Contents of /MITgcm_contrib/cg2d_bench/README

#	$Id: README,v 1.2 2006/05/12 22:32:02 ce107 Exp $	
Benchmarking routine of the CG2D solver in MITgcm (barotropic solve)

To build:

a) Parameterizations:
i) of SIZE.h:
sNx = size of tile in x-direction (ideally fits in cache, 30-60)
sNy = size of tile in y-direction (ideally fits in cache, 30-60)
OLx = overlap size in x-direction (1 or 3 usually)
OLy = overlap size in y-direction (1 or 3 usually)
ii) of ini_parms.F:
nTimeSteps = number of pseudo-timesteps to run for
cg2dMaxIters = maximum number of CG iterations per timestep

b) Compilation
$CC $CFLAGS -c tim.c
$FC $DEFINES $INCLUDES $FCFLAGS -o cg2d *.F tim.o $LIBS -lm

$DEFINES:
1) For single precision add
-DUSE_SINGLE_PRECISION
2) For mixed (single for most ops, double for reductions) precision add
-DUSE_MIXED_PRECISION to -DUSE_SINGLE_PRECISION
3) Parallel (MPI) operation
-DALLOW_MPI -DUSE_MPI_INIT -DUSE_MPI_GSUM -DUSE_MPI_EXCH
4) Use MPI timing routines
-DUSE_MPI_TIME
5) Use of MPI_Sendrecv() instead of MPI_Isend/MPI_Irecv()/MPI_Waitall()
-DUSE_SNDRCV
6) Use of JAM for exchanges (not available without the hardware)
-DUSE_JAM_EXCH
7) Use of JAM for the global sum (not available without the hardware)
-DUSE_JAM_GSUM
8) In order to avoid doing the global sum in MPI do not define
-DUSE_MPI_GSUM
and all processors will see their own residual instead (dangerous)
9) In order to avoid doing the exchanges in MPI do not define
-DUSE_MPI_EXCH
and all processors avoid exchanging shadow regions (dangerous)
10) Performance counters
-DUSE_PAPI_FLOPS	To use PAPI to produce Mflop/s
or
-DUSE_PAPI_FLIPS	To use PAPI to produce Mflip/s
To produce this information for every iteration instead of each "timestep" 
add a -DPAPI_PER_ITERATION to the above
11) Extra (nearest neighbor) exchange steps to stress comms
-DTEN_EXTRA_EXCHS
12) Extra (global) sum steps to stress comms
-DHUNDRED_EXTRA_SUMS
13) 2D (PxQ) vs 1D decomposition
-DDECOMP2D
14) To output the residual every iteration:
-DRESIDUAL_PER_ITERATION

$INCLUDES (if using PAPI)
-I$PAPI_ROOT/include

$LIBS (if using PAPI - depending on the platform extra libs may be needed)
-L$PAPI_ROOT/lib -lpapi

c) Running

1) Allowing the system to choose the PxQ decomposition if setup for it:
mpiexec -n $NPROCS ./cg2d
2) Create a decomp.touse with the P & Q dimensions declared in the first
two lines as two integers, eg.

cat > decomp.touse << EOF
10
20
EOF

mpiexec -n 200 ./cg2d
1	# $Id: README,v 1.2 2006/05/12 22:32:02 ce107 Exp $
2	Benchmarking routine of the CG2D solver in MITgcm (barotropic solve)
3
4	To build:
5
6	a) Parameterizations:
7	i) of SIZE.h:
8	sNx = size of tile in x-direction (ideally fits in cache, 30-60)
9	sNy = size of tile in y-direction (ideally fits in cache, 30-60)
10	OLx = overlap size in x-direction (1 or 3 usually)
11	OLy = overlap size in y-direction (1 or 3 usually)
12	ii) of ini_parms.F:
13	nTimeSteps = number of pseudo-timesteps to run for
14	cg2dMaxIters = maximum number of CG iterations per timestep
15
16	b) Compilation
17	$CC $CFLAGS -c tim.c
18	$FC $DEFINES $INCLUDES $FCFLAGS -o cg2d *.F tim.o $LIBS -lm
19
20	$DEFINES:
21	1) For single precision add
22	-DUSE_SINGLE_PRECISION
23	2) For mixed (single for most ops, double for reductions) precision add
24	-DUSE_MIXED_PRECISION to -DUSE_SINGLE_PRECISION
25	3) Parallel (MPI) operation
26	-DALLOW_MPI -DUSE_MPI_INIT -DUSE_MPI_GSUM -DUSE_MPI_EXCH
27	4) Use MPI timing routines
28	-DUSE_MPI_TIME
29	5) Use of MPI_Sendrecv() instead of MPI_Isend/MPI_Irecv()/MPI_Waitall()
30	-DUSE_SNDRCV
31	6) Use of JAM for exchanges (not available without the hardware)
32	-DUSE_JAM_EXCH
33	7) Use of JAM for the global sum (not available without the hardware)
34	-DUSE_JAM_GSUM
35	8) In order to avoid doing the global sum in MPI do not define
36	-DUSE_MPI_GSUM
37	and all processors will see their own residual instead (dangerous)
38	9) In order to avoid doing the exchanges in MPI do not define
39	-DUSE_MPI_EXCH
40	and all processors avoid exchanging shadow regions (dangerous)
41	10) Performance counters
42	-DUSE_PAPI_FLOPS To use PAPI to produce Mflop/s
43	or
44	-DUSE_PAPI_FLIPS To use PAPI to produce Mflip/s
45	To produce this information for every iteration instead of each "timestep"
46	add a -DPAPI_PER_ITERATION to the above
47	11) Extra (nearest neighbor) exchange steps to stress comms
48	-DTEN_EXTRA_EXCHS
49	12) Extra (global) sum steps to stress comms
50	-DHUNDRED_EXTRA_SUMS
51	13) 2D (PxQ) vs 1D decomposition
52	-DDECOMP2D
53	14) To output the residual every iteration:
54	-DRESIDUAL_PER_ITERATION
55
56	$INCLUDES (if using PAPI)
57	-I$PAPI_ROOT/include
58
59	$LIBS (if using PAPI - depending on the platform extra libs may be needed)
60	-L$PAPI_ROOT/lib -lpapi
61
62	c) Running
63
64	1) Allowing the system to choose the PxQ decomposition if setup for it:
65	mpiexec -n $NPROCS ./cg2d
66	2) Create a decomp.touse with the P & Q dimensions declared in the first
67	two lines as two integers, eg.
68
69	cat > decomp.touse << EOF
70	10
71	20
72	EOF
73
74	mpiexec -n 200 ./cg2d