmp(argv[i], "--uniform_refine"))
uniform_refine = atoi(argv[++i]);
else if (!strcmp(argv[i], "--nx"))
x_block_size = atoi(argv[++i]);
else if (!strcmp(argv[i], "--ny"))
y_block_size = atoi(argv[++i]);
else if (!strcmp(argv[i], "--nz"))
z_block_size = atoi(argv[++i]);
else if (!strcmp(argv[i], "--num_vars"))
num_vars = atoi(argv[++i]);
else if (!strcmp(argv[i], "--comm_vars"))
comm_vars = atoi(argv[++i]);
else if (!strcmp(argv[i], "--init_x"))
init_block_x = atoi(argv[++i]);
else if (!strcmp(argv[i], "--init_y"))
init_block_y = atoi(argv[++i]);
else if (!strcmp(argv[i], "--init_z"))
init_block_z = atoi(argv[++i]);
else if (!strcmp(argv[i], "--refine_freq"))
refine_freq = atoi(argv[++i]);
else if (!strcmp(argv[i], "--report_diffusion"))
report_diffusion = atoi(argv[++i]);
else if (!strcmp(argv[i], "--error_tol"))
error_tol = atoi(argv[++i]);
else if (!strcmp(argv[i], "--num_tsteps"))
num_tsteps = atoi(argv[++i]);
else if (!strcmp(argv[i], "--stages_per_ts"))
stages_per_ts = atoi(argv[++i]);
else if (!strcmp(argv[i], "--checksum_freq"))
checksum_freq = atoi(argv[++i]);
else if (!strcmp(argv[i], "--stencil"))
stencil = atoi(argv[++i]);
else if (!strcmp(argv[i], "--permute"))
permute = 1;
else if (!strcmp(argv[i], "--report_perf"))
report_perf = atoi(argv[++i]);
else if (!strcmp(argv[i], "--plot_freq"))
plot_freq = atoi(argv[++i]);
else if (!strcmp(argv[i], "--code"))
code = atoi(argv[++i]);
else if (!strcmp(argv[i], "--refine_ghost"))
refine_ghost = 1;
else if (!strcmp(argv[i], "--num_objects")) {
num_objects = atoi(argv[++i]);
objects = (object *) ma_malloc(num_objects*sizeof(object),
__FILE__, __LINE__);
object_num = 0;
} else if (!strcmp(argv[i], "--object")) {
if (object_num >= num_objects) {
printf("object number greater than num_objects\n");
exit(-1);
}
objects[object_num].type = atoi(argv[++i]);
objects[object_num].bounce = atoi(argv[++i]);
objects[object_num].cen[0] = atof(argv[++i]);
objects[object_num].cen[1] = atof(argv[++i]);
objects[object_num].cen[2] = atof(argv[++i]);
objects[object_num].move[0] = atof(argv[++i]);
objects[object_num].move[1] = atof(argv[++i]);
objects[object_num].move[2] = atof(argv[++i]);
objects[object_num].size[0] = atof(argv[++i]);
objects[object_num].size[1] = atof(argv[++i]);
objects[object_num].size[2] = atof(argv[++i]);
objects[object_num].inc[0] = atof(argv[++i]);
objects[object_num].inc[1] = atof(argv[++i]);
objects[object_num].inc[2] = atof(argv[++i]);
object_num++;
} else if (!strcmp(argv[i], "--help")) {
print_help_message();
exit(0);
} else {
printf("** Error ** Unknown input parameter %s\n", argv[i]);
print_help_message();
exit(-1);
}
if (check_input())
exit(-1);
if (!block_change)
block_change = num_refine;
for (object_num = 0; object_num < num_objects; object_num++)
for (i = 0; i < 3; i++) {
objects[object_num].orig_cen[i] = objects[object_num].cen[i];
objects[object_num].orig_move[i] = objects[object_num].move[i];
objects[object_num].orig_size[i] = objects[object_num].size[i];
}
allocate();
driver();
profile();
deallocate();
exit(0);
}
// =================================== print_help_message ====================
void print_help_message(void)
{
printf("(Optional) command line input is of the form: \n\n");
printf("--nx - block size x (even && > 0)\n");
printf("--ny - block size y (even && > 0)\n");
printf("--nz - block size z (even && > 0)\n");
printf("--init_x - initial blocks in x (> 0)\n");
printf("--init_y - initial blocks in y (> 0)\n");
printf("--init_z - initial blocks in z (> 0)\n");
printf("--reorder - ordering of blocks if initial number > 1\n");
printf("--max_blocks - maximun number of blocks per core\n");
printf("--num_refine - (>= 0) number of levels of refinement\n");
printf("--block_change - (>= 0) number of levels a block can change in a timestep\n");
printf("--uniform_refine - if 1, then grid is uniformly refined\n");
printf("--refine_freq - frequency (in timesteps) of checking for refinement\n");
printf("--target_active - (>= 0) target number of blocks per core, none if 0\n");
printf("--target_max - (>= 0) max number of blocks per core, none if 0\n");
printf("--target_min - (>= 0) min number of blocks per core, none if 0\n");
printf("--num_vars - number of variables (> 0)\n");
printf("--comm_vars - number of vars to communicate together\n");
printf("--num_tsteps - number of timesteps (> 0)\n");
printf("--stages_per_ts - number of comm/calc stages per timestep\n");
printf("--checksum_freq - number of stages between checksums\n");
printf("--stencil - 7 or 27 point (27 will not work with refinement (except uniform))\n");
printf("--error_tol - (e^{-error_tol} ; >= 0) \n");
printf("--report_diffusion - (>= 0) none if 0\n");
printf("--report_perf - 0, 1, 2\n");
printf("--refine_freq - frequency (timesteps) of plotting (0 for none)\n");
printf("--code - closely minic communication of different codes\n");
printf(" 0 minimal sends, 1 send ghosts, 2 send ghosts and process on send\n");
printf("--permute - altenates directions in communication\n");
printf("--refine_ghost - use full extent of block (including ghosts) to determine if block is refined\n");
printf("--num_objects - (>= 0) number of objects to cause refinement\n");
printf("--object - type, position, movement, size, size rate of change\n");
printf("All associated settings are integers except for objects\n");
}
// =================================== allocate ==============================
void allocate(void)
{
int i, j, k, m, n;
num_blocks = (int *) ma_malloc((num_refine+1)*sizeof(int),
__FILE__, __LINE__);
num_blocks[0] = num_pes*init_block_x*init_block_y*init_block_z;
num_blocks[0] = init_block_x*init_block_y*init_block_z;
blocks = (block *) ma_malloc(max_num_blocks*sizeof(block),
__FILE__, __LINE__);
for (n = 0; n < max_num_blocks; n++) {
blocks[n].number = -1;
blocks[n].array = (double ****) ma_malloc(num_vars*sizeof(double ***),
__FILE__, __LINE__);
for (m = 0; m < num_vars; m++) {
blocks[n].array[m] = (double ***)
ma_malloc((x_block_size+2)*sizeof(double **),
__FILE__, __LINE__);
for (i = 0; i < x_block_size+2; i++) {
blocks[n].array[m][i] = (double **)
ma_malloc((y_block_size+2)*sizeof(double *),
__FILE__, __LINE__);
for (j = 0; j < y_block_size+2; j++)
blocks[n].array[m][i][j] = (double *)
ma_malloc((z_block_size+2)*sizeof(double),
__FILE__, __LINE__);
}
}
}
sorted_list = (sorted_block *)ma_malloc(max_num_blocks*sizeof(sorted_block),
__FILE__, __LINE__);
sorted_index = (int *) ma_malloc((num_refine+2)*sizeof(int),
__FILE__, __LINE__);
max_num_parents = max_num_blocks; // Guess at number needed
parents = (parent *) ma_malloc(max_num_parents*sizeof(parent),
__FILE__, __LINE__);
for (n = 0; n < max_num_parents; n++)
parents[n].number = -1;
grid_sum = (double *)ma_malloc(num_vars*sizeof(double), __FILE__, __LINE__);
p8 = (int *) ma_malloc((num_refine+2)*sizeof(int), __FILE__, __LINE__);
p2 = (int *) ma_malloc((num_refine+2)*sizeof(int), __FILE__, __LINE__);
block_start = (int *) ma_malloc((num_refine+1)*sizeof(int),
__FILE__, __LINE__);
}
// =================================== deallocate ============================
void deallocate(void)
{
int i, j, m, n;
for (n = 0; n < max_num_blocks; n++) {
for (m = 0; m < num_vars; m++) {
for (i = 0; i < x_block_size+2; i++) {
for (j = 0; j < y_block_size+2; j++)
free(blocks[n].array[m][i][j]);
free(blocks[n].array[m][i]);
}
free(blocks[n].array[m]);
}
free(blocks[n].array);
}
free(blocks);
free(sorted_list);
free(sorted_index);
free(objects);
free(grid_sum);
free(p8);
free(p2);
}
int check_input(void)
{
int error = 0;
if (init_block_x < 1 || init_block_y < 1 || init_block_z < 1) {
printf("initial blocks on processor must be positive\n");
error = 1;
}
if (max_num_blocks < init_block_x*init_block_y*init_block_z) {
printf("max_num_blocks not large enough\n");
error = 1;
}
if (x_block_size < 1 || y_block_size < 1 || z_block_size < 1) {
printf("block size must be positive\n");
error = 1;
}
if (((x_block_size/2)*2) != x_block_size) {
printf("block size in x direction must be even\n");
error = 1;
}
if (((y_block_size/2)*2) != y_block_size) {
printf("block size in y direction must be even\n");
error = 1;
}
if (((z_block_size/2)*2) != z_block_size) {
printf("block size in z direction must be even\n");
error = 1;
}
if (target_active && target_max) {
printf("Only one of target_active and target_max can be used\n");
error = 1;
}
if (target_active && target_min) {
printf("Only one of target_active and target_min can be used\n");
error = 1;
}
if (target_active < 0 || target_active > max_num_blocks) {
printf("illegal value for target_active\n");
