Adding local work stealing scheduler that is based on message passing…

… internally - Using uniform_int_distribution with proper bounds - Removing queue index from thread_queues as it was unused - flyby: remove commented out options from .clang-format - Renaming workstealing --> workrequesting - Adding adaptive work stealing (steal half/steal one) - this makes this scheduler consistently (albeit only slightly) faster than the (default) local-priority scheduler - Adding LIFO and FIFO variations of local work-stealing scheduler - flyby: fixing HPX_WITH_SWAP_CONTEXT_EMULATION - flyby: minor changes to fibonacci_local example - Adding high- and low- priority queues - flyby: cache_line_data now does not generate warnings errors if padding is not needed
STEllAR-GROUP · Apr 13, 2022 · a98609d · a98609d
1 parent 85fdb87
commit a98609d
Show file tree

Hide file tree

Showing 44 changed files with 2,115 additions and 282 deletions.
diff --git a/.circleci/config.yml b/.circleci/config.yml
@@ -233,7 +233,8 @@ jobs:
                 -DHPX_WITH_CHECK_MODULE_DEPENDENCIES=On \
                 -DCMAKE_EXPORT_COMPILE_COMMANDS=On \
                 -DHPX_WITH_DOCUMENTATION=On \
-                -DHPX_WITH_DOCUMENTATION_OUTPUT_FORMATS="${DOCUMENTATION_OUTPUT_FORMATS}"
+                -DHPX_WITH_DOCUMENTATION_OUTPUT_FORMATS="${DOCUMENTATION_OUTPUT_FORMATS}" \
+                -DHPX_WITH_TESTS_COMMAND_LINE=--hpx:queuing=local-workrequesting-fifo
       - persist_to_workspace:
           root: /hpx
           paths:

diff --git a/.clang-format b/.clang-format
@@ -1,5 +1,5 @@
 # Copyright (c) 2016 Thomas Heller
-# Copyright (c) 2016-2018 Hartmut Kaiser
+# Copyright (c) 2016-2019 Hartmut Kaiser
 #
 # SPDX-License-Identifier: BSL-1.0
 # Distributed under the Boost Software License, Version 1.0. (See accompanying
@@ -105,7 +105,7 @@ PenaltyExcessCharacter: 1000000
 PenaltyReturnTypeOnItsOwnLine: 200
 PointerAlignment: Left
 ReflowComments: false
-SortIncludes:    true
+SortIncludes: true
 SpaceAfterCStyleCast: true
 SpaceAfterTemplateKeyword: true
 SpaceBeforeAssignmentOperators: true

diff --git a/.jenkins/lsu/env-clang-13.sh b/.jenkins/lsu/env-clang-13.sh
@@ -23,8 +23,9 @@ configure_extra_options+=" -DHPX_WITH_COMPILER_WARNINGS_AS_ERRORS=ON"
 configure_extra_options+=" -DHPX_WITH_PARCELPORT_MPI=ON"
 configure_extra_options+=" -DHPX_WITH_PARCELPORT_LCI=ON"
 configure_extra_options+=" -DHPX_WITH_FETCH_LCI=ON"
-configure_extra_options+=" -DLCI_SERVER=ibv"
 configure_extra_options+=" -DHPX_WITH_LOGGING=OFF"
+configure_extra_options+=" -DHPX_WITH_TESTS_COMMAND_LINE=--hpx:queuing=local-workrequesting-fifo"
+configure_extra_options+=" -DLCI_SERVER=ibv"
 
 # Make sure HWLOC does not report 'cores'. This is purely an option to enable
 # testing the topology code under conditions close to those on FreeBSD.

diff --git a/.jenkins/lsu/env-gcc-11.sh b/.jenkins/lsu/env-gcc-11.sh
@@ -14,6 +14,7 @@ module load openmpi
 export HPXRUN_RUNWRAPPER=srun
 export CXX_STD="20"
 
