def test1d_3(self):
"""
1D, int
"""
dtyp = numpy.int64
# MPI stuff
comm = MPI.COMM_WORLD
rk = comm.Get_rank()
sz = comm.Get_size()
# create the dist array
n = 10
da = pnumpy.mdaZeros( (n,), dtyp ) # all data are valid
# expose the last element
da.expose( slce=(slice(-1, None, None),), winID='left' )
# set data
da[:] = 100*rk + numpy.array([i for i in range(n)], dtyp)
# access remote data
leftRk = (rk - 1) % sz
print('proc %d tries to access data and mask from %d' % (rk, leftRk))
leftData = da.getData(pe=leftRk, winID='left')
leftMask = da.getMask(pe=leftRk, winID='left')
print('leftData for rank %d = %s' % (rk, str(leftData)))
print('leftMask for rank %d = %s' % (rk, str(leftMask)))
# check
if leftRk < rk:
self.assertEqual(leftData[0], da[-1] - 100)
else:
self.assertEqual(leftData[0], da[-1] + 100*(sz-1))
# free
da.free()
def test1d_3(self):
"""
1D, int
"""
dtyp = numpy.int64
# MPI stuff
comm = MPI.COMM_WORLD
rk = comm.Get_rank()
sz = comm.Get_size()
# create the dist array
n = 10
da = pnumpy.mdaZeros((n, ), dtyp) # all data are valid
# expose the last element
da.expose(slce=(slice(-1, None, None), ), winID='left')
# set data
da[:] = 100 * rk + numpy.array([i for i in range(n)], dtyp)
# access remote data
leftRk = (rk - 1) % sz
print('proc %d tries to access data and mask from %d' % (rk, leftRk))
leftData = da.getData(pe=leftRk, winID='left')
leftMask = da.getMask(pe=leftRk, winID='left')
print('leftData for rank %d = %s' % (rk, str(leftData)))
print('leftMask for rank %d = %s' % (rk, str(leftMask)))
# check
if leftRk < rk:
self.assertEqual(leftData[0], da[-1] - 100)
else:
self.assertEqual(leftData[0], da[-1] + 100 * (sz - 1))
# free
da.free()
def test1d_1(self):
"""
1D, float64
"""
dtyp = numpy.float64
# MPI stuff
comm = MPI.COMM_WORLD
rk = comm.Get_rank()
sz = comm.Get_size()
# create the dist array
n = 10
mask = numpy.zeros( (n,), numpy.bool_ )
mask[-1] = 1 # last element is invalid
da = pnumpy.mdaZeros( (n,), dtyp, mask=mask )
# expose the last element
da.expose( slce=(slice(-1, None, None),), winID='left' )
# set data
da[:] = 100*rk + numpy.array([i for i in range(n)], dtyp)
# access remote data
leftRk = (rk - 1) % sz
print('proc {0} tries to access data and mask from {1}'.format(rk, leftRk))
leftData = da.getData(pe=leftRk, winID='left')
leftMask = da.getMask(pe=leftRk, winID='left')
print('leftData for rank {0} = {1}'.format(rk, str(leftData)))
print('leftMask for rank {0} = {1}'.format(rk, str(leftMask)))
# check
if leftRk < rk:
self.assertEqual(leftData[0], da[-1] - 100)
else:
self.assertEqual(leftData[0], da[-1] + 100*(sz-1))
# free
da.free()
def test0(self):
"""
Test constructors
"""
mask = numpy.zeros( (2, 3), numpy.bool_ )
mask[0, 0] = 0
da = pnumpy.mdaZeros( (2,3), numpy.float64, mask=mask )
da = pnumpy.mdaOnes( (2,3), numpy.float64, mask=mask )
da = pnumpy.mdaArray( [1,2,3], mask=numpy.array([0, 1, 1], numpy.bool_) )
def test0(self):
"""
Test constructors
"""
mask = numpy.zeros((2, 3), numpy.bool_)
mask[0, 0] = 0
