def getInstructions(app_name='testEfficiency_default',
functions=['gauss_solve','buildR_kernel','amplitude_capon'],
M=16,L=8):
if not ONLY_TIME:
def run(M=16,L=4,verbose=False):
# print "nvprof -u us --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L)
prof = Popen("nvprof -u ns --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L), shell=True, stdout=PIPE, stderr=PIPE)
output, error = prof.communicate()
if error == '':
if verbose:
print output
return output
else:
print output
raise Exception(error)
events = "inst_issued1_0,inst_issued2_0,inst_issued1_1,inst_issued2_1"#,\
# l2_read_requests,l2_write_requests,l2_read_texture_requests"
events_list = re.split(',',events)
profiler_output = run(M,L)
lines = re.split("\n+", profiler_output)
instructions = np.zeros((functions.__len__(),events_list.__len__()))
for i,l in enumerate(lines):
for j,function in enumerate(functions):
if re.search(".+%s.+"%function, l):
for k,event in enumerate(events_list):
timings = re.sub("^\s*", '', lines[i+k+1])
timings = re.sub("\s*%s.*"%event, '', timings)
columns = re.split("\s+", timings)
instructions[j,k] = float(columns[1])
else:
pass
##From CUPTI Users manual:
## inst_issued1_0 + (inst_issued2_0 * 2) + inst_issued1_1 + (inst_issued2_1 * 2)
# print instructions
instructions = instructions[:,0] + 2*instructions[:,1] + instructions[:,2] + 2*instructions[:,3]
# print "Recorded runtimes [ms]: "
# print runtimes[:,3]/1e3
return instructions
## events = "--events inst_issued1_0,inst_issued2_0,inst_issued1_1,inst_issued2_1"#,\
## l2_read_requests,l2_write_requests,l2_read_texture_requests"
else:
return np.zeros((3,))
def getMemoryOps(app_name='testEfficiency_default',
functions=['gauss_solve', 'buildR_kernel', 'amplitude_capon'],
M=16,
L=4):
if not ONLY_TIME:
def run(M=24, L=8, verbose=False):
# print "nvprof -u us --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L)
prof = Popen(
"nvprof -u ns --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"
% (events, app_name, M, L),
shell=True,
stdout=PIPE,
stderr=PIPE)
output, error = prof.communicate()
if error == '':
if verbose:
print output
return output
else:
print output
raise Exception(error)
events = "gld_request,gst_request,shared_load,shared_store"
events_list = re.split(',', events)
profiler_output = run(M, L)
lines = re.split("\n+", profiler_output)
memops = np.zeros((functions.__len__(), events_list.__len__()))
for i, l in enumerate(lines):
for j, function in enumerate(functions):
if re.search(".+%s.+" % function, l):
for k, event in enumerate(events_list):
timings = re.sub("^\s*", '', lines[i + k + 1])
timings = re.sub("\s*%s.*" % event, '', timings)
columns = re.split("\s+", timings)
memops[j, k] = float(columns[1])
else:
pass
# print "Recorded runtimes [ms]: "
# print runtimes[:,3]/1e3
return memops
else:
return np.zeros((3, ))
def setUp(self):
self.n = 5
self.R = np.zeros((self.n,self.n),dtype=complex)
self.a = np.zeros((self.n,),dtype=complex)
mu = 2
std = 3
for i in range(self.R.shape[1]):
self.a[i] = complex(np.random.normal(mu,std),np.random.normal(mu,std))
for j in range(self.R.shape[0]):
self.R[j,i] = complex(np.random.normal(mu,std),np.random.normal(mu,std))
def setUp(self):
self.n = 5
