Skip to content

P
pytensor
提交 db1ce04e authored 作者: notoraptor's avatar notoraptor
浏览文件
操作

Fix accumulator initialization for GpuCaReduceCuda.

上级 0ae39e6a
隐藏空白字符变更
内嵌 并排
正在显示 包含 39 行增加 和 23 行删除
theano/gpuarray/elemwise.py
浏览文件 @ db1ce04e
...@@ -492,6 +492,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -492,6 +492,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
__props__ = ('axis', 'reduce_mask', 'dtype', 'acc_dtype', 'scalar_op', __props__ = ('axis', 'reduce_mask', 'dtype', 'acc_dtype', 'scalar_op',
'pre_scalar_op') 'pre_scalar_op')
_f16_ok = True _f16_ok = True
verbose = 0
def __init__(self, scalar_op, axis=None, def __init__(self, scalar_op, axis=None,
reduce_mask=None, dtype=None, acc_dtype=None, reduce_mask=None, dtype=None, acc_dtype=None,
...@@ -1092,6 +1093,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1092,6 +1093,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals() """ % locals()
def c_code_reduce_ccontig(self, sio, node, name, x, z, fail): def c_code_reduce_ccontig(self, sio, node, name, x, z, fail):
verbose = self.verbose
in_dtype = "npy_" + node.inputs[0].dtype in_dtype = "npy_" + node.inputs[0].dtype
out_dtype = "npy_" + node.outputs[0].dtype out_dtype = "npy_" + node.outputs[0].dtype
if getattr(self.scalar_op, 'identity', None) == 0: if getattr(self.scalar_op, 'identity', None) == 0:
...@@ -1120,7 +1122,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1120,7 +1122,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
if(PyGpuArray_SIZE(%(x)s)==0){ if(PyGpuArray_SIZE(%(x)s)==0){
%(zero_shp)s; %(zero_shp)s;
}else{ }else{
int verbose = 0; int verbose = %(verbose)s;
size_t numEls = PyGpuArray_SIZE(%(x)s); size_t numEls = PyGpuArray_SIZE(%(x)s);
size_t n_threads = std::min(numEls, (size_t) 256); size_t n_threads = std::min(numEls, (size_t) 256);
size_t n_blocks = 1; size_t n_blocks = 1;
...@@ -1141,10 +1143,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1141,10 +1143,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_1(self, sio, node, name, x, z, fail): def c_code_reduce_1(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 256), 1, 1};
size_t n_blocks[3] = {1, 1, 1}; size_t n_blocks[3] = {1, 1, 1};
%(makecall)s %(makecall)s
...@@ -1152,10 +1155,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1152,10 +1155,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_11(self, sio, node, name, x, z, fail): def c_code_reduce_11(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[1], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[1], (size_t) 256), 1, 1};
while (n_threads[1] * n_threads[0] <= 256) ++n_threads[1]; while (n_threads[1] * n_threads[0] <= 256) ++n_threads[1];
...@@ -1180,6 +1184,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1180,6 +1184,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" """
assert N in [1, 2, 3] assert N in [1, 2, 3]
verbose = self.verbose
in_dtype = "npy_" + node.inputs[0].dtype in_dtype = "npy_" + node.inputs[0].dtype
out_dtype = "npy_" + node.outputs[0].dtype out_dtype = "npy_" + node.outputs[0].dtype
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
...@@ -1221,7 +1226,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1221,7 +1226,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[%(N)s], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[%(N)s], (size_t) 256), 1, 1};
%(threads_y)s %(threads_y)s
%(threads_z)s %(threads_z)s
...@@ -1240,6 +1245,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1240,6 +1245,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
self.c_code_reduce_01X(sio, node, name, x, z, fail, 3) self.c_code_reduce_01X(sio, node, name, x, z, fail, 3)
def c_code_reduce_10(self, sio, node, name, x, z, fail): def c_code_reduce_10(self, sio, node, name, x, z, fail):
verbose = self.verbose
in_dtype = "npy_" + node.inputs[0].dtype in_dtype = "npy_" + node.inputs[0].dtype
out_dtype = "npy_" + node.outputs[0].dtype out_dtype = "npy_" + node.outputs[0].dtype
acc_dtype = "npy_" + self._acc_dtype(node.inputs[0].dtype) acc_dtype = "npy_" + self._acc_dtype(node.inputs[0].dtype)
...@@ -1255,7 +1261,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1255,7 +1261,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
