def _axpbyz(self,
selffac,
other,
otherfac,
out,
add_timer=None,
stream=None):
"""Compute ``out = selffac * self + otherfac*other``,
where `other` is a vector.."""
assert self.shape == other.shape
if not self.flags.forc or not other.flags.forc:
raise RuntimeError("only contiguous arrays may "
"be used as arguments to this operation")
func = elementwise.get_axpbyz_kernel(self.dtype, other.dtype,
out.dtype)
if add_timer is not None:
add_timer(
3 * self.size,
func.prepared_timed_call(self._grid, selffac, self.gpudata,
otherfac, other.gpudata, out.gpudata,
self.mem_size))
else:
func.prepared_async_call(self._grid, self._block, stream, selffac,
self.gpudata, otherfac, other.gpudata,
out.gpudata, self.mem_size)
return out
def _axpbyz(self, selffac, other, otherfac, out, add_timer=None, stream=None):
"""Compute ``out = selffac * self + otherfac*other``,
where `other` is a vector.."""
assert self.shape == other.shape
if not self.flags.forc or not other.flags.forc:
raise RuntimeError("only contiguous arrays may " "be used as arguments to this operation")
func = elementwise.get_axpbyz_kernel(self.dtype, other.dtype, out.dtype)
if add_timer is not None:
add_timer(
3 * self.size,
func.prepared_timed_call(
self._grid, selffac, self.gpudata, otherfac, other.gpudata, out.gpudata, self.mem_size
),
)
else:
func.prepared_async_call(
self._grid,
self._block,
stream,
selffac,
self.gpudata,
otherfac,
other.gpudata,
out.gpudata,
self.mem_size,
)
return out
def _axpbyz(self,
selffac,
other,
otherfac,
out,
add_timer=None,
stream=None):
"""Compute ``out = selffac * self + otherfac*other``,
where `other` is a vector.."""
assert self.shape == other.shape
func = elementwise.get_axpbyz_kernel(self.dtype, other.dtype,
out.dtype)
func.set_block_shape(*self._block)
if add_timer is not None:
add_timer(
3 * self.size,
func.prepared_timed_call(self._grid, selffac, self.gpudata,
otherfac, other.gpudata, out.gpudata,
self.mem_size))
else:
func.prepared_async_call(self._grid, stream, selffac, self.gpudata,
otherfac, other.gpudata, out.gpudata,
self.mem_size)
return out
def _axpbyz(self, selffac, other, otherfac, out, add_timer=None, stream=None):
"""Compute ``out = selffac * self + otherfac*other``,
where `other` is a vector.."""
assert self.shape == other.shape
func = elementwise.get_axpbyz_kernel(self.dtype, other.dtype, out.dtype)
func.set_block_shape(*self._block)
if add_timer is not None:
add_timer(3*self.size, func.prepared_timed_call(self._grid,
selffac, self.gpudata, otherfac, other.gpudata,
out.gpudata, self.mem_size))
else:
func.prepared_async_call(self._grid, stream,
selffac, self.gpudata, otherfac, other.gpudata,
out.gpudata, self.mem_size)
return out