error = 1;
}
if (target_max < 0 || target_max > max_num_blocks ||
target_max < target_active) {
printf("illegal value for target_max\n");
error = 1;
}
if (target_min < 0 || target_min > max_num_blocks ||
target_min > target_active || target_min > target_max) {
printf("illegal value for target_min\n");
error = 1;
}
if (num_refine < 0) {
printf("number of refinement levels must be non-negative\n");
error = 1;
}
if (block_change < 0) {
printf("number of refinement levels must be non-negative\n");
error = 1;
}
if (num_vars < 1) {
printf("number of variables must be positive\n");
error = 1;
}
if (num_pes != npx*npy*npz) {
printf("number of processors used does not match number allocated\n");
error = 1;
}
if (stencil != 7 && stencil != 27) {
printf("illegal value for stencil\n");
error = 1;
}
if (stencil == 27 && num_refine && !uniform_refine)
printf("WARNING: 27 point stencil with non-uniform refinement: answers may diverge\n");
if (comm_vars == 0 || comm_vars > num_vars)
comm_vars = num_vars;
if (code < 0 || code > 2) {
printf("code must be 0, 1, or 2\n");
error = 1;
}
return (error);
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/miniAMR.reference_output

Number of blocks at level 0 at timestep 0 is 1
Number of blocks at level 1 at timestep 0 is 0
Number of blocks at level 2 at timestep 0 is 0
Number of blocks at level 3 at timestep 0 is 0
Number of blocks at level 4 at timestep 0 is 0
Number of blocks at level 5 at timestep 0 is 0
Number of blocks at level 0 at timestep 5 is 1
Number of blocks at level 1 at timestep 5 is 0
Number of blocks at level 2 at timestep 5 is 0
Number of blocks at level 3 at timestep 5 is 0
Number of blocks at level 4 at timestep 5 is 0
Number of blocks at level 5 at timestep 5 is 0
Number of blocks at level 0 at timestep 10 is 1
Number of blocks at level 1 at timestep 10 is 0
Number of blocks at level 2 at timestep 10 is 0
Number of blocks at level 3 at timestep 10 is 0
Number of blocks at level 4 at timestep 10 is 0
Number of blocks at level 5 at timestep 10 is 0
Number of blocks at level 0 at timestep 15 is 1
Number of blocks at level 1 at timestep 15 is 0
Number of blocks at level 2 at timestep 15 is 0
Number of blocks at level 3 at timestep 15 is 0
Number of blocks at level 4 at timestep 15 is 0
Number of blocks at level 5 at timestep 15 is 0
Number of blocks at level 0 at timestep 20 is 1
Number of blocks at level 1 at timestep 20 is 0
Number of blocks at level 2 at timestep 20 is 0
Number of blocks at level 3 at timestep 20 is 0
Number of blocks at level 4 at timestep 20 is 0
Number of blocks at level 5 at timestep 20 is 0
Number of blocks at level 0 at timestep 25 is 1
Number of blocks at level 1 at timestep 25 is 0
Number of blocks at level 2 at timestep 25 is 0
Number of blocks at level 3 at timestep 25 is 0
Number of blocks at level 4 at timestep 25 is 0
Number of blocks at level 5 at timestep 25 is 0
Number of blocks at level 0 at timestep 30 is 1
Number of blocks at level 1 at timestep 30 is 0
Number of blocks at level 2 at timestep 30 is 0
Number of blocks at level 3 at timestep 30 is 0
Number of blocks at level 4 at timestep 30 is 0
Number of blocks at level 5 at timestep 30 is 0
Number of blocks at level 0 at timestep 35 is 1
Number of blocks at level 1 at timestep 35 is 0
Number of blocks at level 2 at timestep 35 is 0
Number of blocks at level 3 at timestep 35 is 0
Number of blocks at level 4 at timestep 35 is 0
Number of blocks at level 5 at timestep 35 is 0
Number of blocks at level 0 at timestep 40 is 1
Number of blocks at level 1 at timestep 40 is 0
Number of blocks at level 2 at timestep 40 is 0
Number of blocks at level 3 at timestep 40 is 0
Number of blocks at level 4 at timestep 40 is 0
Number of blocks at level 5 at timestep 40 is 0
Number of blocks at level 0 at timestep 45 is 1
Number of blocks at level 1 at timestep 45 is 0
Number of blocks at level 2 at timestep 45 is 0
Number of blocks at level 3 at timestep 45 is 0
Number of blocks at level 4 at timestep 45 is 0
Number of blocks at level 5 at timestep 45 is 0
Number of blocks at level 0 at timestep 50 is 1
Number of blocks at level 1 at timestep 50 is 0
Number of blocks at level 2 at timestep 50 is 0
Number of blocks at level 3 at timestep 50 is 0
Number of blocks at level 4 at timestep 50 is 0
Number of blocks at level 5 at timestep 50 is 0
Number of blocks at level 0 at timestep 55 is 1
Number of blocks at level 1 at timestep 55 is 0
Number of blocks at level 2 at timestep 55 is 0
Number of blocks at level 3 at timestep 55 is 0
Number of blocks at level 4 at timestep 55 is 0
Number of blocks at level 5 at timestep 55 is 0
Number of blocks at level 0 at timestep 60 is 1
Number of blocks at level 1 at timestep 60 is 0
Number of blocks at level 2 at timestep 60 is 0
Number of blocks at level 3 at timestep 60 is 0
Number of blocks at level 4 at timestep 60 is 0
Number of blocks at level 5 at timestep 60 is 0
Number of blocks at level 0 at timestep 65 is 1
Number of blocks at level 1 at timestep 65 is 0
Number of blocks at level 2 at timestep 65 is 0
Number of blocks at level 3 at timestep 65 is 0
Number of blocks at level 4 at timestep 65 is 0
Number of blocks at level 5 at timestep 65 is 0
Number of blocks at level 0 at timestep 70 is 1
Number of blocks at level 1 at timestep 70 is 0
Number of blocks at level 2 at timestep 70 is 0
Number of blocks at level 3 at timestep 70 is 0
Number of blocks at level 4 at timestep 70 is 0
Number of blocks at level 5 at timestep 70 is 0
Number of blocks at level 0 at timestep 75 is 1
Number of blocks at level 1 at timestep 75 is 0
Number of blocks at level 2 at timestep 75 is 0
Number of blocks at level 3 at timestep 75 is 0
Number of blocks at level 4 at timestep 75 is 0
Number of blocks at level 5 at timestep 75 is 0
Number of blocks at level 0 at timestep 80 is 1
Number of blocks at level 1 at timestep 80 is 0
Number of blocks at level 2 at timestep 80 is 0
Number of blocks at level 3 at timestep 80 is 0
Number of blocks at level 4 at timestep 80 is 0
Number of blocks at level 5 at timestep 80 is 0
Number of blocks at level 0 at timestep 85 is 1
Number of blocks at level 1 at timestep 85 is 0
Number of blocks at level 2 at timestep 85 is 0
Number of blocks at level 3 at timestep 85 is 0
Number of blocks at level 4 at timestep 85 is 0
Number of blocks at level 5 at timestep 85 is 0
Number of blocks at level 0 at timestep 90 is 1
Number of blocks at level 1 at timestep 90 is 0
Number of blocks at level 2 at timestep 90 is 0
Number of blocks at level 3 at timestep 90 is 0
Number of blocks at level 4 at timestep 90 is 0
Number of blocks at level 5 at timestep 90 is 0
Number of blocks at level 0 at timestep 95 is 1
Number of blocks at level 1 at timestep 95 is 0
Number of blocks at level 2 at timestep 95 is 0
Number of blocks at level 3 at timestep 95 is 0
Number of blocks at level 4 at timestep 95 is 0
Number of blocks at level 5 at timestep 95 is 0
Number of blocks at level 0 at timestep 100 is 1
Number of blocks at level 1 at timestep 100 is 0
Number of blocks at level 2 at timestep 100 is 0
Number of blocks at level 3 at timestep 100 is 0
Number of blocks at level 4 at timestep 100 is 0
Number of blocks at level 5 at timestep 100 is 0
================ Start report ===================
Mantevo miniAMR
version 1.0 provisional
serial run on 1 rank
initial blocks per rank 1 x 1 x 1
block size 8 x 8 x 8
Maximum number of blocks per rank is 500
Number of levels of refinement is 5
Blocks can change by 5 levels per refinement step
Blocks will be refined by 0 objects
Number of timesteps is 100
Communicaion/computation stages per timestep is 20
Will perform checksums every 5 stages
Will refine every 5 timesteps
Will not plot results
Calculate on 40 variables with 7 point stencil
Communicate 40 variables at a time
Error tolorance for variable sums is 10^(-8)
exit 0

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/move.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdio.h>
#include <math.h>
#include "block.h"
#include "proto.h"
// This routine moves the objects that determine the refinement and checks
// the blocks against those objects to determine which blocks will refine.
void move(void)
{
int i, j;
for (i = 0; i < num_objects; i++)
for (j = 0; j < 3; j++) {
objects[i].cen[j] += objects[i].move[j];
if (objects[i].bounce)
if (objects[i].cen[j] >= 1.0) {
objects[i].cen[j] = 2.0 - objects[i].cen[j];
objects[i].move[j] = -objects[i].move[j];
} else if (objects[i].cen[j] <= 0.0) {
objects[i].cen[j] = 0.0 - objects[i].cen[j];