+configure_extra_options+=" -DCMAKE_BUILD_TYPE=${build_type}"
 configure_extra_options+=" -DHPX_WITH_CXX_STANDARD=${CXX_STD}"
 configure_extra_options+=" -DHPX_WITH_MALLOC=system"
 configure_extra_options+=" -DHPX_WITH_FETCH_ASIO=ON"
@@ -22,4 +23,5 @@ configure_extra_options+=" -DHPX_WITH_COMPILER_WARNINGS_AS_ERRORS=ON"
 configure_extra_options+=" -DHPX_WITH_PARCELPORT_MPI=ON"
 configure_extra_options+=" -DHPX_WITH_PARCELPORT_LCI=ON"
 configure_extra_options+=" -DHPX_WITH_FETCH_LCI=ON"
+configure_extra_options+=" -DHPX_WITH_TESTS_COMMAND_LINE=--hpx:queuing=local-workrequesting-fifo"
 configure_extra_options+=" -DLCI_SERVER=ibv"
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -1388,6 +1388,13 @@ hpx_option(
   ADVANCED
 )
 
+hpx_option(
+  HPX_WITH_TESTS_COMMAND_LINE STRING
+  "Add given command line options to all tests run" ""
+  CATEGORY "Debugging"
+  ADVANCED
+)
+
 hpx_option(
   HPX_WITH_TESTS_MAX_THREADS_PER_LOCALITY
   STRING

diff --git a/cmake/HPX_AddTest.cmake b/cmake/HPX_AddTest.cmake
@@ -66,6 +66,11 @@ function(add_hpx_test category name)
     )
   endif()
 
+  # add additional command line arguments to all test executions
+  if(NOT x${HPX_WITH_TESTS_COMMAND_LINE} STREQUAL x"")
+    set(args ${args} "${HPX_WITH_TESTS_COMMAND_LINE}")
+  endif()
+
   if(${HPX_WITH_PARALLEL_TESTS_BIND_NONE}
      AND NOT run_serial
      AND NOT "${name}_RUNWRAPPER"

diff --git a/docs/sphinx/manual/hpx_runtime_and_resources.rst b/docs/sphinx/manual/hpx_runtime_and_resources.rst
@@ -16,13 +16,10 @@
 |hpx| thread scheduling policies
 ================================
 
-The HPX runtime has five thread scheduling policies: local-priority,
-static-priority, local, static and abp-priority. These policies can be specified
-from the command line using the command line option :option:`--hpx:queuing`. In
-order to use a particular scheduling policy, the runtime system must be built
-with the appropriate scheduler flag turned on (e.g. ``cmake
--DHPX_THREAD_SCHEDULERS=local``, see :ref:`cmake_variables` for more
-information).
+The HPX runtime has several thread scheduling policies: local-priority-fifo,
+local-priority-lifo, static-priority, local, static, local-workrequesting-fifo,
+local-workrequesting-lifo, and abp-priority. These policies can be specified
+from the command line using the command line option :option:`--hpx:queuing`. 
 
 Priority local scheduling policy (default policy)
 -------------------------------------------------
@@ -54,8 +51,6 @@ Static priority scheduling policy
 ---------------------------------
 
 * invoke using: :option:`--hpx:queuing`\ ``=static-priority`` (or ``-qs``)
-* flag to turn on for build: ``HPX_THREAD_SCHEDULERS=all`` or
-  ``HPX_THREAD_SCHEDULERS=static-priority``
 
 The static scheduling policy maintains one queue per OS thread from which each
 OS thread pulls its tasks (user threads). Threads are distributed in a round
@@ -65,8 +60,6 @@ Local scheduling policy
 -----------------------
 
 * invoke using: :option:`--hpx:queuing`\ ``=local`` (or ``-ql``)
-* flag to turn on for build: ``HPX_THREAD_SCHEDULERS=all`` or
-  ``HPX_THREAD_SCHEDULERS=local``
 
 The local scheduling policy maintains one queue per OS thread from which each OS
 thread pulls its tasks (user threads).
@@ -75,8 +68,6 @@ Static scheduling policy
 ------------------------
 
 * invoke using: :option:`--hpx:queuing`\ ``=static``
-* flag to turn on for build: ``HPX_THREAD_SCHEDULERS=all`` or
-  ``HPX_THREAD_SCHEDULERS=static``
 
 The static scheduling policy maintains one queue per OS thread from which each
 OS thread pulls its tasks (user threads). Threads are distributed in a round
@@ -86,8 +77,6 @@ Priority ABP scheduling policy
 ------------------------------
 
 * invoke using: :option:`--hpx:queuing`\ ``=abp-priority-fifo``
-* flag to turn on for build: ``HPX_THREAD_SCHEDULERS=all`` or
-  ``HPX_THREAD_SCHEDULERS=abp-priority``
 
 Priority ABP policy maintains a double ended lock free queue for each OS thread.
 By default the number of high priority queues is equal to the number of OS
@@ -153,6 +142,23 @@ policy use the command line option :option:`--hpx:queuing`\
 
     I see both FIFO and double ended queues in ABP policies?
 
