Skip to content

Conversation

Naghasan
Copy link
Member

@Naghasan Naghasan commented Aug 4, 2020

Some devices can manage global values to push and pull different values.
In the CUDA runtime, this is covered by cudaMemcpyToSymbol/cudaMemcpyFromSymbol functions.
But other APIs may offer similar functionalities.

This proposal add the possibilities to manage such global values via a new class and new sycl::handler member functions.

The proposal is still a bit incomplete as regular C++ global variable can potentially be handled by the sycl::handler extension
without requiring the new 'sycl::symbol' class. but limitation have yet to be explored.

Some devices can manage global values to push and pull different values.
In the CUDA runtime, this is covered by cudaMemcpyToSymbol/cudaMemcpyFromSymbol functions.
But other APIs may offer similar functionalities.

This proposal add the possibilities to manage such global values via a new class and new sycl::handler member functions.

The proposal is still a bit incomplete as regular C++ global variable can potentially be handled by the sycl::handler extension
without requiring the new 'sycl::symbol' class. but limitation have yet to be explored.
@Naghasan Naghasan requested review from Ruyk and AerialMantis August 4, 2020 12:22
Copy link
Contributor

@keryell keryell left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That looks quite reasonable.
Even some C++17 here. :-)
I am curious about the mutable kernel lambdas. Is this a side effect of the simple mock-up?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants