1469 Output LB statistics as JSON

ci/docker/ubuntu-18.04-nvidia-cpp.dockerfile

@@ -42,7 +42,19 @@ RUN if test ${compiler} = "nvcc-10"; then \
       rm -rf /var/lib/apt/lists/* && \
       rm -rf cuda-repo-ubuntu1804-10-1-local-10.1.243-418.87.00_1.0-1_amd64.deb && \
       ln -s /usr/local/cuda-10.1 /usr/local/cuda-versioned; \
-    else \
+   elif test ${compiler} = "nvcc-10.2"; then \
+     wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin && \
+     mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600 && \
+     wget https://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_amd64.deb && \
+     dpkg -i cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_amd64.deb && \
+     apt-key add /var/cuda-repo-10-2-local-10.2.89-440.33.01/7fa2af80.pub && \
+     apt-get update && \
+     apt-get -y install cuda && \
+     apt-get clean && \
+     rm -rf /var/lib/apt/lists/* && \
+     rm -rf cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_amd64.deb && \
+     ln -s /usr/local/cuda-10.2 /usr/local/cuda-versioned; \
+   else \
       wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin && \
       mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600 && \
       wget http://developer.download.nvidia.com/compute/cuda/11.0.1/local_installers/cuda-repo-ubuntu1804-11-0-local_11.0.1-450.36.06-1_amd64.deb && \

cmake/link_vt.cmake

@@ -32,6 +32,8 @@ function(link_target_with_vt)
     LINK_DL
     LINK_ZOLTAN
     LINK_FORT
+    LINK_JSON
+    LINK_BROTLI
   )
   set(
     multiValueArg
@@ -49,6 +51,18 @@ function(link_target_with_vt)
     message(STATUS "link_target_with_vt: default link=${ARG_DEFAULT_LINK_SET}")
   endif()
 
+  if (NOT DEFINED ARG_LINK_JSON AND ${ARG_DEFAULT_LINK_SET} OR ARG_LINK_JSON)
+    target_link_libraries(
+      ${ARG_TARGET} PRIVATE ${ARG_BUILD_TYPE} ${JSON_LIBRARY}
+    )
+  endif()
+
+  if (NOT DEFINED ARG_LINK_BROTLI AND ${ARG_DEFAULT_LINK_SET} OR ARG_LINK_BROTLI)
+    target_link_libraries(
+      ${ARG_TARGET} PRIVATE ${ARG_BUILD_TYPE} ${BROTLI_LIBRARY}
+    )
+  endif()
+
   if (NOT DEFINED ARG_LINK_FORT AND ${ARG_DEFAULT_LINK_SET} OR ARG_LINK_FORT)
     if (vt_libfort_enabled)
       target_link_libraries(

cmake/load_bundled_libraries.cmake

@@ -23,6 +23,20 @@ add_subdirectory(${PROJECT_LIB_DIR}/CLI)
 set(FMT_LIBRARY fmt)
 add_subdirectory(${PROJECT_LIB_DIR}/fmt)
 
+# json library always included in the build
+set(JSON_BuildTests OFF)
+set(JSON_MultipleHeaders ON)
+set(JSON_LIBRARY nlohmann_json)
+add_subdirectory(${PROJECT_LIB_DIR}/json)
+
+# brotli library always included in the build
+set(BROTLI_DISABLE_TESTS ON)
+# we need to disable bundled mode so it will install properly
+set(BROTLI_BUNDLED_MODE OFF)
+set(BROTLI_BUILD_PORTABLE ON)
+set(BROTLI_LIBRARY brotlicommon brotlienc brotlidec)
+add_subdirectory(${PROJECT_LIB_DIR}/brotli)
+
 # Optionally include mimalloc (alternative memory allocator)
 if (vt_mimalloc_enabled)
   add_subdirectory(${PROJECT_LIB_DIR}/mimalloc)

docker-compose.yml

@@ -14,7 +14,7 @@
 #             clang-3.9, clang-4.0, clang-5.0, clang-6.0, clang-7, clang-8,
 #             clang-9, clang-10,
 #             icc-18, icc-19,
-#             nvcc-10, nvcc-11}
+#             nvcc-10, nvcc-10.2, nvcc-11}
 #   REPO=lifflander1/vt
 #   UBUNTU={18.04, 20.04}
 #   ULIMIT_CORE=0
@@ -87,6 +87,7 @@ volumes:
   amd64-ubuntu-18.04-icc-19-cache:
   amd64-ubuntu-18.04-icc-20-cache:
   amd64-ubuntu-18.04-nvcc-10-cache:
+  amd64-ubuntu-18.04-nvcc-10.2-cache:
   amd64-ubuntu-18.04-nvcc-11-cache:
   amd64-alpine-clang-3.9-cache:
   amd64-alpine-clang-4.0-cache:
@@ -106,6 +107,7 @@ volumes:
   amd64-alpine-icc-19-cache:
   amd64-alpine-icc-20-cache:
   amd64-alpine-nvcc-10-cache:
+  amd64-alpine-nvcc-10.2-cache:
   amd64-alpine-nvcc-11-cache:
   arm64v8-ubuntu-18.04-gcc-7-cache:
   arm64v8-alpine-gcc-7-cache:

docs/md/node-stats.md

@@ -29,23 +29,161 @@ the statistics and mapping.
 