objects[i].move[j] = -objects[i].move[j];
}
objects[i].size[j] += objects[i].inc[j];
}
}
void check_objects(void)
{
int n, sz, in, c;
double cor[3][2]; /* extent of block */
block *bp;
parent *pp;
/* only need to check corners to refine
* * if boundary is in block refine if not level max_level block.refine = 1
* * else if not level 0 then block.refine = -1 (unrefine)
* * types of objects: spheroid, solid spheroid, rectangle, solid rectangle,
* * hemispheroid, and solid hemispheroid (in six orientations),
* * cylinder, solid cylinder (in three directions)
* * (later) diamond, solid diamond */
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
if ((bp = &blocks[n])->number >= 0) {
sz = p2[num_refine - bp->level]; /* half size of block */
cor[0][0] = ((double) (bp->cen[0] - sz))/((double) mesh_size[0]);
cor[0][1] = ((double) (bp->cen[0] + sz))/((double) mesh_size[0]);
cor[1][0] = ((double) (bp->cen[1] - sz))/((double) mesh_size[1]);
cor[1][1] = ((double) (bp->cen[1] + sz))/((double) mesh_size[1]);
cor[2][0] = ((double) (bp->cen[2] - sz))/((double) mesh_size[2]);
cor[2][1] = ((double) (bp->cen[2] + sz))/((double) mesh_size[2]);
if (refine_ghost) {
cor[0][0] -= 2.0*(((double) sz)/((double) x_block_size))/
((double) mesh_size[0]);
cor[0][1] += 2.0*(((double) sz)/((double) x_block_size))/
((double) mesh_size[0]);
cor[1][0] -= 2.0*(((double) sz)/((double) y_block_size))/
((double) mesh_size[1]);
cor[1][1] += 2.0*(((double) sz)/((double) y_block_size))/
((double) mesh_size[1]);
cor[2][0] -= 2.0*(((double) sz)/((double) z_block_size))/
((double) mesh_size[2]);
cor[2][1] += 2.0*(((double) sz)/((double) z_block_size))/
((double) mesh_size[2]);
}
if (check_block(cor))
bp->refine = 1;
else if (refine_ghost && bp->level) {
/* check if this block would unrefine, but its parent would then
* refine. Then leave it alone */
sz = p2[num_refine - bp->level + 1]; /* half size of parent */
cor[0][0] -= (((double) sz)/((double) x_block_size))/
((double) mesh_size[0]);
cor[0][1] += (((double) sz)/((double) x_block_size))/
((double) mesh_size[0]);
cor[1][0] -= (((double) sz)/((double) y_block_size))/
((double) mesh_size[1]);
cor[1][1] += (((double) sz)/((double) y_block_size))/
((double) mesh_size[1]);
cor[2][0] -= (((double) sz)/((double) z_block_size))/
((double) mesh_size[2]);
cor[2][1] += (((double) sz)/((double) z_block_size))/
((double) mesh_size[2]);
if (check_block(cor))
bp->refine = 0;
}
/* if at max refinement, then can not refine */
if ((bp->level == num_refine && bp->refine == 1) || !bp->refine) {
bp->refine = 0;
if (bp->parent != -1 && bp->parent_node == my_pe) {
pp = &parents[bp->parent];
pp->refine = 0;
for (c = 0; c < 8; c++)
if (pp->child_node[c] == my_pe && pp->child[c] >= 0)
if (blocks[pp->child[c]].refine == -1)
blocks[pp->child[c]].refine = 0;
}
}
/* if 0 level, we can not unrefine */
if (!bp->level && bp->refine == -1)
bp->refine = 0;
}
}
}
int check_block(double cor[3][2])
{
int o, tmp, ca, c1, c2, intersect,
xc, xv, yc, yv, zc, zv; /* where is center of object to block */
object *op;
intersect = 0;
for (o = 0; o < num_objects; o++) {
op = &objects[o];
if (intersect > 0 ||
op->size[0] < 0.0 || op->size[1] < 0.0 || op->size[2] < 0)
/* skip since already determined that it will be refined or
* can not be or object has size less than zero */
;
else if (op->type == 0) { /* surface of rectangle */
if (cor[0][1] > (op->cen[0] - op->size[0]) &&
cor[0][0] < (op->cen[0] + op->size[0])) {
/* some portion of block intersects with rectangle in x
* 4 cases - the two that straddle can be treated the same
* and the two that do not can also be treated the same */
if ((cor[0][0] < (op->cen[0] - op->size[0]) &&
cor[0][1] < (op->cen[0] + op->size[0])) ||
(cor[0][0] > (op->cen[0] - op->size[0]) &&
cor[0][1] > (op->cen[0] + op->size[0]))) {
/* one of rectangle boundary between block sides */
if (cor[1][1] > (op->cen[1] - op->size[1]) &&
cor[1][0] < (op->cen[1] + op->size[1]) &&
cor[2][1] > (op->cen[2] - op->size[2]) &&
cor[2][0] < (op->cen[2] + op->size[2]))
/* some portion of block intersects rectangle in y,z */
intersect = 1;
} else {
/* rectangle in block (or vice-versa) in x */
if (cor[1][1] > (op->cen[1] - op->size[1]) &&
cor[1][0] < (op->cen[1] + op->size[1])) {
/* some portion of block intersects rectangle in y */
if ((cor[1][0] < (op->cen[1] - op->size[1]) &&
cor[1][1] < (op->cen[1] + op->size[1])) ||
(cor[1][0] > (op->cen[1] - op->size[1]) &&
cor[1][1] > (op->cen[1] + op->size[1])))
if (cor[2][1] > (op->cen[2] - op->size[2]) &&
cor[2][0] < (op->cen[2] + op->size[2]))
/* portion of block intersects rectangle in z */
intersect = 1;
} else {
/* rectangle in block (or vice-versa) in x and y */
if (cor[2][1] > (op->cen[2] - op->size[2]) &&
cor[2][0] < (op->cen[2] + op->size[2])) {
if ((cor[2][0] < (op->cen[2] - op->size[2]) &&
cor[2][1] < (op->cen[2] + op->size[2])) ||
(cor[2][0] > (op->cen[2] - op->size[2]) &&
cor[2][1] > (op->cen[2] + op->size[2])))
/* portion of block intersects rectangle in z */
intersect = 1;
} /* else case need not be considered */
}
}
}
} else if (op->type == 1) { /* solid rectangle */
if (cor[0][1] > (op->cen[0] - op->size[0]) &&
cor[0][0] < (op->cen[0] + op->size[0]) &&
cor[1][1] > (op->cen[1] - op->size[1]) &&
cor[1][0] < (op->cen[1] + op->size[1]) &&
cor[2][1] > (op->cen[2] - op->size[2]) &&
cor[2][0] < (op->cen[2] + op->size[2]))
intersect = 1;
} else if (op->type >= 2 && op->type <= 14 && !(op->type%2)) {
/* boundary of spheroid or hemispheroid */
/* determine where center is and nearest and furthest
* verticies and then determine if boundary is between them
* 1 = (x/a)^2 + (y/b)^2 + (z/c)^2 is boundary */
tmp = intersect;
xc = yc = zc = 0;
if (op->cen[0] < cor[0][0])
xv = 0;
else if (op->cen[0] > cor[0][1])
xv = 1;
else {
xc = 1;
if (op->cen[0] < (cor[0][0] + cor[0][1])/2.0)
xv = 0;
else
xv = 1;
}
if (op->cen[1] < cor[1][0])
yv = 0;
else if (op->cen[1] > cor[1][1])
yv = 1;
else {
yc = 1;
if (op->cen[1] < (cor[1][0] + cor[1][1])/2.0)
yv = 0;
else
yv = 1;
}
if (op->cen[2] < cor[2][0])
zv = 0;
else if (op->cen[2] > cor[2][1])
zv = 1;
else {
zc = 1;
if (op->cen[2] < (cor[2][0] + cor[2][1])/2.0)
zv = 0;
else
zv = 1;
}
if (xc) {
if (yc) {
if (zc) { /* xc, yc, zc */
if ((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2]))
> 1.0)
intersect = 1;
} else { /* xc, yc, !zc */
if ((fabs(cor[2][zv] - op->cen[2]) < op->size[2]) &&
((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2])) > 1.0))
intersect = 1;
}
} else {
if (zc) { /* xc, !yc, zc */
if ((fabs(cor[1][yv] - op->cen[1]) < op->size[1]) &&
((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2])) > 1.0))
intersect = 1;
} else { /* xc, !yc, !zc */
if (((((cor[1][yv] - op->cen[1])/op->size[1])*
((cor[1][yv] - op->cen[1])/op->size[1]) +
((cor[2][zv] - op->cen[2])/op->size[2])*
((cor[2][zv] - op->cen[2])/op->size[2])) < 1.0) &&
((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2])) > 1.0))
intersect = 1;
}
}
} else {
if (yc) {
if (zc) { /* !xc, yc, zc */
if ((fabs(cor[0][xv] - op->cen[0]) < op->size[0]) &&
((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2])) > 1.0))
intersect = 1;
} else { /* !xc, yc, !zc */
if (((((cor[0][xv] - op->cen[0])/op->size[0])*
((cor[0][xv] - op->cen[0])/op->size[0]) +
((cor[2][zv] - op->cen[2])/op->size[2])*
((cor[2][zv] - op->cen[2])/op->size[2])) < 1.0) &&
((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2])) > 1.0))
intersect = 1;
}
} else {
if (zc) { /* !xc, !yc, zc */
if (((((cor[0][xv] - op->cen[0])/op->size[0])*
((cor[0][xv] - op->cen[0])/op->size[0]) +
((cor[1][yv] - op->cen[1])/op->size[1])*
((cor[1][yv] - op->cen[1])/op->size[1])) < 1.0) &&
((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2])) > 1.0))
intersect = 1;
} else { /* !xc, !yc, !zc */
if (((((cor[0][xv] - op->cen[0])/op->size[0])*
((cor[0][xv] - op->cen[0])/op->size[0]) +
((cor[1][yv] - op->cen[1])/op->size[1])*
((cor[1][yv] - op->cen[1])/op->size[1]) +
((cor[2][zv] - op->cen[2])/op->size[2])*
((cor[2][zv] - op->cen[2])/op->size[2])) < 1.0) &&
((((cor[0][1-xv] - op->cen[0])/op->size[0])*
((cor[0][1-xv] - op->cen[0])/op->size[0]) +
((cor[1][1-yv] - op->cen[1])/op->size[1])*
((cor[1][1-yv] - op->cen[1])/op->size[1]) +
((cor[2][1-zv] - op->cen[2])/op->size[2])*
((cor[2][1-zv] - op->cen[2])/op->size[2])) > 1.0))
intersect = 1;
}
}
}
if (op->type != 2 && intersect) {
/* exclude halfplane of spheroid to make hemispheroid */
if (op->type == 4 && cor[0][1] < op->cen[0])
intersect = tmp;
else if (op->type == 6 && cor[0][0] > op->cen[0])
intersect = tmp;
