Optimize implementation of cuda::std::conditional_t
#2779
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
miscco
requested review from
wmaxey,
gevtushenko,
bernhardmgruber and
ericniebler
and removed request for
gevtushenko
miscco
force-pushed
the
optimize_conditional
branch
from
612fe7a to
c2c5a18
Compare
The current implementation always instantiates `conditional` at least once. However, we can get aways with a total of 2 class instantiations by using the implementation of the internal `_If` This not only makes `conditional_t` more efficient but also simplifies usage, because no one has to think about what to take
miscco
force-pushed
the
optimize_conditional
branch
from
c2c5a18 to
fc1a121
Compare
ericniebler
approved these changes
Nov 13, 2024
Comment on lines
+458
to
+461
| using BlevBufferSrcsOutT = | ||
| typename ::cuda::std::conditional_t<IsMemcpy, const void*, cub::detail::value_t<InputBufferIt>>; | ||
| using BlevBufferDstOutT = | ||
| typename ::cuda::std::conditional_t<IsMemcpy, void*, cub::detail::value_t<OutputBufferIt>>; |
There was a problem hiding this comment.
Suggested change
| using BlevBufferSrcsOutT = | |
| typename ::cuda::std::conditional_t<IsMemcpy, const void*, cub::detail::value_t<InputBufferIt>>; | |
| using BlevBufferDstOutT = | |
| typename ::cuda::std::conditional_t<IsMemcpy, void*, cub::detail::value_t<OutputBufferIt>>; | |
| using BlevBufferSrcsOutT = | |
| ::cuda::std::conditional_t<IsMemcpy, const void*, cub::detail::value_t<InputBufferIt>>; | |
| using BlevBufferDstOutT = | |
| ::cuda::std::conditional_t<IsMemcpy, void*, cub::detail::value_t<OutputBufferIt>>; |
| @@ -178,8 +178,7 @@ struct csr_matrix | |||
|
|
|||
| void print_internals(std::ostream& out) const | |||
| { | |||
| out << (HostStorage ? "host" : "device") << "_csr_matrix" | |||
| << "(" << m_num_rows << ", " << m_num_columns << ")\n" | |||
| out << (HostStorage ? "host" : "device") << "_csr_matrix" << "(" << m_num_rows << ", " << m_num_columns << ")\n" | |||
There was a problem hiding this comment.
i wonder why this changed
| struct __conditional_impl<false> | ||
| { | ||
| template <class _If, class _Else> | ||
| using type _CCCL_NODEBUG_ALIAS = _Else; |
There was a problem hiding this comment.
it doesn't much matter in this case, but our metaprogramming utilities like to call this nested alias template __call. then conditional_t could be an alias for __type_call2<__conditional_impl<_Cond>, _If, _Else>.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The current implementation always instantiates
conditionalat least once.However, we can get aways with a total of 2 class instantiations by using the implementation of the internal
_Ifalias template.This not only makes
conditional_tmore efficient but also simplifies usage, because no one has to think about what to take