+Work requesting scheduling policies
+-----------------------------------
+
+* invoke using: :option:`--hpx:queuing`\ ``=local-workrequesting-fifo``
+  or using :option:`--hpx:queuing`\ ``=local-workrequesting-lifo``
+
+The work-requesting policies rely on a different mechanism of balancing work
+between cores (compared to the other policies listed above). Instead of actively
+trying to steal work from other cores, requesting work relies on a less
+disruptive mechanism. If a core runs out of work, instead of actively looking at
+the queues of neighboring cores, in this case a request is posted to another
+core. This core now (whenever it is not busy with other work) either responds to
+the original core by sending back work or passes the request on to the next
+possible core in the system. In general, this scheme avoids contention on the
+work queues as those are always accessed by their own cores only.
+
+
 The |hpx| resource partitioner
 ==============================
 

diff --git a/docs/sphinx/manual/launching_and_configuring_hpx_applications.rst b/docs/sphinx/manual/launching_and_configuring_hpx_applications.rst
@@ -1518,7 +1518,9 @@ The predefined command line options for any application using
 
    the queue scheduling policy to use, options are ``local``,
    ``local-priority-fifo``, ``local-priority-lifo``, ``static``,
-   ``static-priority``, ``abp-priority-fifo`` and ``abp-priority-lifo``
+   ``static-priority``, ``abp-priority-fifo``,
+   ``local-workrequesting-fifo``, ``local-workrequesting-lifo``
+   and ``abp-priority-lifo``
    (default: ``local-priority-fifo``)
 
 .. option:: --hpx:high-priority-threads arg

diff --git a/examples/quickstart/fibonacci_local.cpp b/examples/quickstart/fibonacci_local.cpp
@@ -18,6 +18,9 @@
 
 #include <cstdint>
 #include <iostream>
+#include <string>
+#include <utility>
+#include <vector>
 
 ///////////////////////////////////////////////////////////////////////////////
 //[fibonacci
@@ -26,22 +29,20 @@ std::uint64_t fibonacci(std::uint64_t n)
     if (n < 2)
         return n;
 
-    // Invoking the Fibonacci algorithm twice is inefficient.
-    // However, we intentionally demonstrate it this way to create some
-    // heavy workload.
-
     hpx::future<std::uint64_t> n1 = hpx::async(fibonacci, n - 1);
-    hpx::future<std::uint64_t> n2 = hpx::async(fibonacci, n - 2);
+    std::uint64_t n2 = fibonacci(n - 2);
 
-    return n1.get() +
-        n2.get();    // wait for the Futures to return their values
+    return n1.get() + n2;    // wait for the Future to return their values
 }
 //fibonacci]
 
 ///////////////////////////////////////////////////////////////////////////////
 //[hpx_main
 int hpx_main(hpx::program_options::variables_map& vm)
 {
+    hpx::threads::add_scheduler_mode(
+        hpx::threads::policies::scheduler_mode::fast_idle_mode);
+
     // extract command line argument, i.e. fib(N)
     std::uint64_t n = vm["n-value"].as<std::uint64_t>();
 
@@ -67,9 +68,13 @@ int main(int argc, char* argv[])
     hpx::program_options::options_description desc_commandline(
         "Usage: " HPX_APPLICATION_STRING " [options]");
 
-    desc_commandline.add_options()("n-value",
-        hpx::program_options::value<std::uint64_t>()->default_value(10),
-        "n value for the Fibonacci function");
+    // clang-format off
+    desc_commandline.add_options()
+        ("n-value",
+            hpx::program_options::value<std::uint64_t>()->default_value(10),
+            "n value for the Fibonacci function")
+        ;
+    // clang-format on
 