if(PyGpuArray_STRIDES(%(x)s)[0]> if(PyGpuArray_STRIDES(%(x)s)[0]>
PyGpuArray_STRIDES(%(x)s)[1]){ PyGpuArray_STRIDES(%(x)s)[1]){
// If there are a lot of summations to do, then we can use simple parallelization - // If there are a lot of summations to do, then we can use simple parallelization -
...@@ -1333,6 +1339,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1333,6 +1339,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_010(self, sio, node, name, x, z, fail): def c_code_reduce_010(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
makecall_inner = self._makecall(node, name, x, z, fail, makecall_inner = self._makecall(node, name, x, z, fail,
pattern="010_inner") pattern="010_inner")
...@@ -1399,7 +1406,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1399,7 +1406,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
} }
else else
{ {
int verbose = 2; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min((size_t) 32, PyGpuArray_DIMS(%(x)s)[2]), 1, 1}; size_t n_threads[3] = {std::min((size_t) 32, PyGpuArray_DIMS(%(x)s)[2]), 1, 1};
while( (n_threads[0]*(n_threads[1]+1)<=256) while( (n_threads[0]*(n_threads[1]+1)<=256)
...@@ -1441,10 +1448,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1441,10 +1448,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_0101(self, sio, node, name, x, z, fail): def c_code_reduce_0101(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[3], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[3], (size_t) 256), 1, 1};
while (n_threads[0] * n_threads[1] <= 256) while (n_threads[0] * n_threads[1] <= 256)
{ {
...@@ -1458,6 +1466,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1458,6 +1466,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_100(self, sio, node, name, x, z, fail): def c_code_reduce_100(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
in_dtype = "npy_" + node.inputs[0].dtype in_dtype = "npy_" + node.inputs[0].dtype
out_dtype = "npy_" + node.outputs[0].dtype out_dtype = "npy_" + node.outputs[0].dtype
...@@ -1476,7 +1485,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1476,7 +1485,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
# use blockIdx.y for i2 # use blockIdx.y for i2
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
if (PyGpuArray_STRIDES(%(x)s)[2] != sizeof(%(in_dtype)s)){ if (PyGpuArray_STRIDES(%(x)s)[2] != sizeof(%(in_dtype)s)){
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 256), 1, 1};
size_t n_blocks[3] = {std::min(PyGpuArray_DIMS(%(x)s)[1], (size_t)4096), 1, 1}; size_t n_blocks[3] = {std::min(PyGpuArray_DIMS(%(x)s)[1], (size_t)4096), 1, 1};
...@@ -1526,10 +1535,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1526,10 +1535,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_110(self, sio, node, name, x, z, fail): def c_code_reduce_110(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[1], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[1], (size_t) 256), 1, 1};
while (n_threads[0]*n_threads[1] <= 256) while (n_threads[0]*n_threads[1] <= 256)
{ {
...@@ -1545,10 +1555,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1545,10 +1555,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_001(self, sio, node, name, x, z, fail): def c_code_reduce_001(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[2], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[2], (size_t) 256), 1, 1};
size_t n_blocks[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 4096), 1, 1}; size_t n_blocks[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 4096), 1, 1};
while (n_blocks[0] * n_blocks[1] <= 4096) while (n_blocks[0] * n_blocks[1] <= 4096)
...@@ -1563,13 +1574,14 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1563,13 +1574,14 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_101(self, sio, node, name, x, z, fail): def c_code_reduce_101(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail, makecall = self._makecall(node, name, x, z, fail,
extra_dims=[("size_t one = 1;", "(void *) &one")], extra_dims=[("size_t one = 1;", "(void *) &one")],
extra_strides=[("ssize_t sone = 1;", "(void *) &sone")], extra_strides=[("ssize_t sone = 1;", "(void *) &sone")],
pattern="1011") pattern="1011")
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
