Make up commit

CMakeLists.txt

@@ -37,7 +37,7 @@ option(MRC_USE_CONDA "Enables finding dependencies via conda. All dependencies m
   environment" ON)
 option(MRC_USE_IWYU "Enable running include-what-you-use as part of the build process" OFF)
 
-set(MRC_RAPIDS_VERSION "23.12" CACHE STRING "Which version of RAPIDS to build for. Sets default versions for RAPIDS CMake and RMM.")
+set(MRC_RAPIDS_VERSION "24.04" CACHE STRING "Which version of RAPIDS to build for. Sets default versions for RAPIDS CMake and RMM.")
 
 set(MRC_CACHE_DIR "${CMAKE_SOURCE_DIR}/.cache" CACHE PATH "Directory to contain all CPM and CCache data")
 mark_as_advanced(MRC_CACHE_DIR)

cmake/Configure_UCXX.cmake

@@ -20,7 +20,7 @@ function(morpheus_utils_configure_UCXX)
   list(APPEND CMAKE_MESSAGE_CONTEXT "UCXX")
 
   morpheus_utils_assert_cpm_initialized()
-  set(UCXX_VERSION "0.37.00" CACHE STRING "Which version of UCXX to use.")
+  set(UCXX_VERSION "0.38.00" CACHE STRING "Which version of UCXX to use.")
 
   find_package(ucx REQUIRED)
 
@@ -34,8 +34,8 @@ function(morpheus_utils_configure_UCXX)
     INSTALL_EXPORT_SET
       ${PROJECT_NAME}-core-exports
     CPM_ARGS
-      GIT_REPOSITORY          https://github.com/pentschev/ucxx.git
-      GIT_TAG                 mrc-all
+      GIT_REPOSITORY          https://github.com/rapidsai/ucxx.git
+      GIT_TAG                 branch-0.38
       GIT_SHALLOW             TRUE
       SOURCE_SUBDIR           cpp
       OPTIONS                 "UCXX_ENABLE_RMM ON"

conda/environments/all_cuda-118_arch-x86_64.yaml

@@ -38,7 +38,7 @@ dependencies:
 - libclang=17
 - libgrpc=1.54.0
 - libhwloc=2.9.2
-- librmm=23.12
+- librmm=24.04
 - llvmdev=17
 - ninja=1.10
 - nlohmann_json=3.9
@@ -53,6 +53,6 @@ dependencies:
 - pytest-timeout
 - python=3.10
 - scikit-build>=0.17
-- ucx=1.14
+- ucx>=1.15
 - yapf
 name: all_cuda-118_arch-x86_64

conda/environments/ci_cuda-118_arch-x86_64.yaml

@@ -30,7 +30,7 @@ dependencies:
 - include-what-you-use=0.21
 - libgrpc=1.54.0
 - libhwloc=2.9.2
-- librmm=23.12
+- librmm=24.04
 - ninja=1.10
 - nlohmann_json=3.9
 - nodejs=18
@@ -43,6 +43,6 @@ dependencies:
 - pytest-timeout
 - python=3.10
 - scikit-build>=0.17
-- ucx=1.14
+- ucx>=1.15
 - yapf
 name: ci_cuda-118_arch-x86_64

cpp/mrc/src/internal/data_plane/data_plane_resources.cpp

@@ -327,6 +327,15 @@ std::shared_ptr<ucxx::Endpoint> DataPlaneResources2::create_endpoint(const ucx::
     m_endpoints_by_address[address] = endpoint;
     m_endpoints_by_id[instance_id]  = endpoint;
 
+    auto close_request = endpoint->close();
+
+    endpoint->cancelInflightRequests();
+
+    while (close_request->isCompleted() == false)
+    {
+        this->progress();
+    }
+
     DVLOG(10) << "Created endpoint with address: " << address;
 
     return endpoint;

cpp/mrc/src/internal/data_plane/data_plane_resources.hpp

@@ -120,7 +120,7 @@ class DataPlaneResources2
 {
   public:
     DataPlaneResources2();
-    ~DataPlaneResources2();
+    virtual ~DataPlaneResources2();
 
     void set_instance_id(uint64_t instance_id);
     bool has_instance_id() const;
@@ -222,6 +222,8 @@ class DataPlaneResources2
     channel::Egress<std::unique_ptr<runtime::RemoteDescriptor2>>& get_inbound_channel() const;
 
   private:
+    virtual std::shared_ptr<runtime::RemoteDescriptorImpl2> get_descriptor(uint64_t object_id);
+
     std::optional<uint64_t> m_instance_id;  // Global ID used to identify this instance
 
     std::shared_ptr<ucxx::Context> m_context;

cpp/mrc/src/tests/test_network.cpp

@@ -39,9 +39,8 @@
 #include "internal/ucx/registration_cache.hpp"
 