else if (op->type == 8 && cor[1][1] < op->cen[1])
intersect = tmp;
else if (op->type == 10 && cor[1][0] > op->cen[1])
intersect = tmp;
else if (op->type == 12 && cor[2][1] < op->cen[2])
intersect = tmp;
else if (op->type == 14 && cor[2][0] > op->cen[2])
intersect = tmp;
}
} else if (op->type >= 3 && op->type <= 15 && op->type%2) {
/* solid spheroid or hemispheroid */
/* determine if center is in block - if so then refine,
* if not determine nearest vertix and see if that is in
* - if so refine
* 1 = (x/a)^2 + (y/b)^2 + (z/c)^2 is boundary */
tmp = intersect;
xc = yc = zc = 0;
if (op->cen[0] < cor[0][0])
xv = 0;
else if (op->cen[0] > cor[0][1])
xv = 1;
else
xc = 1;
if (op->cen[1] < cor[1][0])
yv = 0;
else if (op->cen[1] > cor[1][1])
yv = 1;
else
yc = 1;
if (op->cen[2] < cor[2][0])
zv = 0;
else if (op->cen[2] > cor[2][1])
zv = 1;
else
zc = 1;
if (xc) {
if (yc) {
if (zc) /* xc, yc, zc */
intersect = 1;
else { /* xc, yc, !zc */
if (fabs(cor[2][zv] - op->cen[2]) < op->size[2])
intersect = 1;
}
} else {
if (zc) { /* xc, !yc, zc */
if (fabs(cor[1][yv] - op->cen[1]) < op->size[1])
intersect = 1;
} else { /* xc, !yc, !zc */
if ((((cor[1][yv] - op->cen[1])/op->size[1])*
((cor[1][yv] - op->cen[1])/op->size[1]) +
((cor[2][zv] - op->cen[2])/op->size[2])*
((cor[2][zv] - op->cen[2])/op->size[2])) < 1.0)
intersect = 1;
}
}
} else {
if (yc) {
if (zc) { /* !xc, yc, zc */
if (fabs(cor[0][xv] - op->cen[0]) < op->size[0])
intersect = 1;
} else { /* !xc, yc, !zc */
if ((((cor[0][xv] - op->cen[0])/op->size[0])*
((cor[0][xv] - op->cen[0])/op->size[0]) +
((cor[2][zv] - op->cen[2])/op->size[2])*
((cor[2][zv] - op->cen[2])/op->size[2])) < 1.0)
intersect = 1;
}
} else {
if (zc) { /* !xc, !yc, zc */
if ((((cor[0][xv] - op->cen[0])/op->size[0])*
((cor[0][xv] - op->cen[0])/op->size[0]) +
((cor[1][yv] - op->cen[1])/op->size[1])*
((cor[1][yv] - op->cen[1])/op->size[1])) < 1.0)
intersect = 1;
} else { /* !xc, !yc, !zc */
if ((((cor[0][xv] - op->cen[0])/op->size[0])*
((cor[0][xv] - op->cen[0])/op->size[0]) +
((cor[1][yv] - op->cen[1])/op->size[1])*
((cor[1][yv] - op->cen[1])/op->size[1]) +
((cor[2][zv] - op->cen[2])/op->size[2])*
((cor[2][zv] - op->cen[2])/op->size[2])) < 1.0)
intersect = 1;
}
}
}
if (op->type != 3 && intersect) {
/* exclude halfplane of spheroid to make hemispheroid */
if (op->type == 5 && cor[0][1] < op->cen[0])
intersect = tmp;
else if (op->type == 7 && cor[0][0] > op->cen[0])
intersect = tmp;
else if (op->type == 9 && cor[1][1] < op->cen[1])
intersect = tmp;
else if (op->type == 11 && cor[1][0] > op->cen[1])
intersect = tmp;
else if (op->type == 13 && cor[2][1] < op->cen[2])
intersect = tmp;
else if (op->type == 15 && cor[2][0] > op->cen[2])
intersect = tmp;
}
} else if (op->type == 20 || op->type == 22 || op->type == 24) {
/* boundary of cylinder, ca is axis of cylinder */
if (op->type == 20) {
ca = 0;
c1 = 1;
c2 = 2;
} else if (op->type == 22) {
ca = 1;
c1 = 2;
c2 = 0;
} else {
ca = 2;
c1 = 0;
c2 = 1;
}
if (cor[ca][1] > (op->cen[ca] - op->size[ca]) &&
cor[ca][0] < (op->cen[ca] + op->size[ca])) {
/* some part of block between planes that define ends */
/* use y and z for directions perpendicular to axis */
yc = zc = 0;
if (op->cen[c1] < cor[c1][0])
yv = 0;
else if (op->cen[c1] > cor[c1][1])
yv = 1;
else {
yc = 1;
if (op->cen[c1] < (cor[c1][0] + cor[c1][1])/2.0)
yv = 0;
else
yv = 1;
}
if (op->cen[c2] < cor[c2][0])
zv = 0;
else if (op->cen[c2] > cor[c2][1])
zv = 1;
else {
zc = 1;
if (op->cen[c2] < (cor[c2][0] + cor[c2][1])/2.0)
zv = 0;
else
zv = 1;
}
if ((cor[0][0] < (op->cen[0] - op->size[0]) &&
cor[0][1] < (op->cen[0] + op->size[0])) ||
(cor[0][0] > (op->cen[0] - op->size[0]) &&
cor[0][1] > (op->cen[0] + op->size[0]))) {
/* block overlaps cylinder ends in aixs dir */
if (yc) {
if (zc) {
intersect = 1;
} else {
if (fabs(cor[c2][zv] - op->cen[c2]) < op->size[c2])
intersect = 1;
}
} else {
if (zc) {
if (fabs(cor[c1][yv] - op->cen[c1]) < op->size[c1])
intersect = 1;
} else {
if ((((cor[c1][yv] - op->cen[c1])/op->size[c1])*
((cor[c1][yv] - op->cen[c1])/op->size[c1]) +
((cor[c2][zv] - op->cen[c2])/op->size[c2])*
((cor[c2][zv] - op->cen[c2])/op->size[c2]))
< 1.0)
intersect = 1;
}
}
} else {
/* block in cylinder or cylinder in block in aixs dir */
/* in c1 c2 plane need block point in and out of circle */
if (yc) {
if (zc) {
if ((((cor[c1][1-yv] - op->cen[c1])/op->size[c1])*
((cor[c1][1-yv] - op->cen[c1])/op->size[c1]) +
((cor[c2][1-zv] - op->cen[c2])/op->size[c2])*
((cor[c2][1-zv] - op->cen[c2])/op->size[c2]))
> 1.0)
intersect = 1;
} else {
if ((fabs(cor[c2][zv]-op->cen[c2]) < op->size[c2])&&
((((cor[c1][1-yv] - op->cen[c1])/op->size[c1])*
((cor[c1][1-yv] - op->cen[c1])/op->size[c1]) +
((cor[c2][1-zv] - op->cen[c2])/op->size[c2])*
((cor[c2][1-zv] - op->cen[c2])/op->size[c2]))
> 1.0))
intersect = 1;
}
} else {
if (zc) {
if ((fabs(cor[c1][yv]-op->cen[c1]) < op->size[c1])&&
((((cor[c1][1-yv] - op->cen[c1])/op->size[c1])*
((cor[c1][1-yv] - op->cen[c1])/op->size[c1]) +
((cor[c2][1-zv] - op->cen[c2])/op->size[c2])*
((cor[c2][1-zv] - op->cen[c2])/op->size[c2]))
> 1.0))
intersect = 1;
} else {
if (((((cor[c1][yv] - op->cen[c1])/op->size[c1])*
((cor[c1][yv] - op->cen[c1])/op->size[c1]) +
((cor[c2][zv] - op->cen[c2])/op->size[c2])*
((cor[c2][zv] - op->cen[c2])/op->size[c2]))
< 1.0) &&
((((cor[c1][1-yv] - op->cen[c1])/op->size[c1])*
((cor[c1][1-yv] - op->cen[c1])/op->size[c1]) +
((cor[c2][1-zv] - op->cen[c2])/op->size[c2])*
((cor[c2][1-zv] - op->cen[c2])/op->size[c2]))
> 1.0))
intersect = 1;
}
}
}
}
} else if (op->type == 21 || op->type == 23 || op->type == 25) {
/* volume of cylinder, ca is axis of cylinder */
if (op->type == 21) {
ca = 0;
c1 = 1;
c2 = 2;
} else if (op->type == 23) {
ca = 1;
c1 = 2;
c2 = 0;
} else {
ca = 2;
c1 = 0;
c2 = 1;
}
if (cor[ca][1] > (op->cen[ca] - op->size[ca]) &&
cor[ca][0] < (op->cen[ca] + op->size[ca])) {
/* some part of block between planes that define ends */
/* use y and z for directions perpendicular to axis */
yc = zc = 0;
if (op->cen[c1] < cor[c1][0])
yv = 0;
else if (op->cen[c1] > cor[c1][1])
yv = 1;
else
yc = 1;
if (op->cen[c2] < cor[c2][0])
zv = 0;
else if (op->cen[c2] > cor[c2][1])
zv = 1;
else
zc = 1;
if (yc) {
if (zc) {
intersect = 1;
} else {
if (fabs(cor[c2][zv] - op->cen[c2]) < op->size[c2])
intersect = 1;
}
} else {
if (zc) {
if (fabs(cor[c1][yv] - op->cen[c1]) < op->size[c1])
intersect = 1;
} else {
if ((((cor[c1][yv] - op->cen[c1])/op->size[c1])*
((cor[c1][yv] - op->cen[c1])/op->size[c1]) +
((cor[c2][zv] - op->cen[c2])/op->size[c2])*
((cor[c2][zv] - op->cen[c2])/op->size[c2])) < 1.0)
intersect = 1;
}
}
}
} else {
printf("undefined object %d\n", op->type);
}
}
return(intersect);
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/param.h

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
max_num_blocks = 500;
target_active = 0;
target_max = 0;
target_min = 0;
num_refine = 5;
uniform_refine = 0;
x_block_size = 10;
y_block_size = 10;
z_block_size = 10;
num_vars = 40;
comm_vars = 0;
init_block_x = 1;
init_block_y = 1;
init_block_z = 1;
reorder = 1;
npx = 1;
npy = 1;
npz = 1;
inbalance = 0;
refine_freq = 5;
report_diffusion = 0;
error_tol = 8;
num_tsteps = 20;
stages_per_ts = 20;
checksum_freq = 5;
stencil = 7;
report_perf = 12;
plot_freq = 0;
num_objects = 0;
lb_opt = 1;
block_change = 0;
code = 0;
permute = 0;
nonblocking = 1;
refine_ghost = 0;

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/plot.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdlib.h>
#include <stdio.h>
#include "block.h"
#include "proto.h"
// Write block information (level and center) to plot file.
void plot(int ts)
{
int i, j, n, total_num_blocks, *buf, buf_size, size;
char fname[20];
block *bp;
FILE *fp;
fname[0] = 'p';
fname[1] = 'l';
fname[2] = 'o';
fname[3] = 't';
fname[4] = '.';
for (n = 1, j = 0; n < num_tsteps; j++, n *= 10) ;
for (n = 1, i = 0; i <= j; i++, n *= 10)
fname[5+j-i] = (char) ('0' + (ts/n)%10);
fname[6+j] = '\0';
fp = fopen(fname, "w");
total_num_blocks = 0;
for (i = 0; i <= num_refine; i++)
total_num_blocks += num_blocks[i];
fprintf(fp, "%d %d %d %d %d\n", total_num_blocks, num_refine,
npx*init_block_x, npy*init_block_y,
npz*init_block_z);
buf_size = 0;
fprintf(fp, "%d\n", num_active);
for (n = 0; n < max_active_block; n++)
if ((bp = &blocks[n])->number >= 0)
fprintf(fp, "%d %d %d %d\n", bp->level, bp->cen[0],
bp->cen[1], bp->cen[2]);
fclose(fp);
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/profile.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdio.h>
#include <math.h>
#include "block.h"
#include "proto.h"
#include "timer.h"
// Profiling output.
