-
Notifications
You must be signed in to change notification settings - Fork 304
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
Rename multiple .cuh (.cu) files to .hpp (.cpp) #2501
Rename multiple .cuh (.cu) files to .hpp (.cpp) #2501
Conversation
Codecov Report
@@ Coverage Diff @@
## branch-22.10 #2501 +/- ##
===============================================
Coverage ? 61.26%
===============================================
Files ? 106
Lines ? 5404
Branches ? 0
===============================================
Hits ? 3311
Misses ? 2093
Partials ? 0 Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
template <typename Iterator, typename TupleType, size_t I, size_t N> | ||
struct atomic_accumulate_thrust_tuple_impl { | ||
__device__ constexpr void compute(Iterator iter, TupleType const& value) const | ||
{ | ||
atomic_accumulate_impl(thrust::raw_reference_cast(thrust::get<I>(*iter)), | ||
thrust::get<I>(value)); | ||
atomic_accumulate_thrust_tuple_impl<Iterator, TupleType, I + 1, N>().compute(iter, value); | ||
} | ||
}; | ||
|
||
template <typename Iterator, typename TupleType, size_t I> | ||
struct atomic_accumulate_thrust_tuple_impl<Iterator, TupleType, I, I> { | ||
__device__ constexpr void compute(Iterator iter, TupleType const& value) const {} | ||
}; | ||
|
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.
@seunghwak If the tuple has too many element (not sure if it can be the case), would it be a bit inefficient in terms of number of function instantiates it would generate? (Not related to this PR though)
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.
As we discussed offline, this will end up with N atomicAdd instructions. So, no in terms of execution time, but yes, this can increase compile time.
@gpucibot merge |
Pre-requisites for #2479
Several .cuh header files (.cuh indicates nvcc is required to compile) can be renamed to .hpp (.hpp indicates the file can be compiled with a host compiler such as g++) with no or minimum modifications.
This PR performs multiple renaming (eventually to separate edge_partition_src_dst_property.cuh to .hpp and .cuh files).
Breaking as public header files are renamed.