Branch: refs/heads/staging
Home: https://github.com/STEllAR-GROUP/hpx
Commit: 512c6d23cebe6a40848a2e9521750e83bcba9394
https://github.com/STEllAR-GROUP/hpx/commit/512c6d23cebe6a40848a2e9521750e83bcba9394
Author: Hartmut Kaiser <[email protected]>
Date: 2022-11-21 (Mon, 21 Nov 2022)
Changed paths:
M libs/core/concurrency/CMakeLists.txt
M
libs/core/concurrency/include/hpx/concurrency/detail/contiguous_index_queue.hpp
A
libs/core/concurrency/include/hpx/concurrency/detail/non_contiguous_index_queue.hpp
M libs/core/concurrency/tests/unit/CMakeLists.txt
M libs/core/concurrency/tests/unit/contiguous_index_queue.cpp
A libs/core/concurrency/tests/unit/non_contiguous_index_queue.cpp
M libs/core/coroutines/include/hpx/coroutines/thread_enums.hpp
M libs/core/executors/include/hpx/executors/thread_pool_scheduler_bulk.hpp
M libs/core/executors/tests/unit/explicit_scheduler_executor.cpp
Log Message:
-----------
Add non-contiguous index queue and use it in thread_pool_bulk_scheduler
- add new thread_placement_hint to thread_schedule_hint structure (breadth-first
and depth-first)
- change thread_pool_scheduler_bulk to support placement hint
Commit: de1c7478e483ad371cf47c426391ff637a955ecc
https://github.com/STEllAR-GROUP/hpx/commit/de1c7478e483ad371cf47c426391ff637a955ecc
Author: Hartmut Kaiser <[email protected]>
Date: 2022-11-21 (Mon, 21 Nov 2022)
Changed paths:
M libs/core/algorithms/CMakeLists.txt
M libs/core/algorithms/include/hpx/parallel/algorithms/adjacent_find.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/all_any_none.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/detail/search.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/detail/upper_lower_bound.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/equal.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/find.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/is_heap.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/is_sorted.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/lexicographical_compare.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/mismatch.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_copy.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_default_construct.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_fill.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_move.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_value_construct.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/adjacent_find.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/equal.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/find.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/mismatch.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/transform_loop.hpp
A libs/core/algorithms/include/hpx/parallel/util/adapt_placement_mode.hpp
M libs/core/algorithms/include/hpx/parallel/util/loop.hpp
M libs/core/algorithms/include/hpx/parallel/util/partitioner.hpp
M libs/core/algorithms/include/hpx/parallel/util/transform_loop.hpp
M libs/core/algorithms/tests/unit/algorithms/CMakeLists.txt
A libs/core/algorithms/tests/unit/algorithms/find_sender.cpp
M libs/core/algorithms/tests/unit/algorithms/find_tests.hpp
M
libs/core/concurrency/include/hpx/concurrency/detail/contiguous_index_queue.hpp
M
libs/core/concurrency/include/hpx/concurrency/detail/non_contiguous_index_queue.hpp
M libs/core/concurrency/tests/unit/contiguous_index_queue.cpp
M libs/core/concurrency/tests/unit/non_contiguous_index_queue.cpp
M libs/core/coroutines/include/hpx/coroutines/thread_enums.hpp
M libs/core/executors/CMakeLists.txt
M libs/core/executors/include/hpx/executors/execution_policy_annotation.hpp
A libs/core/executors/include/hpx/executors/execution_policy_hint.hpp
M libs/core/executors/include/hpx/executors/thread_pool_scheduler_bulk.hpp
M libs/core/executors/tests/unit/explicit_scheduler_executor.cpp
Log Message:
-----------
Using breadth-first placement mode for algorithms that use cancellation
- flyby: add missing #includes
- flyby: adapt hpx::find for S/R
Commit: 7445e8ad10ad7dcae01038921bc6a6c178512907
https://github.com/STEllAR-GROUP/hpx/commit/7445e8ad10ad7dcae01038921bc6a6c178512907
Author: StellarBot <[email protected]>
Date: 2022-11-23 (Wed, 23 Nov 2022)
Changed paths:
M libs/core/algorithms/CMakeLists.txt
M libs/core/algorithms/include/hpx/parallel/algorithms/adjacent_find.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/all_any_none.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/detail/search.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/detail/upper_lower_bound.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/equal.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/find.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/is_heap.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/is_sorted.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/lexicographical_compare.hpp
M libs/core/algorithms/include/hpx/parallel/algorithms/mismatch.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_copy.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_default_construct.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_fill.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_move.hpp
M
libs/core/algorithms/include/hpx/parallel/algorithms/uninitialized_value_construct.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/adjacent_find.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/equal.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/find.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/mismatch.hpp
M libs/core/algorithms/include/hpx/parallel/datapar/transform_loop.hpp
A libs/core/algorithms/include/hpx/parallel/util/adapt_placement_mode.hpp
M libs/core/algorithms/include/hpx/parallel/util/loop.hpp
M libs/core/algorithms/include/hpx/parallel/util/partitioner.hpp
M libs/core/algorithms/include/hpx/parallel/util/transform_loop.hpp
M libs/core/algorithms/tests/unit/algorithms/CMakeLists.txt
A libs/core/algorithms/tests/unit/algorithms/find_sender.cpp
M libs/core/algorithms/tests/unit/algorithms/find_tests.hpp
M libs/core/concurrency/CMakeLists.txt
M
libs/core/concurrency/include/hpx/concurrency/detail/contiguous_index_queue.hpp
A
libs/core/concurrency/include/hpx/concurrency/detail/non_contiguous_index_queue.hpp
M libs/core/concurrency/tests/unit/CMakeLists.txt
M libs/core/concurrency/tests/unit/contiguous_index_queue.cpp
A libs/core/concurrency/tests/unit/non_contiguous_index_queue.cpp
M libs/core/coroutines/include/hpx/coroutines/thread_enums.hpp
M libs/core/executors/CMakeLists.txt
M libs/core/executors/include/hpx/executors/execution_policy_annotation.hpp
A libs/core/executors/include/hpx/executors/execution_policy_hint.hpp
M libs/core/executors/include/hpx/executors/thread_pool_scheduler_bulk.hpp
M libs/core/executors/tests/unit/explicit_scheduler_executor.cpp
Log Message:
-----------
Merge #6051
6051: Add non-contiguous index queue and use it in thread_pool_bulk_scheduler
r=hkaiser a=hkaiser
This changes the sequence of partitions scheduled on the same core.
`@hcq9102` Could you please run your performance scripts for the S/R executor
using PR to see if this behaves differently for larger numbers of cores?
Co-authored-by: Hartmut Kaiser <[email protected]>
Compare:
https://github.com/STEllAR-GROUP/hpx/compare/09c47d373f45...7445e8ad10ad
_______________________________________________
hpx-commits mailing list
[email protected]
https://mail.cct.lsu.edu/mailman/listinfo/hpx-commits