// size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[3], // size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[3],
// (size_t) 256), 1, 1}; // (size_t) 256), 1, 1};
size_t n_threads[3] = {1, 1, 1}; size_t n_threads[3] = {1, 1, 1};
...@@ -1591,10 +1603,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1591,10 +1603,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_111(self, sio, node, name, x, z, fail): def c_code_reduce_111(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[2], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[2], (size_t) 256), 1, 1};
//get as many y threads as we can fit //get as many y threads as we can fit
...@@ -1623,13 +1636,14 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1623,13 +1636,14 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_0011(self, sio, node, name, x, z, fail): def c_code_reduce_0011(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
in_dtype = "npy_" + node.inputs[0].dtype in_dtype = "npy_" + node.inputs[0].dtype
out_dtype = "npy_" + node.outputs[0].dtype out_dtype = "npy_" + node.outputs[0].dtype
acc_dtype = "npy_" + self._acc_dtype(node.inputs[0].dtype) acc_dtype = "npy_" + self._acc_dtype(node.inputs[0].dtype)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_blocks[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 4096), 1, 1}; size_t n_blocks[3] = {std::min(PyGpuArray_DIMS(%(x)s)[0], (size_t) 4096), 1, 1};
...@@ -1652,10 +1666,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1652,10 +1666,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_1111(self, sio, node, name, x, z, fail): def c_code_reduce_1111(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[2], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[2], (size_t) 256), 1, 1};
//get as many y threads as we can fit //get as many y threads as we can fit
...@@ -1685,10 +1700,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1685,10 +1700,11 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_reduce_1011(self, sio, node, name, x, z, fail): def c_code_reduce_1011(self, sio, node, name, x, z, fail):
verbose = self.verbose
makecall = self._makecall(node, name, x, z, fail) makecall = self._makecall(node, name, x, z, fail)
print(""" print("""
{ {
int verbose = 0; int verbose = %(verbose)s;
size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[3], (size_t) 256), 1, 1}; size_t n_threads[3] = {std::min(PyGpuArray_DIMS(%(x)s)[3], (size_t) 256), 1, 1};
while (n_threads[0] * (n_threads[1]+1) <= 256) ++n_threads[1]; while (n_threads[0] * (n_threads[1]+1) <= 256) ++n_threads[1];
...@@ -1707,7 +1723,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1707,7 +1723,7 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
""" % locals(), file=sio) """ % locals(), file=sio)
def c_code_cache_version_apply(self, node): def c_code_cache_version_apply(self, node):
version = [23] # the version corresponding to the c code in this Op version = [24, self.verbose] # the version corresponding to the c code in this Op
# now we insert versions for the ops on which we depend... # now we insert versions for the ops on which we depend...
scalar_node = Apply( scalar_node = Apply(
...@@ -1758,9 +1774,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1758,9 +1774,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
const int threadCount = blockDim.x; const int threadCount = blockDim.x;
const int threadNum = threadIdx.x; const int threadNum = threadIdx.x;
extern __shared__ %(acc_type)s buf[]; extern __shared__ %(acc_type)s buf[];
%(acc_type)s myresult = %(reduce_init)s;
A = (const %(in_type)s *)(((char *)A)+offset_A); A = (const %(in_type)s *)(((char *)A)+offset_A);
Z = (%(out_type)s *)(((char *)Z)+offset_Z); Z = (%(out_type)s *)(((char *)Z)+offset_Z);
%(acc_type)s myresult = %(reduce_init)s;
for (int i0 = threadIdx.x; i0 < d0; i0 += blockDim.x) for (int i0 = threadIdx.x; i0 < d0; i0 += blockDim.x)
{ {
...@@ -1798,9 +1814,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1798,9 +1814,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
const int threadCount = blockDim.x; const int threadCount = blockDim.x;
const int threadNum = threadIdx.x; const int threadNum = threadIdx.x;