 \subsection stats-file-format File Format
 
-Each line in the file will one of two formats. The first line is a computation
-time line for each phase, that breaks time down into subphases:
+The VOM files are output in JSON format, either compressed with brotli
+compression (default on) or pure JSON if the argument `--vt_lb_stats_compress`
+is set to `false`.
 
-\code
-<phase>, <object-id>, <time-in-seconds> <#-of-subphases> '[' [<subphase-time-1>] ... [<subphase-time-N>] ']'
+The JSON files contain an array of `phases` that have been captured by \vt and
+output to the file. Each phase has an `id` indicating which phase it was while
+the application was running. Each phase also has an array of `tasks` that
+represent work that was done during that phase. Each task has a `time`,
+`resource`, `node`, `entity`, and optionally a list of `subphases`. The `entity`
+contains information about the task that performed this work. If that `entity`
+is a virtual collection object, it will specify the unique `id` for the object,
+and optionally the `index`, `home`, and `collection_id` for that object.
+
+\code{.json}
+{
+    "phases": [
+        {
+            "id": 0,
+            "tasks": [
+                {
+                    "entity": {
+                        "collection_id": 7,
+                        "home": 0,
+                        "id": 12884901888,
+                        "index": [
+                            3
+                        ],
+                        "type": "object"
+                    },
+                    "node": 0,
+                    "resource": "cpu",
+                    "subphases": [
+                        {
+                            "id": 0,
+                            "time": 0.014743804931640625
+                        }
+                    ],
+                    "time": 0.014743804931640625
+                },
+                {
+                    "entity": {
+                        "collection_id": 7,
+                        "home": 0,
+                        "id": 4294967296,
+                        "index": [
+                            1
+                        ],
+                        "type": "object"
+                    },
+                    "node": 0,
+                    "resource": "cpu",
+                    "subphases": [
+                        {
+                            "id": 0,
+                            "time": 0.013672113418579102
+                        }
+                    ],
+                    "time": 0.013672113418579102
+                }
+            ]
+        },
+        {
+            "id": 1,
+            "tasks": [
+                {
+                    "entity": {
+                        "collection_id": 7,
+                        "home": 0,
+                        "id": 12884901888,
+                        "index": [
+                            3
+                        ],
+                        "type": "object"
+                    },
+                    "node": 0,
+                    "resource": "cpu",
+                    "subphases": [
+                        {
+                            "id": 0,
+                            "time": 0.014104127883911133
+                        }
+                    ],
+                    "time": 0.014104127883911133
+                }
+            ]
+        }
+    ]
+}
 \endcode
 
-The second line format is a communication line:
+Each phase in the file may also have a `communications` array that specify any
+communication between tasks that occurred during the phase. Each communication
+has `type`, which is described below in the following table. Additionally, it
+specifies the `bytes`, number of `messages`, and the two entities that were
+involved in the operator as `to` and `from`. The entities may be of different
+types, like an `object` or `node` depending on the type of communication.
 
-\code
-<phase>, <object-id1-to/recv>, <object-id2-from/send>, <num-bytes>, <comm-type={1..6}>
+\code{.json}
+{
+    "phases": [
+        {
+            "communications": [
+                {
+                    "bytes": 262.0,
+                    "from": {
+                        "home": 1,
+                        "id": 1,
+                        "type": "object"
+                    },
+                    "messages": 1,
+                    "to": {
+                        "home": 0,
+                        "id": 4294967296,
+                        "type": "object"
+                    },
+                    "type": "SendRecv"
+                },
+                {
+                    "bytes": 96.0,
+                    "from": {
+                        "home": 0,
+                        "id": 4294967296,
+                        "type": "object"
+                    },
+                    "messages": 1,
+                    "to": {
+                        "id": 1,
+                        "type": "node"
+                    },
+                    "type": "CollectionToNode"
+                },
+                {
+                    "bytes": 259.0,
+                    "from": {
+                        "id": 0,
+                        "type": "node"
+                    },
+                    "messages": 1,
+                    "to": {
+                        "home": 0,
+                        "id": 0,
+                        "type": "object"
+                    },
+                    "type": "NodeToCollection"
+                }
+            ],
+            "id": 0
+        }
+    ]
+}
 \endcode
 
 
-Where `<comm-type>` is the type of communication occurred. The type of
-communication lines up the enum `vt::vrt::collection::balance::CommCategory` in
-the code.
+The type of communication lines up with the enum
+`vt::vrt::collection::balance::CommCategory` in the code.
 
 | Value | Enum entry | Description |
 | ----- | ---------- | ----------- |