void profile(void)
{
int i;
double total_gflops, gflops_rank, total_fp_ops, total_fp_adds,
total_fp_divs;
object *op;
char *version = "1.0 provisional";
FILE *fp;
calculate_results();
total_fp_divs = ((double) total_blocks)*((double) x_block_size)*
((double) y_block_size)*((double) z_block_size);
if (stencil == 7)
total_fp_adds = 6*total_fp_divs;
else
total_fp_adds = 26*total_fp_divs;
total_fp_ops = total_fp_divs + total_fp_adds;
total_gflops = total_fp_ops/(average[38]*1024.0*1024.0*1024.0);
gflops_rank = total_gflops/((double) num_pes);
if (!my_pe) {
if (report_perf & 1) {
fp = fopen("results.yaml", "w");
fprintf(fp, "code: miniAMR\n");
fprintf(fp, "version: %s\n", version);
fprintf(fp, "ranks: %d\n", num_pes);
fprintf(fp, "init_block_x: %d\n", init_block_x);
fprintf(fp, "init_block_y: %d\n", init_block_y);
fprintf(fp, "init_block_z: %d\n", init_block_z);
fprintf(fp, "x_block_size: %d\n", x_block_size);
fprintf(fp, "y_block_size: %d\n", y_block_size);
fprintf(fp, "z_block_size: %d\n", z_block_size);
fprintf(fp, "permute: %d\n", permute);
fprintf(fp, "max_blocks_allowed: %d\n", max_num_blocks);
fprintf(fp, "code: %d\n", code);
fprintf(fp, "num_refine: %d\n", num_refine);
fprintf(fp, "block_change: %d\n", block_change);
fprintf(fp, "refine_ghost: %d\n", refine_ghost);
fprintf(fp, "uniform_refine: %d\n", uniform_refine);
fprintf(fp, "num_objects: %d\n", num_objects);
for (i = 0; i < num_objects; i++) {
op = &objects[i];
fprintf(fp, "obj%dtype: %d\n", i, op->type);
fprintf(fp, "obj%dbounce: %d\n", i, op->bounce);
fprintf(fp, "obj%dcenter_x: %lf\n", i, op->cen[0]);
fprintf(fp, "obj%dcenter_y: %lf\n", i, op->cen[1]);
fprintf(fp, "obj%dcenter_z: %lf\n", i, op->cen[2]);
fprintf(fp, "obj%dmove_x: %lf\n", i, op->move[0]);
fprintf(fp, "obj%dmove_y: %lf\n", i, op->move[1]);
fprintf(fp, "obj%dmove_z: %lf\n", i, op->move[2]);
fprintf(fp, "obj%dsize_x: %lf\n", i, op->size[0]);
fprintf(fp, "obj%dsize_y: %lf\n", i, op->size[1]);
fprintf(fp, "obj%dsize_z: %lf\n", i, op->size[2]);
fprintf(fp, "obj%dinc_x: %lf\n", i, op->inc[0]);
fprintf(fp, "obj%dinc_y: %lf\n", i, op->inc[1]);
fprintf(fp, "obj%dinc_z: %lf\n", i, op->inc[2]);
}
fprintf(fp, "num_tsteps: %d\n", num_tsteps);
fprintf(fp, "stages_per_timestep: %d\n", stages_per_ts);
fprintf(fp, "checksum_freq: %d\n", checksum_freq);
fprintf(fp, "refine_freq: %d\n", refine_freq);
fprintf(fp, "plot_freq: %d\n", plot_freq);
fprintf(fp, "num_vars: %d\n", num_vars);
fprintf(fp, "stencil: %d\n", stencil);
fprintf(fp, "comm_vars: %d\n", comm_vars);
fprintf(fp, "error_tol: %d\n", error_tol);
fprintf(fp, "total_time: %lf\n", average[0]);
fprintf(fp, "memory_used: %lf\n", average[111]);
fprintf(fp, "compute_time: %lf\n", average[38]);
fprintf(fp, "total_gflops: %lf\n", total_gflops);
fprintf(fp, "ave_gflops: %lf\n", gflops_rank);
fprintf(fp, "total_comm: %lf\n", average[37]);
fprintf(fp, " total_exch_same: %lf\n", average[5]);
fprintf(fp, " total_exch_diff: %lf\n", average[6]);
fprintf(fp, " total_apply_bc: %lf\n", average[7]);
fprintf(fp, " total_face_exch_same: %lf\n", average[75]);
fprintf(fp, " total_face_exch_diff: %lf\n", average[76]);
fprintf(fp, " total_face_bc_apply: %lf\n", average[74]);
fprintf(fp, " x_comm: %lf\n", average[10]);
fprintf(fp, " x_exch_same: %lf\n", average[14]);
fprintf(fp, " x_exch_diff: %lf\n", average[15]);
fprintf(fp, " x_apply_bc: %lf\n", average[16]);
fprintf(fp, " x_face_exch_same: %lf\n", average[84]);
fprintf(fp, " x_face_exch_diff: %lf\n", average[85]);
fprintf(fp, " x_face_bc_apply: %lf\n", average[83]);
fprintf(fp, " y_comm: %lf\n", average[19]);
fprintf(fp, " y_exch_same: %lf\n", average[23]);
fprintf(fp, " y_exch_diff: %lf\n", average[24]);
fprintf(fp, " y_apply_bc: %lf\n", average[25]);
fprintf(fp, " y_face_exch_same: %lf\n", average[93]);
fprintf(fp, " y_face_exch_diff: %lf\n", average[94]);
fprintf(fp, " y_face_bc_apply: %lf\n", average[92]);
fprintf(fp, " z_comm: %lf\n", average[28]);
fprintf(fp, " z_exch_same: %lf\n", average[32]);
fprintf(fp, " z_exch_diff: %lf\n", average[33]);
fprintf(fp, " z_apply_bc: %lf\n", average[34]);
fprintf(fp, " z_face_exch_same: %lf\n", average[102]);
fprintf(fp, " z_face_exch_diff: %lf\n", average[103]);
fprintf(fp, " z_face_bc_apply: %lf\n", average[101]);
fprintf(fp, "gridsum_time: %lf\n", average[39]);
fprintf(fp, " gridsum_calc: %lf\n", average[41]);
fprintf(fp, "refine_time: %lf\n", average[42]);
fprintf(fp, " total_blocks_ts: %lf\n",
((double) total_blocks)/((double) (num_tsteps*stages_per_ts)));
fprintf(fp, " total_blocks_ts_min: %d\n", nb_min);
fprintf(fp, " total_blocks_ts_max: %d\n", nb_max);
fprintf(fp, " blocks_split: %lf\n", average[104]);
fprintf(fp, " blocks_reformed: %lf\n", average[105]);
fprintf(fp, " time_compare_obj: %lf\n", average[43]);
fprintf(fp, " time_mark_refine: %lf\n", average[44]);
fprintf(fp, " time_split_block: %lf\n", average[46]);
fprintf(fp, " time_total_coarsen: %lf\n", average[47]);
fprintf(fp, " time_misc: %lf\n", average[45]);
fprintf(fp, "plot_time: %lf\n", average[67]);
fclose(fp);
}
if (report_perf & 2) {
fp = fopen("results.txt", "w");
fprintf(fp, "\n ================ Start report ===================\n\n");
fprintf(fp, " Mantevo miniAMR\n");
fprintf(fp, " version %s\n\n", version);
fprintf(fp, "serial run on 1 rank\n");
fprintf(fp, "initial blocks per rank %d x %d x %d\n", init_block_x,
init_block_y, init_block_z);
fprintf(fp, "block size %d x %d x %d\n", x_block_size, y_block_size,
z_block_size);
if (permute)
fprintf(fp, "Order of exchanges permuted\n");
fprintf(fp, "Maximum number of blocks per rank is %d\n",
max_num_blocks);
if (target_active)
fprintf(fp, "Target number of blocks per rank is %d\n",
target_active);
if (target_max)
fprintf(fp, "Target max number of blocks per rank is %d\n",
target_max);
if (target_min)
fprintf(fp, "Target min number of blocks per rank is %d\n",
target_min);
if (code)
fprintf(fp, "Code set to code %d\n", code);
fprintf(fp, "Number of levels of refinement is %d\n", num_refine);
fprintf(fp, "Blocks can change by %d levels per refinement step\n",
block_change);
if (refine_ghost)
fprintf(fp, "Ghost cells will be used determine is block is refined\n");
if (uniform_refine)
fprintf(fp, "\nBlocks will be uniformly refined\n");
else {
fprintf(fp, "\nBlocks will be refined by %d objects\n\n", num_objects);
for (i = 0; i < num_objects; i++) {
op = &objects[i];
if (op->type == 0)
fprintf(fp, "Object %d is the surface of a rectangle\n", i);
else if (op->type == 1)
fprintf(fp, "Object %d is the volume of a rectangle\n", i);
else if (op->type == 2)
fprintf(fp, "Object %d is the surface of a spheroid\n", i);
else if (op->type == 3)
fprintf(fp, "Object %d is the volume of a spheroid\n", i);
else if (op->type == 4)
fprintf(fp, "Object %d is the surface of x+ hemispheroid\n", i);
else if (op->type == 5)
fprintf(fp, "Object %d is the volume of x+ hemispheroid\n", i);
else if (op->type == 6)
fprintf(fp, "Object %d is the surface of x- hemispheroid\n", i);
else if (op->type == 7)
fprintf(fp, "Object %d is the volume of x- hemispheroid\n", i);
else if (op->type == 8)
fprintf(fp, "Object %d is the surface of y+ hemispheroid\n", i);
else if (op->type == 9)
fprintf(fp, "Object %d is the volume of y+ hemispheroid\n", i);
else if (op->type == 10)
fprintf(fp, "Object %d is the surface of y- hemispheroid\n", i);
else if (op->type == 11)
fprintf(fp, "Object %d is the volume of y- hemispheroid\n", i);
else if (op->type == 12)
fprintf(fp, "Object %d is the surface of z+ hemispheroid\n", i);
else if (op->type == 13)
fprintf(fp, "Object %d is the volume of z+ hemispheroid\n", i);
else if (op->type == 14)
fprintf(fp, "Object %d is the surface of z- hemispheroid\n", i);
else if (op->type == 15)
fprintf(fp, "Object %d is the volume of z- hemispheroid\n", i);
else if (op->type == 20)
fprintf(fp, "Object %d is the surface of x axis cylinder\n", i);
else if (op->type == 21)
fprintf(fp, "Object %d is the volune of x axis cylinder\n", i);
else if (op->type == 22)
fprintf(fp, "Object %d is the surface of y axis cylinder\n", i);
else if (op->type == 23)
fprintf(fp, "Object %d is the volune of y axis cylinder\n", i);
else if (op->type == 24)
fprintf(fp, "Object %d is the surface of z axis cylinder\n", i);
else if (op->type == 25)
fprintf(fp, "Object %d is the volune of z axis cylinder\n", i);
if (op->bounce == 0)
fprintf(fp, "Oject may leave mesh\n");
else
fprintf(fp, "Oject center will bounce off of walls\n");
fprintf(fp, "Center starting at %lf %lf %lf\n",
op->orig_cen[0], op->orig_cen[1], op->orig_cen[2]);
fprintf(fp, "Center end at %lf %lf %lf\n",
op->cen[0], op->cen[1], op->cen[2]);
fprintf(fp, "Moving at %lf %lf %lf per timestep\n",
op->orig_move[0], op->orig_move[1], op->orig_move[2]);
fprintf(fp, " Rate relative to smallest cell size %lf %lf %lf\n",
op->orig_move[0]*((double) (mesh_size[0]*x_block_size)),
op->orig_move[1]*((double) (mesh_size[1]*y_block_size)),
op->orig_move[2]*((double) (mesh_size[2]*z_block_size)));
fprintf(fp, "Initial size %lf %lf %lf\n",
op->orig_size[0], op->orig_size[1], op->orig_size[2]);
fprintf(fp, "Final size %lf %lf %lf\n",
op->size[0], op->size[1], op->size[2]);