self.R = np.zeros((self.n, self.n), dtype=complex)
self.a = np.zeros((self.n, ), dtype=complex)
mu = 2
std = 3
for i in range(self.R.shape[1]):
self.a[i] = complex(np.random.normal(mu, std),
np.random.normal(mu, std))
for j in range(self.R.shape[0]):
self.R[j, i] = complex(np.random.normal(mu, std),
np.random.normal(mu, std))
def getTimings(app_name='testEfficiency_default',
functions=['gauss_solve','buildR_kernel','amplitude_capon'],
M=8,L=4):
if USE_CODE_COUNTERS:
prof = Popen("./%s testEfficiency testMVDRKernelPerformance %d %d"%(app_name,M,L), shell=True, stdout=PIPE, stderr=PIPE)
output, error = prof.communicate()
return np.zeros((3,))
def run(M=24,L=8,verbose=False):
# print "nvprof -u us -t 0.1 --devices 0 %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L
prof = Popen("nvprof -u ns %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L), shell=True, stdout=PIPE, stderr=PIPE)
output, error = prof.communicate()
if error == '':
if verbose:
print output
return output
else:
print output
raise Exception(error)
events=''
profiler_output = run(M,L)
lines = re.split("\n+", profiler_output)
runtimes = np.zeros([])
entered_once = False
for l in lines:
for i,function in enumerate(functions):
if re.search(".+%s.+"%function, l):
timings = re.sub("^\s*", '', l)
timings = re.sub("[\sa-zA-Z]*%s.+"%function, '', timings)
columns = re.split("\s+", timings)
if not entered_once:
entered_once = True
runtimes = np.zeros((functions.__len__(),columns.__len__()))
runtimes[i] = map(float,columns)
break
else:
pass
return runtimes[:,3]/1e3
def getMemoryOps(app_name='testEfficiency_default',
functions=['gauss_solve','buildR_kernel','amplitude_capon'],
M=16,L=4):
if not ONLY_TIME:
def run(M=24,L=8,verbose=False):
# print "nvprof -u us --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L)
prof = Popen("nvprof -u ns --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L), shell=True, stdout=PIPE, stderr=PIPE)
output, error = prof.communicate()
if error == '':
if verbose:
print output
return output
else:
print output
raise Exception(error)
events = "gld_request,gst_request,shared_load,shared_store"
events_list = re.split(',',events)
profiler_output = run(M,L)
lines = re.split("\n+", profiler_output)
memops = np.zeros((functions.__len__(),events_list.__len__()))
for i,l in enumerate(lines):
for j,function in enumerate(functions):
if re.search(".+%s.+"%function, l):
for k,event in enumerate(events_list):
timings = re.sub("^\s*", '', lines[i+k+1])
timings = re.sub("\s*%s.*"%event, '', timings)
columns = re.split("\s+", timings)
memops[j,k] = float(columns[1])
else:
pass
# print "Recorded runtimes [ms]: "
# print runtimes[:,3]/1e3
return memops
else:
return np.zeros((3,))
def uhdu(A, n):
'''
Calculates the UDUH decomposition of the Hermitian matrix A
such that U is unit upper triangular, D is diagonal and UDU'=A
(' = H = symmetric conjugated)
Now we avoid using the complex sqrt by instead introducing two complex add
Returns [U D]
'''
U = eye(n, dtype=A.dtype)
D = zeros(n, dtype=A.dtype)
for i in range(n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i] * U[k, i].conjugate() * D[k]
D[i] = A[i, i] - upperColSum
for j in range(i + 1, n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i].conjugate() * U[k, j] * D[k]