     // Initialize and run HPX
     hpx::local::init_params init_args;

diff --git a/libs/core/command_line_handling_local/src/parse_command_line_local.cpp b/libs/core/command_line_handling_local/src/parse_command_line_local.cpp
@@ -327,18 +327,24 @@ namespace hpx { namespace local { namespace detail {
                   "run on (default: 0), valid for "
                   "--hpx:queuing=local, --hpx:queuing=abp-priority, "
                   "--hpx:queuing=static, --hpx:queuing=static-priority, "
+                  "--hpx:queuing=local-workrequesting-fifo, "
+                  "--hpx:queuing=local-workrequesting-lifo, "
                   "and --hpx:queuing=local-priority only")
                 ("hpx:pu-step", value<std::size_t>(),
                   "the step between used processing unit numbers for this "
                   "instance of HPX (default: 1), valid for "
                   "--hpx:queuing=local, --hpx:queuing=abp-priority, "
-                  "--hpx:queuing=static, --hpx:queuing=static-priority "
+                  "--hpx:queuing=static, --hpx:queuing=static-priority, "
+                  "--hpx:queuing=local-workrequesting-fifo, "
+                  "--hpx:queuing=local-workrequesting-lifo, "
                   "and --hpx:queuing=local-priority only")
                 ("hpx:affinity", value<std::string>(),
                   "the affinity domain the OS threads will be confined to, "
                   "possible values: pu, core, numa, machine (default: pu), valid for "
                   "--hpx:queuing=local, --hpx:queuing=abp-priority, "
-                  "--hpx:queuing=static, --hpx:queuing=static-priority "
+                  "--hpx:queuing=static, --hpx:queuing=static-priority, "
+                  "--hpx:queuing=local-workrequesting-fifo, "
+                  "--hpx:queuing=local-workrequesting-lifo, "
                   " and --hpx:queuing=local-priority only")
                 ("hpx:bind", value<std::vector<std::string> >()->composing(),
                   "the detailed affinity description for the OS threads, see "
@@ -363,13 +369,16 @@ namespace hpx { namespace local { namespace detail {
                 ("hpx:queuing", value<std::string>(),
                   "the queue scheduling policy to use, options are "
                   "'local', 'local-priority-fifo','local-priority-lifo', "
+                  "'local-workrequesting-fifo', 'local-workrequesting-lifo' "
                   "'abp-priority-fifo', 'abp-priority-lifo', 'static', and "
-                  "'static-priority' (default: 'local-priority'; "
+                  "'static-priority' (default: 'local-priority-fifo'; "
                   "all option values can be abbreviated)")
                 ("hpx:high-priority-threads", value<std::size_t>(),
                   "the number of operating system threads maintaining a high "
                   "priority queue (default: number of OS threads), valid for "
                   "--hpx:queuing=local-priority,--hpx:queuing=static-priority, "
+                  "--hpx:queuing=local-workrequesting-fifo, "
+                  "--hpx:queuing=local-workrequesting-lifo, "
                   " and --hpx:queuing=abp-priority only)")
                 ("hpx:numa-sensitive", value<std::size_t>()->implicit_value(0),
                   "makes the local-priority scheduler NUMA sensitive ("

diff --git a/libs/core/concurrency/include/hpx/concurrency/detail/contiguous_index_queue.hpp b/libs/core/concurrency/include/hpx/concurrency/detail/contiguous_index_queue.hpp
@@ -139,7 +139,7 @@ namespace hpx { namespace concurrency { namespace detail {
             {
                 if (expected_range.empty())
                 {
-                    return hpx::nullopt;
+                    return hpx::optional<T>(hpx::nullopt);
                 }
 
                 index = expected_range.first;
@@ -166,7 +166,7 @@ namespace hpx { namespace concurrency { namespace detail {
             {
                 if (expected_range.empty())
                 {
-                    return hpx::nullopt;
+                    return hpx::optional<T>(hpx::nullopt);
                 }
 
                 desired_range = expected_range.decrement_last();

diff --git a/libs/core/concurrency/include/hpx/concurrency/spinlock.hpp b/libs/core/concurrency/include/hpx/concurrency/spinlock.hpp
@@ -13,9 +13,11 @@
 #include <hpx/modules/itt_notify.hpp>
 #include <hpx/thread_support/spinlock.hpp>
 
