merge changes from develop branch and resolve merge conflicts

.github/workflows/ci.yaml

@@ -19,7 +19,7 @@ jobs:
     # In this step, this action saves a list of existing images,
     # the cache is created without them in the post run.
     # It also restores the cache if it exists.
-    - uses: satackey/action-docker-layer-caching@v0.0.8
+    - uses: satackey/action-docker-layer-caching@v0.0.11
       # Ignore the failure of a step and avoid terminating the job.
       continue-on-error: true
 
@@ -65,7 +65,7 @@ jobs:
     # In this step, this action saves a list of existing images,
     # the cache is created without them in the post run.
     # It also restores the cache if it exists.
-    - uses: satackey/action-docker-layer-caching@v0.0.8
+    - uses: satackey/action-docker-layer-caching@v0.0.11
       # Ignore the failure of a step and avoid terminating the job.
       continue-on-error: true
 
@@ -80,10 +80,10 @@ jobs:
       uses: pat-s/always-upload-cache@v2.1.3
       with:
         path: cppcheck-cache
-        key: cppcheck-cache-${{ steps.cache_timestamp.outputs.timestamp }}
+        key: cppcheck-cache-${{ hashFiles('cppcheck.rules', 'CMakeLists.txt') }}-${{ steps.cache_timestamp.outputs.timestamp }}
         restore-keys: |
-            cppcheck-cache-${{ steps.cache_timestamp.outputs.timestamp }}
-            cppcheck-cache-
+            cppcheck-cache-${{ hashFiles('cppcheck.rules', 'CMakeLists.txt') }}-${{ steps.cache_timestamp.outputs.timestamp }}
+            cppcheck-cache-${{ hashFiles('cppcheck.rules', 'CMakeLists.txt') }}-
 
     - name: Build the Docker image
       run: docker build . --file hip-clang.docker --tag migraphx
@@ -108,7 +108,7 @@ jobs:
     # In this step, this action saves a list of existing images,
     # the cache is created without them in the post run.
     # It also restores the cache if it exists.
-    - uses: satackey/action-docker-layer-caching@v0.0.8
+    - uses: satackey/action-docker-layer-caching@v0.0.11
       # Ignore the failure of a step and avoid terminating the job.
       continue-on-error: true
 
@@ -142,10 +142,12 @@ jobs:
       with:
         python-version: 3.6
     - name: Install pyflakes
-      run: pip install pyflakes==2.3.1
+      run: pip install pyflakes==2.3.1 mypy==0.931
 
     - name: Run pyflakes
-      run: pyflakes examples/ tools/ src/ test/ doc/
+      run: |
+        pyflakes examples/ tools/ src/ test/ doc/
+        mypy tools/api.py
 
   linux:
 
@@ -218,7 +220,7 @@ jobs:
       run: |
         echo "leak:dnnl::impl::malloc" > suppressions.txt
         export LSAN_OPTIONS="suppressions=$(pwd)/suppressions.txt"
-        rbuild build -d cget -s gh -t check \
+        rbuild build -d cget -s gh -T check \
           -DCMAKE_BUILD_TYPE=${{matrix.configuration}} \
           -DMIGRAPHX_ENABLE_PYTHON=${{matrix.configuration == 'release' && 'On' || 'Off'}} \
           -DCMAKE_CXX_FLAGS_DEBUG="-g1 -Os -fdebug-prefix-map=$PWD=. -fdebug-types-section -fno-omit-frame-pointer -fsanitize=undefined -fno-sanitize-recover=undefined" \

CMakeLists.txt

@@ -36,7 +36,7 @@ find_package(nlohmann_json 3.8.0 REQUIRED)
 
 include(ROCMSetupVersion)
 
-rocm_setup_version(VERSION 1.3)
+rocm_setup_version(VERSION 2.0)
 set(MIGRAPHX_SO_VERSION ${PROJECT_VERSION_MAJOR}.${PROJECT_VERSION_MINOR})
 
 option( BUILD_SHARED_LIBS "Build as a shared library" ON )
@@ -190,6 +190,7 @@ rocm_enable_cppcheck(
         shadowVariable
         unsafeClassDivZero
         definePrefix:*test/include/test.hpp
+        ctuOneDefinitionRuleViolation:*test/*
         useSmartPointer:*src/api/api.cpp
         useSmartPointer:*make_shared_array.hpp
         constParameter:*src/targets/gpu/*.cpp
@@ -199,6 +200,7 @@ rocm_enable_cppcheck(
     RULE_FILE
         ${CMAKE_CURRENT_SOURCE_DIR}/cppcheck.rules
     SOURCES
+        examples/
         src/
         test/
     INCLUDE

Dockerfile

@@ -6,7 +6,7 @@ ARG PREFIX=/usr/local
 RUN dpkg --add-architecture i386
 
 # Add rocm repository
-RUN sh -c 'echo deb [arch=amd64 trusted=yes] http://repo.radeon.com/rocm/apt/4.2/ xenial main > /etc/apt/sources.list.d/rocm.list'
+RUN sh -c 'echo deb [arch=amd64 trusted=yes] http://repo.radeon.com/rocm/apt/4.5/ ubuntu main > /etc/apt/sources.list.d/rocm.list'
 