 #include "mrc/channel/status.hpp"
-#include "mrc/codable/codable_protocol.hpp"
 #include "mrc/codable/decode.hpp"
-#include "mrc/codable/fundamental_types.hpp"
+#include "mrc/codable/fundamental_types.hpp"  // IWYU pragma: keep
 #include "mrc/codable/type_traits.hpp"
 #include "mrc/edge/edge_builder.hpp"
 #include "mrc/memory/adaptors.hpp"
@@ -82,7 +81,6 @@
 #include <optional>
 #include <ostream>
 #include <stop_token>
-#include <thread>
 #include <utility>
 
 using namespace mrc;
@@ -91,7 +89,7 @@ using namespace mrc::memory::literals;
 class DataPlaneResources2Tester : public data_plane::DataPlaneResources2
 {
   public:
-    std::shared_ptr<runtime::RemoteDescriptorImpl2> get_descriptor(uint64_t object_id)
+    std::shared_ptr<runtime::RemoteDescriptorImpl2> get_descriptor(uint64_t object_id) override
     {
         return m_remote_descriptor_by_id[object_id];
     }

cpp/mrc/tests/test_executor.cpp

@@ -116,6 +116,8 @@ class TestExecutor : public ::testing::Test
                     // #endif
                 }
 
+                VLOG(10) << "Sent all messages";
+
                 s.on_completed();
             });
 

dependencies.yaml

@@ -60,7 +60,7 @@ dependencies:
           - gxx=11.2
           - libgrpc=1.54.0
           - libhwloc=2.9.2
-          - librmm=23.12
+          - librmm=24.04
           - ninja=1.10
           - nlohmann_json=3.9
           - nodejs=18
@@ -69,7 +69,7 @@ dependencies:
           - pybind11-stubgen=0.10
           - python=3.10
           - scikit-build>=0.17
-          - ucx=1.14
+          - ucx>=1.15
 
   checks:
     common:

mrc.code-workspace

@@ -258,6 +258,10 @@
         "python.testing.pytestEnabled": true,
         "python.testing.unittestEnabled": false,
         "testMate.cpp.debug.configTemplate": {
+            "name": "Debug Test Runner (${parentLabel} > ${label})",
+            "request": "launch",
+            "type": "lldb",
+            "program": "${exec}",
             "args": "${argsArray}",
             "cwd": "${cwd}",
             "darwin": {
@@ -266,44 +270,14 @@
             },
             "env": "${envObj}",
             "environment": "${envObjArray}",
-            "linux": {
-                "MIMode": "gdb",
-                "symbolLoadInfo": {
-                    "exceptionList": "*libmrc*.so",
-                    "loadAll": false
-                },
-                "type": "cppdbg"
-            },
-            "program": "${exec}",
-            "setupCommands": [
-                {
-                    "description": "Enable pretty-printing for gdb",
-                    "text": "-enable-pretty-printing"
-                },
-                {
-                    "description": "Skip stdio-common files",
-                    "text": "-interpreter-exec console \"skip -gfi **/bits/*.h\""
-                }
-                // {
-                //     "description": "Stay on same thread when debugging",
-                //     "text": "-interpreter-exec console \"set scheduler-locking step\""
-                // }
-            ],
-            "sourceFileMap": {
-                "${workspaceFolder}": {
-                    "editorPath": "${workspaceFolder}",
-                    "useForBreakpoints": "true"
-                }
-            },
             "testMate.cpp.debug.setEnv": {
                 "GLOG_v": "10",
                 "GTEST_CATCH_EXCEPTIONS": "0", // Allow the debugger to catch exceptions
                 "UCX_ERROR_SIGNALS": "" // Prevent UCX from capturing errors
             },
-            "type": "cppvsdbg",
             "win32": {
                 "type": "cppvsdbg"
-            }
+            },
         },
         "testMate.cpp.log.logpanel": true,
         "testMate.cpp.test.executables": "{build,Build,BUILD,out,Out,OUT}/**/*{test,Test,TEST}_*.x",
@@ -312,6 +286,8 @@
         "typescript.referencesCodeLens.enabled": true,
         "yapf.args": [
             "--style=${workspaceFolder}/python/setup.cfg"
-        ]
+        ],
+        // "lldb.library": "/home/mdemoret/mambaforge/envs/mrc_dev2/lib/liblldb.so",
+        "lldb.launch.terminal": "integrated"
     }
 }