+#include <utility>
+
 namespace hpx { namespace util {
 
-    /// Lockable spinlock class
+    // Lockable spinlock class
     struct spinlock
     {
     public:
@@ -25,25 +27,32 @@ namespace hpx { namespace util {
         hpx::util::detail::spinlock m;
 
     public:
-        spinlock(char const* /*desc*/ = nullptr)
+        spinlock() noexcept
         {
-            HPX_ITT_SYNC_CREATE(this, "util::spinlock", "");
+            HPX_ITT_SYNC_CREATE(this, "util::spinlock", nullptr);
+        }
+
+        explicit spinlock(char const* desc) noexcept
+        {
+            HPX_ITT_SYNC_CREATE(this, "util::spinlock", desc);
         }
 
         ~spinlock()
         {
             HPX_ITT_SYNC_DESTROY(this);
         }
 
-        void lock() noexcept
+        void lock() noexcept(
+            noexcept(util::register_lock(std::declval<spinlock*>())))
         {
             HPX_ITT_SYNC_PREPARE(this);
             m.lock();
             HPX_ITT_SYNC_ACQUIRED(this);
             util::register_lock(this);
         }
 
-        bool try_lock() noexcept
+        bool try_lock() noexcept(
+            noexcept(util::register_lock(std::declval<spinlock*>())))
         {
             HPX_ITT_SYNC_PREPARE(this);
             if (m.try_lock())
@@ -56,13 +65,13 @@ namespace hpx { namespace util {
             return false;
         }
 
-        void unlock() noexcept
+        void unlock() noexcept(
+            noexcept(util::unregister_lock(std::declval<spinlock*>())))
         {
             HPX_ITT_SYNC_RELEASING(this);
             m.unlock();
             HPX_ITT_SYNC_RELEASED(this);
             util::unregister_lock(this);
         }
     };
-
 }}    // namespace hpx::util
diff --git a/libs/core/resource_partitioner/include/hpx/resource_partitioner/partitioner_fwd.hpp b/libs/core/resource_partitioner/include/hpx/resource_partitioner/partitioner_fwd.hpp
@@ -71,5 +71,7 @@ namespace hpx { namespace resource {
         abp_priority_fifo = 5,
         abp_priority_lifo = 6,
         shared_priority = 7,
+        local_workrequesting_fifo = 8,
+        local_workrequesting_lifo = 9,
     };
 }}    // namespace hpx::resource
diff --git a/libs/core/resource_partitioner/src/detail_partitioner.cpp b/libs/core/resource_partitioner/src/detail_partitioner.cpp
@@ -134,6 +134,12 @@ namespace hpx { namespace resource { namespace detail {
         case resource::local_priority_lifo:
             sched = "local_priority_lifo";
             break;
+        case resource::local_workrequesting_fifo:
+            sched = "local_workrequesting_fifo";
+            break;
+        case resource::local_workrequesting_lifo:
+            sched = "local_workrequesting_lifo";
+            break;
         case resource::static_:
             sched = "static";
             break;
@@ -451,6 +457,18 @@ namespace hpx { namespace resource { namespace detail {
         {
             default_scheduler = scheduling_policy::local_priority_lifo;
         }
+        else if (0 ==
+            std::string("local-workrequesting-fifo")
+                .find(default_scheduler_str))
+        {
+            default_scheduler = scheduling_policy::local_workrequesting_fifo;
+        }
+        else if (0 ==
+            std::string("local-workrequesting-lifo")
+                .find(default_scheduler_str))
+        {
+            default_scheduler = scheduling_policy::local_workrequesting_lifo;
+        }
         else if (0 == std::string("static").find(default_scheduler_str))
         {
             default_scheduler = scheduling_policy::static_;

diff --git a/libs/core/runtime_configuration/src/runtime_configuration.cpp b/libs/core/runtime_configuration/src/runtime_configuration.cpp
@@ -156,7 +156,7 @@ namespace hpx { namespace util {
             "localities = 1",
             "first_pu = 0",
             "runtime_mode = console",
-            "scheduler = local-priority-fifo",
+            "scheduler = local-workrequesting-fifo",
             "affinity = core",
             "pu_step = 1",
             "pu_offset = 0",