fprintf(fp, "Size increasing %lf %lf %lf per timestep\n",
op->inc[0], op->inc[1], op->inc[2]);
fprintf(fp, " Rate relative to smallest cell size %lf %lf %lf\n\n",
op->inc[0]*((double) (mesh_size[0]*x_block_size)),
op->inc[1]*((double) (mesh_size[1]*y_block_size)),
op->inc[2]*((double) (mesh_size[2]*z_block_size)));
}
}
fprintf(fp, "\nNumber of timesteps is %d\n", num_tsteps);
fprintf(fp, "Communicaion/computation stages per timestep is %d\n",
stages_per_ts);
fprintf(fp, "Will perform checksums every %d stages\n", checksum_freq);
fprintf(fp, "Will refine every %d timesteps\n", refine_freq);
if (plot_freq)
fprintf(fp, "Will plot results every %d timesteps\n", plot_freq);
else
fprintf(fp, "Will not plot results\n");
fprintf(fp, "Calculate on %d variables with %d point stencil\n",
num_vars, stencil);
fprintf(fp, "Communicate %d variables at a time\n", comm_vars);
fprintf(fp, "Error tolorance for variable sums is 10^(-%d)\n", error_tol);
fprintf(fp, "\nTotal time for test: (sec): %lf\n\n", average[0]);
fprintf(fp, "\nNumber of malloc calls: %lf\n", average[110]);
fprintf(fp, "\nAmount malloced: %lf\n", average[111]);
fprintf(fp, "---------------------------------------------\n");
fprintf(fp, " Computational Performance\n");
fprintf(fp, "---------------------------------------------\n\n");
fprintf(fp, " Time: ave, stddev, min, max (sec): %lf\n\n",
average[38]);
fprintf(fp, " total GFLOPS: %lf\n", total_gflops);
fprintf(fp, " Average GFLOPS per rank: %lf\n\n", gflops_rank);
fprintf(fp, " Total floating point ops: %lf\n\n", total_fp_ops);
fprintf(fp, " Adds: %lf\n", total_fp_adds);
fprintf(fp, " Divides: %lf\n\n", total_fp_divs);
fprintf(fp, "---------------------------------------------\n");
fprintf(fp, " Interblock communication\n");
fprintf(fp, "---------------------------------------------\n\n");
fprintf(fp, " Time: ave, stddev, min, max (sec): %lf\n\n",
average[37]);
for (i = 0; i < 4; i++) {
if (i == 0)
fprintf(fp, "\nTotal communication:\n\n");
else if (i == 1)
fprintf(fp, "\nX direction communication statistics:\n\n");
else if (i == 2)
fprintf(fp, "\nY direction communication statistics:\n\n");
else
fprintf(fp, "\nZ direction communication statistics:\n\n");
fprintf(fp, " average stddev minimum maximum\n");
fprintf(fp, " Total time : %lf\n", average[1+9*i]);
fprintf(fp, " Exchange same level : %lf\n", average[5+9*i]);
fprintf(fp, " Exchange diff level : %lf\n", average[6+9*i]);
fprintf(fp, " Apply BC : %lf\n", average[7+9*i]);
fprintf(fp, " Faces exchanged same : %lf\n", average[75+9*i]);
fprintf(fp, " Faces exchanged diff : %lf\n", average[76+9*i]);
fprintf(fp, " Faces with BC applied : %lf\n", average[74+9*i]);
}
fprintf(fp, "\n---------------------------------------------\n");
fprintf(fp, " Gridsum performance\n");
fprintf(fp, "---------------------------------------------\n\n");
fprintf(fp, " Time: ave, stddev, min, max (sec): %lf\n\n",
average[39]);
fprintf(fp, " calc: ave, stddev, min, max (sec): %lf\n\n",
average[41]);
fprintf(fp, " total number: %d\n", total_red);
fprintf(fp, " number per timestep: %d\n\n", num_vars);
fprintf(fp, "---------------------------------------------\n");
fprintf(fp, " Mesh Refinement\n");
fprintf(fp, "---------------------------------------------\n\n");
fprintf(fp, " Time: ave, stddev, min, max (sec): %lf\n\n",
average[42]);
fprintf(fp, " Number of refinement steps: %d\n\n", nrs);
fprintf(fp, " Total blocks : %ld\n", total_blocks);
fprintf(fp, " Blocks/timestep ave, min, max : %lf %d %d\n",
((double) total_blocks)/((double) (num_tsteps*stages_per_ts)),
nb_min, nb_max);
fprintf(fp, " Max blocks on a processor at any time: %d\n",
global_max_b);
fprintf(fp, " total blocks split : %lf\n", average[104]);
fprintf(fp, " total blocks reformed : %lf\n\n", average[105]);
fprintf(fp, " Time:\n");
fprintf(fp, " compare objects : %lf\n", average[43]);
fprintf(fp, " mark refine/coarsen : %lf\n", average[44]);
fprintf(fp, " split blocks : %lf\n", average[46]);
fprintf(fp, " total coarsen blocks: %lf\n", average[47]);
fprintf(fp, " misc time : %lf\n", average[45]);
if (target_active) {
fprintf(fp, " total target active : %lf\n", average[52]);
fprintf(fp, " reduce blocks : %lf\n", average[53]);
fprintf(fp, " decide and comm : %lf\n", average[54]);
fprintf(fp, " coarsen blocks : %lf\n", average[58]);
fprintf(fp, " add blocks : %lf\n", average[59]);
fprintf(fp, " decide and comm : %lf\n", average[60]);
fprintf(fp, " split blocks : %lf\n", average[61]);
}
fprintf(fp, "---------------------------------------------\n");
fprintf(fp, " Plot\n");
fprintf(fp, "---------------------------------------------\n\n");
fprintf(fp, " Time: ave, stddev, min, max (sec): %lf\n\n",
average[67]);
fprintf(fp, " Number of plot steps: %d\n", nps);
fprintf(fp, "\n ================== End report ===================\n");
fclose(fp);
}
if (report_perf & 4) {
printf("\n ================ Start report ===================\n\n");
printf(" Mantevo miniAMR\n");
printf(" version %s\n\n", version);
printf("serial run on 1 rank\n");
printf("initial blocks per rank %d x %d x %d\n", init_block_x,
init_block_y, init_block_z);
printf("block size %d x %d x %d\n", x_block_size, y_block_size,
z_block_size);
if (permute)
printf("Order of exchanges permuted\n");
printf("Maximum number of blocks per rank is %d\n", max_num_blocks);
if (target_active)
printf("Target number of blocks per rank is %d\n", target_active);
if (target_max)
printf("Target max number of blocks per rank is %d\n", target_max);
if (target_min)
printf("Target min number of blocks per rank is %d\n", target_min);
if (code)
printf("Code set to code %d\n", code);
printf("Number of levels of refinement is %d\n", num_refine);
printf("Blocks can change by %d levels per refinement step\n",
block_change);
if (refine_ghost)
printf("Ghost cells will be used determine is block is refined\n");
if (uniform_refine)
printf("\nBlocks will be uniformly refined\n");
else {
printf("\nBlocks will be refined by %d objects\n\n", num_objects);
for (i = 0; i < num_objects; i++) {
op = &objects[i];
if (op->type == 0)
printf("Object %d is the surface of a rectangle\n", i);
else if (op->type == 1)
printf("Object %d is the volume of a rectangle\n", i);
else if (op->type == 2)
printf("Object %d is the surface of a spheroid\n", i);
else if (op->type == 3)
printf("Object %d is the volume of a spheroid\n", i);
else if (op->type == 4)
printf("Object %d is the surface of x+ hemispheroid\n", i);
else if (op->type == 5)
printf("Object %d is the volume of x+ hemispheroid\n", i);
else if (op->type == 6)
printf("Object %d is the surface of x- hemispheroid\n", i);
else if (op->type == 7)
printf("Object %d is the volume of x- hemispheroid\n", i);
else if (op->type == 8)
printf("Object %d is the surface of y+ hemispheroid\n", i);
else if (op->type == 9)
printf("Object %d is the volume of y+ hemispheroid\n", i);
else if (op->type == 10)
printf("Object %d is the surface of y- hemispheroid\n", i);
else if (op->type == 11)
printf("Object %d is the volume of y- hemispheroid\n", i);
else if (op->type == 12)
printf("Object %d is the surface of z+ hemispheroid\n", i);
else if (op->type == 13)
printf("Object %d is the volume of z+ hemispheroid\n", i);
else if (op->type == 14)
printf("Object %d is the surface of z- hemispheroid\n", i);
else if (op->type == 15)
printf("Object %d is the volume of z- hemispheroid\n", i);
else if (op->type == 20)
printf("Object %d is the surface of x axis cylinder\n", i);
else if (op->type == 21)
printf("Object %d is the volune of x axis cylinder\n", i);
else if (op->type == 22)
printf("Object %d is the surface of y axis cylinder\n", i);
else if (op->type == 23)
printf("Object %d is the volune of y axis cylinder\n", i);
else if (op->type == 24)
printf("Object %d is the surface of z axis cylinder\n", i);
else if (op->type == 25)
printf("Object %d is the volune of z axis cylinder\n", i);
if (op->bounce == 0)
printf("Oject may leave mesh\n");
else
printf("Oject center will bounce off of walls\n");
printf("Center starting at %lf %lf %lf\n",
op->orig_cen[0], op->orig_cen[1], op->orig_cen[2]);
printf("Center end at %lf %lf %lf\n",
op->cen[0], op->cen[1], op->cen[2]);
printf("Moving at %lf %lf %lf per timestep\n",
op->orig_move[0], op->orig_move[1], op->orig_move[2]);
printf(" Rate relative to smallest cell size %lf %lf %lf\n",
op->orig_move[0]*((double) (mesh_size[0]*x_block_size)),
op->orig_move[1]*((double) (mesh_size[1]*y_block_size)),
op->orig_move[2]*((double) (mesh_size[2]*z_block_size)));
printf("Initial size %lf %lf %lf\n",
op->orig_size[0], op->orig_size[1], op->orig_size[2]);
printf("Final size %lf %lf %lf\n",
op->size[0], op->size[1], op->size[2]);
printf("Size increasing %lf %lf %lf per timestep\n",
op->inc[0], op->inc[1], op->inc[2]);
printf(" Rate relative to smallest cell size %lf %lf %lf\n\n",
op->inc[0]*((double) (mesh_size[0]*x_block_size)),
op->inc[1]*((double) (mesh_size[1]*y_block_size)),
op->inc[2]*((double) (mesh_size[2]*z_block_size)));
}
}
printf("\nNumber of timesteps is %d\n", num_tsteps);
printf("Communicaion/computation stages per timestep is %d\n",
stages_per_ts);
printf("Will perform checksums every %d stages\n", checksum_freq);