U[i, j] = (A[i, j] - upperColSum) / D[i]
return [U, D]
def uhdu(A, n):
"""
Calculates the UDUH decomposition of the Hermitian matrix A
such that U is unit upper triangular, D is diagonal and UDU'=A
(' = H = symmetric conjugated)
Now we avoid using the complex sqrt by instead introducing two complex add
Returns [U D]
"""
U = eye(n, dtype=A.dtype)
D = zeros(n, dtype=A.dtype)
for i in range(n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i] * U[k, i].conjugate() * D[k]
D[i] = A[i, i] - upperColSum
for j in range(i + 1, n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i].conjugate() * U[k, j] * D[k]
U[i, j] = (A[i, j] - upperColSum) / D[i]
return [U, D]
def diagonalSolve(A, b, n):
''' Solve the diagonal system Ax = b, A is sparse, hence a vector '''
x = np.zeros(n, A.dtype)
for i in range(n):
x[i] = b[i] / A[i]
return x
def diagonalSolve(A, b, n):
''' Solve the diagonal system Ax = b, A is sparse, hence a vector '''
x = np.zeros(n, A.dtype)
for i in range(n):
x[i] = b[i] / A[i]
return x
def testSolveBiCG(self):
x = ls.solveBiCG(self.A, self.b1, self.x0_zero, 0, 0)
self.assertMatrixAlmosteEqual(self.x1, x, 14)
x = ls.solveBiCG(self.complexA, self.complexb, self.x0_zero, 0, 0)
self.assertMatrixAlmosteEqual(self.complexx, x, 13)
x_ref = np.linalg.solve(self.randA, self.randb)
x = ls.solveBiCG(self.randA, self.randb, np.zeros(self.L, dtype=complex), 0, 0)
self.assertMatrixAlmosteEqual(x_ref, x, 15)
def uhduGPUProto(A, n):
"""
Calculates the UDUH decomposition of the Hermitian matrix A
such that U is unit upper triangular, D is diagonal and UDU'=A
(' = H = symmetric conjugated)
Now we avoid using the complex sqrt by instead introducing two complex add
Returns [U D]
prototype CPU-code for how uhdu-composition should be done on the GPU
"""
U = eye([n, n], A.dtype)
D = zeros([n, 1], A.dtype)
upperColSum = zeros([n, 1], A.dtype) # shared column sum buffer
Ai = zeros([n, 1], A.dtype) # shared A row buffer
for i in range(n):
# read one row into "shared" memory
for k in range(n):
Ai[k, 0] = A[i, k]
upperColSum = 0
for k in range(i):
upperColSum += U[k, i] * U[k, i].conjugate() * D[k, 0]
D[i, 0] = A[i, i] - upperColSum
for j in range(i + 1, n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i].conjugate() * U[k, j] * D[k, 0]
U[i, j] = (A[i, j] - upperColSum) / D[i, 0]
return [U, D]
def uhduGPUProto(A, n):
'''
Calculates the UDUH decomposition of the Hermitian matrix A
such that U is unit upper triangular, D is diagonal and UDU'=A
(' = H = symmetric conjugated)
Now we avoid using the complex sqrt by instead introducing two complex add
Returns [U D]
prototype CPU-code for how uhdu-composition should be done on the GPU
'''
U = eye([n, n], A.dtype)
D = zeros([n, 1], A.dtype)
upperColSum = zeros([n, 1], A.dtype) # shared column sum buffer
Ai = zeros([n, 1], A.dtype) # shared A row buffer
for i in range(n):
# read one row into "shared" memory
for k in range(n):
Ai[k, 0] = A[i, k]
upperColSum = 0
for k in range(i):
upperColSum += U[k, i] * U[k, i].conjugate() * D[k, 0]
D[i, 0] = A[i, i] - upperColSum
for j in range(i + 1, n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i].conjugate() * U[k, j] * D[k, 0]
U[i, j] = (A[i, j] - upperColSum) / D[i, 0]
