Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

shared memory update #117

Merged
merged 4 commits into from
Nov 10, 2023
Merged

shared memory update #117

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
3617c56
hide shared memory from user
pkicsiny Sep 8, 2023
cacffac
Merge branch 'main' of https://github.com/pkicsiny/xobjects into main
pkicsiny Sep 8, 2023
22e351a
Merge branch 'xsuite:main' into main
pkicsiny Sep 30, 2023
29e0969
Merge branch 'release/0.2.10' into main
giadarol Nov 10, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 10 additions & 12 deletions xobjects/context_cupy.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -632,14 +632,14 @@ def to_pointer_arg(self, offset, nbytes):


class KernelCupy(object):
def __init__(
self, function, description, block_size, context, shared_mem_size_bytes
):
def __init__(self, function, description, block_size, context,
shared_mem_size_bytes):

self.function = function
self.description = description
self.block_size = block_size
self.context = context
self.shared_mem_size_bytes = shared_mem_size_bytes
self.context = context

def to_function_arg(self, arg, value):
if arg.pointer:
Expand Down Expand Up @@ -671,7 +671,7 @@ def to_function_arg(self, arg, value):
def num_args(self):
return len(self.description.args)

def __call__(self, **kwargs):
def __call__(self, shared_mem_size_bytes=None, **kwargs):
assert len(kwargs.keys()) == self.num_args
arg_list = []
for arg in self.description.args:
Expand All @@ -683,14 +683,12 @@ def __call__(self, **kwargs):
else:
n_threads = self.description.n_threads

grid_size = int(np.ceil(n_threads / self.block_size))
self.function(
(grid_size,),
(self.block_size,),
arg_list,
shared_mem=self.shared_mem_size_bytes,
)
if shared_mem_size_bytes is None:
shared_mem_size_bytes = self.shared_mem_size_bytes

grid_size = int(np.ceil(n_threads / self.block_size))
self.function((grid_size,), (self.block_size,), arg_list,
shared_mem=shared_mem_size_bytes)

class FFTCupy(object):
def __init__(self, context, data, axes):
Expand Down
Loading