printf("Will refine every %d timesteps\n", refine_freq);
if (plot_freq)
printf("Will plot results every %d timesteps\n", plot_freq);
else
printf("Will not plot results\n");
printf("Calculate on %d variables with %d point stencil\n",
num_vars, stencil);
printf("Communicate %d variables at a time\n", comm_vars);
printf("Error tolorance for variable sums is 10^(-%d)\n", error_tol);
/*
printf("\nTotal time for test (sec): %lf\n\n", average[0]);
printf("\nNumber of malloc calls: %lf\n", average[110]);
printf("\nAmount malloced: %lf\n", average[111]);
printf("---------------------------------------------\n");
printf(" Computational Performance\n");
printf("---------------------------------------------\n\n");
printf(" Time: ave, stddev, min, max (sec): %lf\n\n", average[38]);
printf(" total GFLOPS: %lf\n", total_gflops);
printf(" Average GFLOPS per rank: %lf\n\n", gflops_rank);
printf(" Total floating point ops: %lf\n\n", total_fp_ops);
printf(" Adds: %lf\n", total_fp_adds);
printf(" Divides: %lf\n\n", total_fp_divs);
printf("---------------------------------------------\n");
printf(" Interblock communication\n");
printf("---------------------------------------------\n\n");
printf(" Time: ave, stddev, min, max (sec): %lf\n\n", average[37]);
for (i = 0; i < 4; i++) {
if (i == 0)
printf("\nTotal communication:\n\n");
else if (i == 1)
printf("\nX direction communication statistics:\n\n");
else if (i == 2)
printf("\nY direction communication statistics:\n\n");
else
printf("\nZ direction communication statistics:\n\n");
printf(" Total time : %lf\n", average[1+9*i]);
printf(" Exchange same level : %lf\n", average[5+9*i]);
printf(" Exchange diff level : %lf\n", average[6+9*i]);
printf(" Apply BC : %lf\n", average[7+9*i]);
printf(" Faces exchanged same : %lf\n", average[75+9*i]);
printf(" Faces exchanged diff : %lf\n", average[76+9*i]);
printf(" Faces with BC applied : %lf\n", average[74+9*i]);
}
printf("\n---------------------------------------------\n");
printf(" Gridsum performance\n");
printf("---------------------------------------------\n\n");
printf(" Time: ave, stddev, min, max (sec): %lf\n\n", average[39]);
printf(" calc: ave, stddev, min, max (sec): %lf\n\n",
average[41]);
printf(" total number: %d\n", total_red);
printf(" number per timestep: %d\n\n", num_vars);
printf("---------------------------------------------\n");
printf(" Mesh Refinement\n");
printf("---------------------------------------------\n\n");
printf(" Time: ave, stddev, min, max (sec): %lf\n\n", average[42]);
printf(" Number of refinement steps: %d\n\n", nrs);
printf(" Total blocks : %ld\n", total_blocks);
printf(" Blocks/timestep ave, min, max : %lf %d %d\n",
((double) total_blocks)/((double) (num_tsteps*stages_per_ts)),
nb_min, nb_max);
printf(" Max blocks on a processor at any time: %d\n",
global_max_b);
printf(" total blocks split : %lf\n", average[104]);
printf(" total blocks reformed : %lf\n\n", average[105]);
printf(" Time:\n");
printf(" compare objects : %lf\n", average[43]);
printf(" mark refine/coarsen : %lf\n", average[44]);
printf(" split blocks : %lf\n", average[46]);
printf(" total coarsen blocks: %lf\n", average[47]);
printf(" misc time : %lf\n", average[45]);
if (target_active) {
printf(" total target active : %lf\n", average[52]);
printf(" reduce blocks : %lf\n", average[53]);
printf(" decide and comm : %lf\n", average[54]);
printf(" coarsen blocks : %lf\n", average[58]);
printf(" add blocks : %lf\n", average[59]);
printf(" decide and comm : %lf\n", average[60]);
printf(" split blocks : %lf\n", average[61]);
}
printf("---------------------------------------------\n");
printf(" Plot\n");
printf("---------------------------------------------\n\n");
printf(" Time: ave, stddev, min, max (sec): %lf\n\n", average[67]);
printf(" Number of plot steps: %d\n", nps);
printf("\n ================== End report ===================\n");
*/
}
}
}
void calculate_results(void)
{
double results[128];
int i;
results[0] = timer_all;
for (i = 0; i < 9; i++)
results[i+1] = 0.0;
for (i = 0; i < 3; i++) {
results[1] += results[10+9*i] = timer_comm_dir[i];
results[5] += results[14+9*i] = timer_comm_same[i];
results[6] += results[15+9*i] = timer_comm_diff[i];
results[7] += results[16+9*i] = timer_comm_bc[i];
}
results[37] = timer_comm_all;
results[38] = timer_calc_all;
results[39] = timer_cs_all;
results[41] = timer_cs_calc;
results[42] = timer_refine_all;
results[43] = timer_refine_co;
results[44] = timer_refine_mr;
results[45] = timer_refine_cc;
results[46] = timer_refine_sb;
results[47] = timer_cb_all;
results[52] = timer_target_all;
results[53] = timer_target_rb;
results[54] = timer_target_dc;
results[58] = timer_target_cb;
results[59] = timer_target_ab;
results[60] = timer_target_da;
results[61] = timer_target_sb;
results[67] = timer_plot;
for (i = 0; i < 9; i++)
results[68+i] = 0.0;
for (i = 0; i < 3; i++) {
results[74] += results[83+9*i] = (double) counter_bc[i];
results[75] += results[84+9*i] = (double) counter_same[i];
results[76] += results[85+9*i] = (double) counter_diff[i];
}
results[104] = (double) num_refined;
results[105] = (double) num_reformed;
results[110] = (double) counter_malloc;
results[111] = size_malloc;
results[112] = (double) counter_malloc_init;
results[113] = size_malloc_init;
results[114] = (double) (counter_malloc - counter_malloc_init);
results[115] = size_malloc - size_malloc_init;
for (i = 0; i < 128; i++)
average[i] = results[i];
}
void init_profile(void)
{
int i;
timer_all = 0.0;
timer_comm_all = 0.0;
for (i = 0; i < 3; i++) {
timer_comm_dir[i] = 0.0;
timer_comm_same[i] = 0.0;
timer_comm_diff[i] = 0.0;
timer_comm_bc[i] = 0.0;
}
timer_calc_all = 0.0;
timer_cs_all = 0.0;
timer_cs_calc = 0.0;
timer_refine_all = 0.0;
timer_refine_co = 0.0;
timer_refine_mr = 0.0;
timer_refine_cc = 0.0;
timer_refine_sb = 0.0;
timer_cb_all = 0.0;
timer_target_all = 0.0;
timer_target_rb = 0.0;
timer_target_dc = 0.0;
timer_target_cb = 0.0;
timer_target_ab = 0.0;
timer_target_da = 0.0;
timer_target_sb = 0.0;
timer_plot = 0.0;
total_blocks = 0;
nrs = 0;
nps = 0;
num_refined = 0;
num_reformed = 0;
for (i = 0; i < 3; i++) {
counter_bc[i] = 0;
counter_same[i] = 0;
counter_diff[i] = 0;
}
total_red = 0;
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/proto.h

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stddef.h>
// main.c
void print_help_message(void);
void allocate(void);
void deallocate(void);
int check_input(void);
// block.c
void split_blocks(void);
void consolidate_blocks(void);
void add_sorted_list(int, int, int);
void del_sorted_list(int, int);
int find_sorted_list(int, int);
// check_sum.c
double check_sum(int);
// comm.c
void comm(int, int, int);
void on_proc_comm(int, int, int, int, int);
void on_proc_comm_diff(int, int, int, int, int, int, int);
void apply_bc(int, block *, int, int);
// driver.c
void driver(void);
// init.c
void init(void);
// move.c
void move(void);
void check_objects(void);
int check_block(double cor[3][2]);
// plot.c
void plot(int);
// profile.c
void profile(void);
void calculate_results(void);
void init_profile(void);
// refine.c
void refine(int);
int refine_level(void);
void reset_all(void);
// stencil.c
void stencil_calc(int);
// target.c
int reduce_blocks();
void add_blocks();
void zero_refine(void);
// util.c
double timer(void);
void *ma_malloc(size_t, char *, int);
// debug.c
void print_par(void);
void print_comm(int);
void print_block(int, int);
void print_blocks(int);
void print_parents(int);

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/refine.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdio.h>
#include "block.h"
#include "proto.h"
#include "timer.h"
// This file contains routines that determine which blocks are going to
// be refined and which are going to be coarsened.
void refine(int ts)
{
int i, j, n, in, min_b, max_b, sum_b, num_refine_step, num_split,
nm_r, nm_c, nm_t;
double ratio, t1, t2, t3, t4, t5;
block *bp;
t4 = 0.0;
t1 = timer();
if (ts)
num_refine_step = block_change;
else
num_refine_step = num_refine;
for (i = 0; i < num_refine_step; i++) {
for (j = num_refine; j >= 0; j--)
if (num_blocks[j]) {
cur_max_level = j;
break;
}
reset_all();
if (uniform_refine) {
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
bp = &blocks[n];
if (bp->number >= 0)
if (bp->level < num_refine)
bp->refine = 1;
else
bp->refine = 0;
}
} else {
t2 = timer();
check_objects();
timer_refine_co += timer() - t2;
t4 += timer() - t2;
}
t2 = timer();
num_split = refine_level();
t5 = timer();
timer_refine_mr += t5 - t2;
t4 += t5 - t2;
t2 = timer();
split_blocks();
t5 = timer();
timer_refine_sb += t5 - t2;
t4 += t5 - t2;
t2 = timer();
consolidate_blocks();
t5 = timer();
timer_cb_all += t5 - t2;
t4 += t5 - t2;
}
if (target_active || target_max || target_min) {
if (!my_pe) {
for (j = 0; j <= num_refine; j++)
printf("Number of blocks at level %d before target %d is %d\n",
j, ts, num_blocks[j]);
printf("\n");
}
t2 = timer();
global_active = num_blocks[0];
for (i = 1; i <= num_refine; i++)
global_active += num_blocks[i];
// Will not be able to get to target in all cases, but can get to
// a range of target +- 3 since can add or subtract in units of
// 7 blocks.