return [U, D]
def testSolveBiCG(self):
n = len(self.b1)
x0 = mynp.zeros(n, dtype=complex)
x = ls.solveBiCG(self.A, self.b1, x0, 0, n)
self.assertMatrixAlmosteEqual(self.x1, x, 14)
n = len(self.complexb)
x = ls.solveBiCG(self.complexA, self.complexb, x0, 0, n)
self.assertMatrixAlmosteEqual(self.complexx, x, 12)
def forwardSolve(A, b, n):
''' Forward solve the lower triangular system Ax = b '''
x = np.zeros(n, A.dtype)
for i in range(n):
x[i] = b[i]
for k in range(i):
x[i] -= A[i, k] * x[k]
x[i] /= A[i, i]
return x
def backtrackSolve(A, b, n):
''' Backtrack solve the upper triangular system Ax = b '''
x = np.zeros(n, A.dtype)
for i in reversed(range(n)):
x[i] = b[i]
for k in reversed(range(i+1, n)):
x[i] -= A[i, k] * x[k]
x[i] /= A[i, i]
return x
def backtrackSolve(A, b, n):
''' Backtrack solve the upper triangular system Ax = b '''
x = np.zeros(n, A.dtype)
for i in reversed(range(n)):
x[i] = b[i]
for k in reversed(range(i + 1, n)):
x[i] -= A[i, k] * x[k]
x[i] /= A[i, i]
return x
def forwardSolve(A, b, n):
''' Forward solve the lower triangular system Ax = b '''
x = np.zeros(n, A.dtype)
for i in range(n):
x[i] = b[i]
for k in range(i):
x[i] -= A[i, k] * x[k]
x[i] /= A[i, i]
return x
def butlerMatrix(m, n, ix = 0):
''' Returns the mxn buttler matrix used for beamspace processing '''
''' The matrix is equal to the normalized n-point DFT-matrix '''
' An optional list argument ix can be specified to select different beams than the first m'
B = mynp.zeros([m, n], dtype=complex)
r = range(m)
if ix != 0:
r = ix
for i in r:
for j in range(n):
B[i, j] = 1/ma.sqrt(n) * cm.exp(-1j*2*cm.pi*i*j/n)
return B
def butlerMatrix(m, n, ix=0):
''' Returns the mxn buttler matrix used for beamspace processing '''
''' The matrix is equal to the normalized n-point DFT-matrix '''
' An optional list argument ix can be specified to select different beams than the first m'
B = mynp.zeros([m, n], dtype=complex)
r = range(m)
if ix != 0:
r = ix
for i in r:
for j in range(n):
B[i, j] = 1 / ma.sqrt(n) * cm.exp(-1j * 2 * cm.pi * i * j / n)
return B
def upinterpolate(img):
#
# img Ny x Nx
Ny,Nx = img.shape
Kx = 2 # Must be even!!!
new_img = np.zeros((Ny,Kx*Nx-Kx),dtype=img.dtype)
for i in range(Kx):
new_img[:,i::Kx] = img[:,0:-1] + (i+0.5)*(img[:,1:] - img[:,0:-1])/Kx
#
# new_img[:,0::Kx] = img[:,0:-1] + a*(img[:,1:] - img[:,0:-1])
# new_img[:,1::Kx] = img[:,1:] + 0.75*(img[:,0:-1] - img[:,1:] )
return new_img
def upinterpolate(img):
#
# img Ny x Nx
Ny, Nx = img.shape
Kx = 2 # Must be even!!!
new_img = np.zeros((Ny, Kx * Nx - Kx), dtype=img.dtype)
for i in range(Kx):
new_img[:, i::Kx] = img[:, 0:-1] + (i + 0.5) * (img[:, 1:] -
img[:, 0:-1]) / Kx
#
# new_img[:,0::Kx] = img[:,0:-1] + a*(img[:,1:] - img[:,0:-1])
# new_img[:,1::Kx] = img[:,1:] + 0.75*(img[:,0:-1] - img[:,1:] )
return new_img
def cholesky(A, n):
''' Calculates the Cholesky decomposition of the Hermitian matrix A such that U is upper triangular and U'*U = A '''
U = zeros([n, n], A.dtype)
for i in range(n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i] * (U[k, i]).conjugate()
#U[i, i] = cmath.sqrt(A[i, i] - upperColSum)
U[i, i] = (A[i, i] - upperColSum)**0.5
for j in range(i + 1, n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i].conjugate() * U[k, j]