extern __shared__ %(acc_type)s buf[]; extern __shared__ %(acc_type)s buf[];
%(acc_type)s myresult = %(reduce_init)s;
A = (const %(in_type)s *)(((char *)A)+offset_A); A = (const %(in_type)s *)(((char *)A)+offset_A);
Z = (%(out_type)s *)(((char *)Z)+offset_Z); Z = (%(out_type)s *)(((char *)Z)+offset_Z);
%(acc_type)s myresult = %(reduce_init)s;
for (int i0 = threadIdx.x; i0 < d0; i0 += blockDim.x) for (int i0 = threadIdx.x; i0 < d0; i0 += blockDim.x)
{ {
...@@ -1839,9 +1855,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -1839,9 +1855,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
const int threadCount = blockDim.x * blockDim.y; const int threadCount = blockDim.x * blockDim.y;
const int threadNum = threadIdx.y*blockDim.x + threadIdx.x; const int threadNum = threadIdx.y*blockDim.x + threadIdx.x;
extern __shared__ %(acc_type)s buf[]; extern __shared__ %(acc_type)s buf[];
%(acc_type)s myresult = %(reduce_init)s;
A = (const %(in_type)s *)(((char *)A)+offset_A); A = (const %(in_type)s *)(((char *)A)+offset_A);
Z = (%(out_type)s *)(((char *)Z)+offset_Z); Z = (%(out_type)s *)(((char *)Z)+offset_Z);
%(acc_type)s myresult = %(reduce_init)s;
for (int i0 = threadIdx.y; i0 < d0; i0 += blockDim.y) for (int i0 = threadIdx.y; i0 < d0; i0 += blockDim.y)
{ {
...@@ -2112,9 +2128,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -2112,9 +2128,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
const int threadCount = blockDim.x * blockDim.y; const int threadCount = blockDim.x * blockDim.y;
const int threadNum = threadIdx.y * blockDim.x + threadIdx.x; const int threadNum = threadIdx.y * blockDim.x + threadIdx.x;
extern __shared__ %(acc_type)s buf[]; extern __shared__ %(acc_type)s buf[];
%(acc_type)s myresult = %(reduce_init)s;
A = (const %(in_type)s *)(((char *)A)+offset_A); A = (const %(in_type)s *)(((char *)A)+offset_A);
Z = (%(out_type)s *)(((char *)Z)+offset_Z); Z = (%(out_type)s *)(((char *)Z)+offset_Z);
%(acc_type)s myresult = %(reduce_init)s;
for (int i0 = threadIdx.y; i0 < d0; i0 += blockDim.y) for (int i0 = threadIdx.y; i0 < d0; i0 += blockDim.y)
{ {
...@@ -2375,9 +2391,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype): ...@@ -2375,9 +2391,9 @@ class GpuCAReduceCuda(GpuKernelBase, HideC, CAReduceDtype):
const int threadCount = blockDim.x * blockDim.y * blockDim.z; const int threadCount = blockDim.x * blockDim.y * blockDim.z;
const int threadNum = threadIdx.z * blockDim.x * blockDim.y + threadIdx.y * blockDim.x + threadIdx.x; const int threadNum = threadIdx.z * blockDim.x * blockDim.y + threadIdx.y * blockDim.x + threadIdx.x;
extern __shared__ %(acc_type)s buf[]; extern __shared__ %(acc_type)s buf[];
%(acc_type)s myresult = %(reduce_init)s;
A = (const %(in_type)s *)(((char *)A)+offset_A); A = (const %(in_type)s *)(((char *)A)+offset_A);
Z = (%(out_type)s *)(((char *)Z)+offset_Z); Z = (%(out_type)s *)(((char *)Z)+offset_Z);
%(acc_type)s myresult = %(reduce_init)s;
for (int i0 = threadIdx.z; i0 < d0; i0 += blockDim.z) for (int i0 = threadIdx.z; i0 < d0; i0 += blockDim.z)
{ {
......
theano/tensor/tests/test_elemwise.py
浏览文件 @ db1ce04e
...@@ -20,6 +20,7 @@ from theano.tensor.elemwise import (CAReduce, Elemwise, DimShuffle, ...@@ -20,6 +20,7 @@ from theano.tensor.elemwise import (CAReduce, Elemwise, DimShuffle,
Prod, ProdWithoutZeros) Prod, ProdWithoutZeros)
from theano.tests import unittest_tools from theano.tests import unittest_tools
from theano.tests.unittest_tools import attr from theano.tests.unittest_tools import attr
import theano.tests.unittest_tools as utt
def FunctionGraph(i, o): def FunctionGraph(i, o):
...@@ -482,8 +483,7 @@ class test_CAReduce(unittest_tools.InferShapeTester): ...@@ -482,8 +483,7 @@ class test_CAReduce(unittest_tools.InferShapeTester):
try: try:
f_xv = f(xv) f_xv = f(xv)
self.assertTrue((f_xv.shape == zv.shape), (f_xv, zv)) self.assertTrue((f_xv.shape == zv.shape), (f_xv, zv))
self.assertTrue(np.allclose(f_xv, zv), utt.assert_allclose(zv, f_xv)
(f_xv, zv, xsh, tosum))
except NotImplementedError: except NotImplementedError:
# GpuCAReduce don't implement all cases when size is 0 # GpuCAReduce don't implement all cases when size is 0
assert xv.size == 0 assert xv.size == 0
......
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论