-
Notifications
You must be signed in to change notification settings - Fork 407
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.
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
Allow Kokkos::resize to take an execution space #4844
Conversation
23e2621
to
96fc533
Compare
Retest this please. |
4b02500
to
776ab5e
Compare
776ab5e
to
b2f11e4
Compare
f055b8c
to
98b7b67
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
A few questions.
containers/src/Kokkos_DualView.hpp
Outdated
static_assert( | ||
!alloc_prop_input::allow_padding, | ||
"The view constructor arguments passed to Kokkos::resize must " | ||
"not explicitly allow padding!"); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why this one?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In the discussion around #5035 (comment) we said that we shouldn't be allowed to explicitly request padding but I'm happy to change if that's what we want.
Seems to work with |
Retest this please |
Related to #4792, #4823, and #4826. Specifying an execution space, we could at least get rid of a global fence (and possibly some instance fences depending on #4823 and #4826).
If the View's count is 1, we of course still deallocate data which implies a global fence.
So far tests are missing andthis is mostly intended to discuss if we want to provide these overloads.