U[i, j] = (A[i, j] - upperColSum) / U[i, i]
return U
def cholesky(A, n):
""" Calculates the Cholesky decomposition of the Hermitian matrix A such that U is upper triangular and U'*U = A """
U = zeros([n, n], A.dtype)
for i in range(n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i] * (U[k, i]).conjugate()
# U[i, i] = cmath.sqrt(A[i, i] - upperColSum)
U[i, i] = (A[i, i] - upperColSum) ** 0.5
for j in range(i + 1, n):
upperColSum = 0
for k in range(i):
upperColSum += U[k, i].conjugate() * U[k, j]
U[i, j] = (A[i, j] - upperColSum) / U[i, i]
return U
##
##app_name = "testEfficiency_mathcheck"
##math_run = run()
##events = "--events gld_request,gst_request,shared_load,shared_store"
##math_event_run = run()
#
#
#return info
##export PYTHONPATH="/home/me/Work/UiO/Phd/Code/Profile"
##python testEfficiency.py testMVDRKernelPerformance
#
##nvprof ./testEfficiency testEfficiency testMVDRKernelPerformance
M = 8
L_list = np.arange(M - 1) + 2
time = np.zeros((L_list.shape[0], 6))
for l, L in enumerate(L_list):
time[l] = collectResults(M, L)
np.savetxt('time-M%d.txt' % M, time)
M = 16
L_list = np.arange(M - 1) + 2
time = np.zeros((L_list.shape[0], 6))
for l, L in enumerate(L_list):
time[l] = collectResults(M, L)
np.savetxt('time-M%d.txt' % M, time)
M = 32
L_list = np.arange(M - 1) + 2
time = np.zeros((L_list.shape[0], 6))
for l, L in enumerate(L_list):
def setUp(self):
self.n = 3
self.A = np.array([[2.0, -1.0, 0.0], [-1.0, 2.0, -1.0], [0.0, -1.0, 2.0]],dtype=complex)
self.A2 = np.array([[4.0, -2.0, 0.0], [-2.0, 4.0, -2.0], [0.0, -2.0, 4.0]],dtype=complex)
self.R = np.array([[1.414213562373095, -0.707106781186547, 0.0],
[0.0, 1.224744871391589, -0.816496580927726],
[0.0, 0.0, 1.154700538379252]],dtype=complex)
self.AA = np.array([[5, -4, 1], [-4, 6, -4], [1, -4, 5]],dtype=complex)
self.B = np.array([[1, 2, 3], [1, 2, 3], [1, 2, 3]],dtype=complex)
self.BT = np.array([[1, 1, 1], [2, 2, 2], [3, 3, 3]],dtype=complex)
self.b1 = np.array([1.0, 1.0, 1.0],dtype=complex)
self.x1 = np.array([3.0/2, 2.0, 3.0/2],dtype=complex)
self.b2 = np.array([1, 2, 3],dtype=complex)
self.x2 = np.array([5.0/2, 4, 7.0/2],dtype=complex)
self.C = np.array([[1, 2, 3], [0, 1, 1], [0, 0, 1]],dtype=complex)
self.b1c = np.array([1, 1, 1],dtype=complex)
self.x1c = np.array([-2, 0, 1],dtype=complex)
self.x1cT = np.array([1, -1, -1],dtype=complex)
self.x0_zero = np.zeros((3,), dtype=complex)
self.Ab2 = np.array([0, 0, 4],dtype=complex)
self.Ab1 = np.array([1, 0, 1],dtype=complex)
self.Ab1b1T = np.array([[3, 1, 3], [1, 6, 5], [3, 5, 11]],dtype=complex)
self.invA = np.array([[0.750, 0.50, 0.250], [0.50, 1.0, 0.50], [0.250, 0.50, 0.750]],dtype=complex)
self.invAb1b1T = np.array([[0.465909090909091, 0.045454545454545, -0.147727272727273],
[0.045454545454545, 0.272727272727273, -0.136363636363636],
[-0.147727272727273, -0.136363636363636, 0.193181818181818]],dtype=complex)
self.complexA = np.array([[2.0, 3.0 + 1.0j, 2.0 - 2.0j],
[3.0 - 1.0j, 9.0, -2.0j],
[2.0 + 2.0j, 2.0j, 14.0]], dtype=complex)
self.complexR = np.array([[1.414213562373095, 2.121320343559642 + 0.707106781186547j, 1.414213562373095 - 1.414213562373095j],
[0.0, 2.0, -1.0 + 1.0j],
[0.0, 0.0, 2.828427124746190]], dtype=complex)