if ((target_active && global_active > num_pes*target_active + 3) ||
(target_max && global_active > num_pes*target_max))
nm_t += reduce_blocks();
else if ((target_active && global_active < num_pes*target_active - 3) ||
(target_min && global_active < num_pes*target_min))
add_blocks();
t5 = timer();
timer_target_all += t5 - t2;
t4 += t5 - t2;
}
t2 = timer();
if (num_active > local_max_b)
local_max_b = num_active;
if (local_max_b > global_max_b)
global_max_b = local_max_b;
for (j = 0; j <= num_refine; j++) {
if (!j)
global_active = num_blocks[0];
else
global_active += num_blocks[j];
if (!my_pe && report_perf & 8)
printf("Number of blocks at level %d at timestep %d is %d\n",
j, ts, num_blocks[j]);
}
if (!my_pe && report_perf & 8) printf("\n");
t4 += timer() - t2;
t5 = timer();
timer_refine_cc += t5 - t1 - t4;
}
int refine_level(void)
{
int level, nei, n, i, j, b, c, c1, change, unrefine, sib, p, in;
block *bp, *bp1;
parent *pp;
/* block states:
* 1 block should be refined
* -1 block could be unrefined
* 0 block at level 0 and can not be unrefined or
* at max level and can not be refined
*/
// get list of neighbor blocks (indirect links in from blocks)
for (level = cur_max_level; level >= 0; level--) {
/* check for blocks at this level that will refine
their neighbors at this level can not unrefine
their neighbors at a lower level must refine
*/
do {
change = 0;
for (in = sorted_index[level]; in < sorted_index[level+1]; in++) {
n = sorted_list[in].n;
bp = &blocks[n];
if (bp->number >= 0 && bp->level == level) {
if (bp->refine == 1) {
if (bp->parent != -1 && bp->parent_node == my_pe) {
pp = &parents[bp->parent];
if (pp->refine == -1)
pp->refine = 0;
for (b = 0; b < 8; b++)
if (pp->child_node[b] == my_pe && pp->child[b] >= 0)
if (blocks[pp->child[b]].refine == -1) {
blocks[pp->child[b]].refine = 0;
change++;
}
}
for (i = 0; i < 6; i++)
/* neighbors in level above taken care of already */
/* neighbors in this level can not unrefine */
if (bp->nei_level[i] == level) {
if ((nei = bp->nei[i][0][0]) >= 0) { /* on core */
if (blocks[nei].refine == -1) {
blocks[nei].refine = 0;
change++;
if ((p = blocks[nei].parent) != -1 &&
blocks[nei].parent_node == my_pe) {
if ((pp = &parents[p])->refine == -1)
pp->refine = 0;
for (b = 0; b < 8; b++)
if (pp->child_node[b] == my_pe &&
pp->child[b] >= 0)
if (blocks[pp->child[b]].refine == -1) {
blocks[pp->child[b]].refine = 0;
change++;
}
}
}
}
/* neighbors in level below must refine */
} else if (bp->nei_level[i] == level-1)
if ((nei = bp->nei[i][0][0]) >= 0) {
if (blocks[nei].refine != 1) {
blocks[nei].refine = 1;
change++;
}
}
} else if (bp->refine == -1) {
// check if block can be unrefined
for (i = 0; i < 6; i++)
if (bp->nei_level[i] == level+1) {
bp->refine = 0;
change++;
if ((p = bp->parent) != -1 &&
bp->parent_node == my_pe) {
if ((pp = &parents[p])->refine == -1)
pp->refine = 0;
for (b = 0; b < 8; b++)
if (pp->child_node[b] == my_pe &&
pp->child[b] >= 0 &&
blocks[pp->child[b]].refine == -1)
blocks[pp->child[b]].refine = 0;
}
}
}
}
}
} while (change);
/* Check for blocks at this level that will remain at this level
their neighbors at a lower level can not unrefine
*/
do {
change = 0;
for (in = sorted_index[level]; in < sorted_index[level+1]; in++) {
n = sorted_list[in].n;
bp = &blocks[n];
if (bp->number >= 0)
if (bp->level == level && bp->refine == 0)
for (c = 0; c < 6; c++)
if (bp->nei_level[c] == level-1) {
if ((nei = bp->nei[c][0][0]) >= 0) {
if (blocks[nei].refine == -1) {
blocks[nei].refine = 0;
change++;
if ((p = blocks[nei].parent) != -1 &&
blocks[nei].parent_node == my_pe)
if ((pp = &parents[p])->refine == -1) {
pp->refine = 0;
for (b = 0; b < 8; b++)
if (pp->child_node[b] == my_pe &&
pp->child[b] >= 0 &&
blocks[pp->child[b]].refine == -1)
blocks[pp->child[b]].refine = 0;
}
}
}
} else if (bp->nei_level[c] == level) {
if ((nei = bp->nei[c][0][0]) >= 0)
blocks[nei].nei_refine[(c/2)*2+(c+1)%2] = 0;
} else if (bp->nei_level[c] == level+1) {
c1 = (c/2)*2 + (c+1)%2;
for (i = 0; i < 2; i++)
for (j = 0; j < 2; j++)
if ((nei = bp->nei[c][i][j]) >= 0)
blocks[nei].nei_refine[c1] = 0;
}
}
} while (change);
}
for (i = in = 0; in < sorted_index[num_refine+1]; in++)
if (blocks[sorted_list[in].n].refine == 1)
i++;
return(i);
}
// Reset the neighbor lists on blocks so that matching them against objects
// can set those which can be refined.
void reset_all(void)
{
int n, c, in;
block *bp;
parent *pp;
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
if ((bp= &blocks[n])->number >= 0)
bp->refine = -1;
}
for (n = 0; n < max_active_parent; n++)
if ((pp = &parents[n])->number >= 0) {
pp->refine = -1;
for (c = 0; c < 8; c++)
if (pp->child[c] < 0)
pp->refine = 0;
if (pp->refine == 0)
for (c = 0; c < 8; c++)
if (pp->child[c] >= 0)
if (blocks[pp->child[c]].refine == -1)
blocks[pp->child[c]].refine = 0;
}
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/stencil.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include "block.h"
#include "proto.h"
// This routine does the stencil calculations.
void stencil_calc(int var)
{
int n, i, j, k, in;
double sb, sm, sf, work[x_block_size+2][y_block_size+2][z_block_size+2];
block *bp;
if (stencil == 7) {
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
bp = &blocks[n];
if (bp->number >= 0) {
for (i = 1; i <= x_block_size; i++)
for (j = 1; j <= y_block_size; j++)
for (k = 1; k <= z_block_size; k++)
work[i][j][k] = (bp->array[var][i-1][j ][k ] +
bp->array[var][i ][j-1][k ] +
bp->array[var][i ][j ][k-1] +
bp->array[var][i ][j ][k ] +
bp->array[var][i ][j ][k+1] +
bp->array[var][i ][j+1][k ] +
bp->array[var][i+1][j ][k ])/7.0;
for (i = 1; i <= x_block_size; i++)
for (j = 1; j <= y_block_size; j++)
for (k = 1; k <= z_block_size; k++)
bp->array[var][i][j][k] = work[i][j][k];
}
}
} else {
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
bp = &blocks[n];
if (bp->number >= 0) {
for (i = 1; i <= x_block_size; i++)
for (j = 1; j <= y_block_size; j++)
for (k = 1; k <= z_block_size; k++) {
sb = bp->array[var][i-1][j-1][k-1] +
bp->array[var][i-1][j-1][k ] +
bp->array[var][i-1][j-1][k+1] +
bp->array[var][i-1][j ][k-1] +
bp->array[var][i-1][j ][k ] +
bp->array[var][i-1][j ][k+1] +
bp->array[var][i-1][j+1][k-1] +
bp->array[var][i-1][j+1][k ] +
bp->array[var][i-1][j+1][k+1];
sm = bp->array[var][i ][j-1][k-1] +
bp->array[var][i ][j-1][k ] +
bp->array[var][i ][j-1][k+1] +
bp->array[var][i ][j ][k-1] +
bp->array[var][i ][j ][k ] +
bp->array[var][i ][j ][k+1] +
bp->array[var][i ][j+1][k-1] +
bp->array[var][i ][j+1][k ] +
bp->array[var][i ][j+1][k+1];
sf = bp->array[var][i+1][j-1][k-1] +
bp->array[var][i+1][j-1][k ] +
bp->array[var][i+1][j-1][k+1] +
bp->array[var][i+1][j ][k-1] +
bp->array[var][i+1][j ][k ] +
bp->array[var][i+1][j ][k+1] +
bp->array[var][i+1][j+1][k-1] +
bp->array[var][i+1][j+1][k ] +
bp->array[var][i+1][j+1][k+1];
work[i][j][k] = (sb + sm + sf)/27.0;
}
for (i = 1; i <= x_block_size; i++)
for (j = 1; j <= y_block_size; j++)
for (k = 1; k <= z_block_size; k++)
bp->array[var][i][j][k] = work[i][j][k];
}
}
}
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/target.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include "block.h"
#include "proto.h"
#include "timer.h"
// This file contains routines that modify the number of blocks so that the
// number is close (+- 3) to the target number of blocks for the problem.
int reduce_blocks()
{
int l, i, j, p, c, num_comb, comb, num_parents, nm_t;
double t1, t2, t3;
parent *pp;
nm_t = 0;
t3 = 0.0;
t1 = timer();
zero_refine();
if (target_active)
num_comb = (global_active - num_pes*target_active + 3)/7;
else
num_comb = (global_active - num_pes*target_active)/7;
for (comb = 0, l = num_refine-1; comb < num_comb; l--) {
for (p = 0; p < max_active_parent; p++)
if ((pp = &parents[p])->number >= 0)
if (pp->level == l)
num_parents++;
for (p = 0; p < max_active_parent && comb < num_comb; p++)
if ((pp = &parents[p])->number >= 0)
if (pp->level == l) {
pp->refine = -1;
comb++;
for (c = 0; c < 8; c++)
if (pp->child_node[c] == my_pe && pp->child[c] >= 0)
blocks[pp->child[c]].refine = -1;
}
t2 = timer() - t2;
consolidate_blocks();
t3 += timer() - t2;
}
timer_target_rb += timer() - t1;
timer_target_dc += timer() - t1 - t3;
timer_target_cb += t3;
return(nm_t);
}
void add_blocks()
{
int l, i, j, n, in, num_split, split;
double t1, t2, t3;
block *bp;
t3 = 0.0;
t1 = timer();
if (target_active)
num_split = (num_pes*target_active + 3 - global_active)/7;
else
num_split = (num_pes*target_active - global_active)/7;
for (split = l = 0; split < num_split; l++) {
zero_refine();
for (j = num_refine; j >= 0; j--)
if (num_blocks[j]) {
cur_max_level = j;
break;
}
for (in = 0; split < num_split && in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
if ((bp = &blocks[n])->number >= 0)
if (bp->level == l) {
bp->refine = 1;
split++;
}
}
t2 = timer();
split_blocks();
t3 += timer() - t2;
}
timer_target_ab += timer() - t1;
timer_target_da += timer() - t1 - t3;
timer_target_sb += t3;
}
void zero_refine(void)
{
int n, c, in;
block *bp;
parent *pp;
for (in = 0; in < sorted_index[num_refine+1]; in++) {
n = sorted_list[in].n;
if ((bp= &blocks[n])->number >= 0) {
bp->refine = 0;
for (c = 0; c < 6; c++)
if (bp->nei_level[c] >= 0)
bp->nei_refine[c] = 0;
}
}
for (n = 0; n < max_active_parent; n++)
if ((pp = &parents[n])->number >= 0)
pp->refine = 0;
}

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/timer.h

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
double average[128];
double stddev[128];
double minimum[128];
double maximum[128];
double timer_all;
double timer_comm_all;
double timer_comm_dir[3];
double timer_comm_same[3];
double timer_comm_diff[3];
double timer_comm_bc[3];
double timer_calc_all;
double timer_cs_all;
double timer_cs_calc;
double timer_refine_all;
double timer_refine_co;
double timer_refine_mr;
double timer_refine_sb;
double timer_refine_cc;
double timer_cb_all;
double timer_target_all;
double timer_target_rb;
double timer_target_dc;
double timer_target_cb;
double timer_target_ab;
double timer_target_da;
double timer_target_sb;
double timer_plot;
long total_blocks;
int nb_min;
int nb_max;
int nrs;
int nps;
int num_refined;
int num_reformed;
int counter_bc[3];
int counter_same[3];
int counter_diff[3];
int counter_malloc;
double size_malloc;
int counter_malloc_init;
double size_malloc_init;
int total_red;

test-suite/trunk/MultiSource/Benchmarks/DOE-ProxyApps-C/miniAMR/util.c

// ************************************************************************
//
// miniAMR: stencil computations with boundary exchange and AMR.
//
// Copyright (2014) Sandia Corporation. Under the terms of Contract
// DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
// retains certain rights in this software.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// Questions? Contact Courtenay T. Vaughan (ctvaugh@sandia.gov)
// Richard F. Barrett (rfbarre@sandia.gov)
//
// ************************************************************************
#include <stdlib.h>
#include <stdio.h>
#include <time.h>
#include "block.h"
#include "proto.h"
#include "timer.h"
double timer(void)
{
return((((double) clock())/((double) CLOCKS_PER_SEC)));
}
void *ma_malloc(size_t size, char *file, int line)
{
void *ptr;
ptr = (void *) malloc(size);
if (ptr == NULL) {
printf("NULL pointer from malloc call in %s at %d\n", file, line);
exit(-1);
}
counter_malloc++;
size_malloc += (double) size;
return(ptr);
}