da = pnumpy.mdaZeros((2, 3), numpy.float64, mask=mask)
da = pnumpy.mdaOnes((2, 3), numpy.float64, mask=mask)
da = pnumpy.mdaArray([1, 2, 3],
mask=numpy.array([0, 1, 1], numpy.bool_))
def test2d_1(self):
# create the dist array, the sizes are local to each processor
mask = numpy.zeros( (2,3), numpy.bool_ )
mask[1, 2] = 1
da = pnumpy.mdaZeros( (2,3), numpy.float32, mask=mask )
# processor rank and number of processes
rk = da.rk
nprocs = da.sz
# expose sub-domains
northSlab = ( slice(-1, None, None), slice(0, None, None) )
da.expose( slce=northSlab, winID='n' )
# set the data
da[:] = rk
# access the remote data, collective operation
northData = da.getData( (rk-1) % nprocs, winID='n' )
northMask = da.getMask( (rk-1) % nprocs, winID='n' )
# check sum
checkSum = da.reduce(lambda x,y:x+y, rootPe=0)
if checkSum is not None:
exactCheckSum = reduce(lambda x,y:x+y,
[rank for rank in range(nprocs)])
ntot = reduce(lambda x,y:x*y, da.shape)
exactCheckSum *= ntot
self.assertEqual(checkSum, exactCheckSum)
# check
self.assertEqual(northData.min(), (rk - 1) % nprocs)
self.assertEqual(northData.max(), (rk - 1) % nprocs)
# clean up
da.free()
def test2d_1(self):
# create the dist array, the sizes are local to each processor
mask = numpy.zeros((2, 3), numpy.bool_)
mask[1, 2] = 1
da = pnumpy.mdaZeros((2, 3), numpy.float32, mask=mask)
# processor rank and number of processes
rk = da.rk
nprocs = da.sz
# expose sub-domains
northSlab = (slice(-1, None, None), slice(0, None, None))
da.expose(slce=northSlab, winID='n')
# set the data
da[:] = rk
# access the remote data, collective operation
northData = da.getData((rk - 1) % nprocs, winID='n')
northMask = da.getMask((rk - 1) % nprocs, winID='n')
# check sum
checkSum = da.reduce(lambda x, y: x + y, rootPe=0)
if checkSum is not None:
exactCheckSum = reduce(lambda x, y: x + y,
[rank for rank in range(nprocs)])
ntot = reduce(lambda x, y: x * y, da.shape)
exactCheckSum *= ntot
self.assertEqual(checkSum, exactCheckSum)
# check
self.assertEqual(northData.min(), (rk - 1) % nprocs)
self.assertEqual(northData.max(), (rk - 1) % nprocs)
# clean up
da.free()
def test1d_1(self):
"""
1D, float64
"""
dtyp = numpy.float64
# MPI stuff
comm = MPI.COMM_WORLD
rk = comm.Get_rank()
sz = comm.Get_size()
# create the dist array
n = 10
mask = numpy.zeros((n, ), numpy.bool_)
mask[-1] = 1 # last element is invalid
da = pnumpy.mdaZeros((n, ), dtyp, mask=mask)
# expose the last element
da.expose(slce=(slice(-1, None, None), ), winID='left')
# set data
da[:] = 100 * rk + numpy.array([i for i in range(n)], dtyp)
# access remote data
leftRk = (rk - 1) % sz
print('proc {0} tries to access data and mask from {1}'.format(
rk, leftRk))
leftData = da.getData(pe=leftRk, winID='left')
leftMask = da.getMask(pe=leftRk, winID='left')
print('leftData for rank {0} = {1}'.format(rk, str(leftData)))
print('leftMask for rank {0} = {1}'.format(rk, str(leftMask)))
# check
if leftRk < rk:
self.assertEqual(leftData[0], da[-1] - 100)
else:
self.assertEqual(leftData[0], da[-1] + 100 * (sz - 1))
# free
da.free()