self.complexb = np.array([1.0, 1.0 + 1.0j, 1.0 - 2.0j], dtype=complex)
self.complexy = np.array([0.707106781186547 - 0.0j,
-0.250 + 0.750j,
-0.353553390593273 - 0.883883476483184j], dtype=complex)
self.complexx = np.array([1.593749999999999 - 0.06250j,
-0.343750 + 0.281250j,
-0.1250 - 0.31250j], dtype=complex)
self.complexAb = np.array([2.0 - 2.0j, 8.0 + 6.0j, 14.0 - 24.0j], dtype=complex)
self.but4 = np.array([[0.5]*4, [0.5, -0.5j, -0.5, 0.5j], [0.5, -0.5]*2, [0.5, 0.5j, -0.5, -0.5j]], dtype=complex)
self.but3 = np.array([[0.577350269189626, 0.577350269189626, 0.577350269189626],
[0.577350269189626, -0.288675134594813 - 0.50j, -0.288675134594813 + 0.50j],
[0.577350269189626, -0.288675134594813 + 0.50j, -0.288675134594813 - 0.50j]], dtype=complex)
self.bsComplexb = np.array([1.732050807568878 - 0.577350269189626j, 1.50 + 0.288675134594813j], dtype=complex)
self.diag = 0.2
self.x = np.array([1.0, 1.0 + 1.0j, 1.0 - 2.0j, 2.0 + 1.0j], dtype=complex)
self.complexAbbH = np.array([[10.0, 11.0 + 1.0j, 10.0 - 2.0j],
[11.0 - 1.0j, 17.0, 8.0 - 2.0j],
[10.0 + 2.0j, 8.0 + 2.0j, 22.0]], dtype=complex)
self.complexInvAbbH = np.array([[1.067010309278351, -0.407216494845361 - 0.015463917525773j, -0.190721649484536 + 0.020618556701031j],
[-0.407216494845361 + 0.015463917525773j, 0.247422680412371, 0.077319587628866 - 0.015463917525773j],
[-0.190721649484536 - 0.020618556701031j, 0.077319587628866 + 0.015463917525773j, 0.087628865979381]], dtype=complex)
self.complexInvA = np.array([[1.906250, -0.593750 - 0.156250j, -0.250 + 0.18750j],
[-0.593750 + 0.156250j, 0.31250, 0.06250 - 0.06250j],
[-0.250 - 0.18750j, 0.06250 + 0.06250j, 0.1250]], dtype=complex)
self.complexA4x4 = np.array([[22.0, 8.0, 11.0 - 11.0j, 22.0 - 7.0j],
[8.0, 22.0, 17.0 - 2.0j, 11.0 - 7.0j],
[11.0 + 11.0j, 17.0 + 2.0j, 45.0, 23.0 - 5.0j],
[22.0 + 7.0j, 11.0 + 7.0j, 23.0 + 5.0j, 37.0]], dtype=complex)
self.U4x4 = np.array([[1.0000, 0.3636, 0.50 - 0.50j, 1.0 - 0.3182j],
[0.0, 1.0, 0.6810 + 0.1048j, 0.1571 - 0.2333j],
[0.0, 0.0, 1.0, 0.2776 - 0.3670j],
[0.0, 0.0, 0.0, 1.0]], dtype=complex)
self.D4x4 = np.array([22.0, 19.0909, 24.9381, 5.9806])
self.sonardata_R = np.array(np.load('./data/data_R.npy')) # created without diagonal loading
self.sonardata_a = np.array(np.load('./data/data_a.npy'))
self.sonardata_Ria = np.array(np.load('./data/data_Ria.npy'))
self.sonardata_ar = np.array(np.load('./data/data_ar.npy'))
self.sonardata_n = 32
# random data for testing
self.L = L = 24
self.d = d = 100
U = np.triu(np.random.randn(L,L) + np.random.randn(L,L)*1j) + np.eye(L)*d
self.randA = np.dot(U.conjugate().T, U)
self.randb = np.random.randn(L) + np.random.randn(L)*1j
##
##app_name = "testEfficiency_mathcheck"
##math_run = run()
##events = "--events gld_request,gst_request,shared_load,shared_store"
##math_event_run = run()
#
#
#return info
##export PYTHONPATH="/home/me/Work/UiO/Phd/Code/Profile"
##python testEfficiency.py testMVDRKernelPerformance
#
##nvprof ./testEfficiency testEfficiency testMVDRKernelPerformance
M = 8
L_list = np.arange(M-1)+2
time = np.zeros((L_list.shape[0],6))
for l,L in enumerate(L_list):
time[l] = collectResults(M,L)
np.savetxt('time-M%d.txt'%M,time)
M = 16
L_list = np.arange(M-1)+2
time = np.zeros((L_list.shape[0],6))
for l,L in enumerate(L_list):
time[l] = collectResults(M,L)
np.savetxt('time-M%d.txt'%M,time)
M = 32
L_list = np.arange(M-1)+2
time = np.zeros((L_list.shape[0],6))
for l,L in enumerate(L_list):
def getTimings(app_name='testEfficiency_default',
functions=['gauss_solve', 'buildR_kernel', 'amplitude_capon'],
M=8,
L=4):
if USE_CODE_COUNTERS:
prof = Popen("./%s testEfficiency testMVDRKernelPerformance %d %d" %
(app_name, M, L),
shell=True,
stdout=PIPE,
stderr=PIPE)
output, error = prof.communicate()
return np.zeros((3, ))
def run(M=24, L=8, verbose=False):
# print "nvprof -u us -t 0.1 --devices 0 %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L
prof = Popen(
"nvprof -u ns %s ./%s testEfficiency testMVDRKernelPerformance %d %d"
% (events, app_name, M, L),
shell=True,
stdout=PIPE,
stderr=PIPE)
output, error = prof.communicate()
if error == '':
if verbose:
print output
return output
else:
print output
raise Exception(error)
events = ''
profiler_output = run(M, L)
lines = re.split("\n+", profiler_output)
runtimes = np.zeros([])
entered_once = False
for l in lines:
for i, function in enumerate(functions):
if re.search(".+%s.+" % function, l):
timings = re.sub("^\s*", '', l)
timings = re.sub("[\sa-zA-Z]*%s.+" % function, '', timings)
columns = re.split("\s+", timings)
if not entered_once:
entered_once = True
runtimes = np.zeros(
(functions.__len__(), columns.__len__()))
runtimes[i] = map(float, columns)
break
else:
pass
return runtimes[:, 3] / 1e3
def getInstructions(
app_name='testEfficiency_default',
functions=['gauss_solve', 'buildR_kernel', 'amplitude_capon'],
M=16,
L=8):
if not ONLY_TIME:
def run(M=16, L=4, verbose=False):
# print "nvprof -u us --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"%(events,app_name,M,L)
prof = Popen(
"nvprof -u ns --devices 0 --events %s ./%s testEfficiency testMVDRKernelPerformance %d %d"
% (events, app_name, M, L),
shell=True,
stdout=PIPE,
stderr=PIPE)
output, error = prof.communicate()
if error == '':
if verbose:
print output
return output
else:
print output
raise Exception(error)
events = "inst_issued1_0,inst_issued2_0,inst_issued1_1,inst_issued2_1" #,\
# l2_read_requests,l2_write_requests,l2_read_texture_requests"
events_list = re.split(',', events)
profiler_output = run(M, L)
lines = re.split("\n+", profiler_output)
instructions = np.zeros((functions.__len__(), events_list.__len__()))
for i, l in enumerate(lines):
for j, function in enumerate(functions):
if re.search(".+%s.+" % function, l):
for k, event in enumerate(events_list):
timings = re.sub("^\s*", '', lines[i + k + 1])
timings = re.sub("\s*%s.*" % event, '', timings)
columns = re.split("\s+", timings)
instructions[j, k] = float(columns[1])
else:
pass
##From CUPTI Users manual:
## inst_issued1_0 + (inst_issued2_0 * 2) + inst_issued1_1 + (inst_issued2_1 * 2)
# print instructions
instructions = instructions[:,
0] + 2 * instructions[:,
1] + instructions[:,
2] + 2 * instructions[:,
3]
# print "Recorded runtimes [ms]: "
# print runtimes[:,3]/1e3
return instructions
## events = "--events inst_issued1_0,inst_issued2_0,inst_issued1_1,inst_issued2_1"#,\
## l2_read_requests,l2_write_requests,l2_read_texture_requests"
else:
return np.zeros((3, ))