 # Install dependencies
 RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \
@@ -32,6 +32,8 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-
     software-properties-common \
     wget \
     rocm-device-libs \
+    hip-base \
+    libnuma-dev \
     miopen-hip \
     rocblas \
     zlib1g-dev && \

Jenkinsfile

@@ -20,7 +20,7 @@ def rocmtestnode(Map conf) {
             rm -rf build
             mkdir build
             cd build
-            CXX=${compiler} CXXFLAGS='-Werror -Wno-fallback' cmake -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache ${flags} .. 
+            CXX=${compiler} CXXFLAGS='-Werror' cmake -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache ${flags} .. 
             make -j\$(nproc) generate all doc package check VERBOSE=1
         """
         echo cmd
@@ -75,6 +75,8 @@ def rocmnodename(name) {
         node_name = "${rocmtest_name} && fiji";
     } else if(name == "vega") {
         node_name = "${rocmtest_name} && vega";
+    } else if(name == "navi21") {
+        node_name = "${rocmtest_name} && navi21";
     } else if(name == "nogpu") {
         return rocmtest_name;
     }
@@ -110,6 +112,10 @@ rocmtest clang_debug: rocmnode('vega') { cmake_build ->
         def debug_flags = "-g -O2 -fno-omit-frame-pointer -fsanitize=${sanitizers} -fno-sanitize-recover=${sanitizers}"
         cmake_build("/opt/rocm/llvm/bin/clang++", "-DCMAKE_BUILD_TYPE=debug -DMIGRAPHX_ENABLE_PYTHON=Off -DMIGRAPHX_ENABLE_GPU=Off -DMIGRAPHX_ENABLE_CPU=On -DCMAKE_CXX_FLAGS_DEBUG='${debug_flags}'")
     }
+}, clang_release_navi: rocmnode('navi21') { cmake_build ->
+    stage('HIP Clang Release Navi') {
+        cmake_build("/opt/rocm/llvm/bin/clang++", "-DCMAKE_BUILD_TYPE=release")
+    }
 }
 
 def onnxnode(name, body) {

cmake/TargetFlags.cmake

 
+function(eval_and_strip_genex OUTPUT_VAR INPUT)
+    string(REPLACE "$<LINK_LANGUAGE:CXX>" "1" INPUT "${INPUT}")
+    string(REPLACE "$<COMPILE_LANGUAGE:CXX>" "1" INPUT "${INPUT}")
+    string(REPLACE "SHELL:" "" INPUT "${INPUT}")
+    string(REPLACE "$<BOOL:>" "0" INPUT "${INPUT}")
+    string(REGEX REPLACE "\\$<BOOL:(0|FALSE|false|OFF|off|N|n|IGNORE|ignore|NOTFOUND|notfound)>" "0" INPUT "${INPUT}")
+    string(REGEX REPLACE "\\$<BOOL:[^<>]*-NOTFOUND>" "0" INPUT "${INPUT}")
+    string(REGEX REPLACE "\\$<BOOL:[^$<>]*>" "1" INPUT "${INPUT}")
+    string(REPLACE "$<NOT:0>" "1" INPUT "${INPUT}")
+    string(REPLACE "$<NOT:1>" "0" INPUT "${INPUT}")
+    string(REGEX REPLACE "\\$<0:[^<>]*>" "" INPUT "${INPUT}")
+    string(REGEX REPLACE "\\$<1:([^<>]*)>" "\\1" INPUT "${INPUT}")
+    string(GENEX_STRIP "${INPUT}" INPUT)
+    set(${OUTPUT_VAR} "${INPUT}" PARENT_SCOPE)
+endfunction()
+
 function(get_target_property2 VAR TARGET PROPERTY)
     get_target_property(_pflags ${TARGET} ${PROPERTY})
     if(_pflags)
+        eval_and_strip_genex(_pflags "${_pflags}")
         set(${VAR} ${_pflags} PARENT_SCOPE)
     else()
         set(${VAR} "" PARENT_SCOPE)
     endif()
 endfunction()
 
+function(flags_requires_arg OUTPUT_VAR FLAG)
+    set(_args -x -isystem)
+    if(FLAG IN_LIST _args)
+        set(${OUTPUT_VAR} 1 PARENT_SCOPE)
+    else()
+        set(${OUTPUT_VAR} 0 PARENT_SCOPE)
+    endif()
+endfunction()
 
 macro(append_flags FLAGS TARGET PROPERTY PREFIX)
     get_target_property2(_pflags ${TARGET} ${PROPERTY})
+    set(_requires_arg 0)
     foreach(FLAG ${_pflags})
-        if(TARGET ${FLAG})
+        string(STRIP "${FLAG}" FLAG)
+        if(FLAG)
+            if(TARGET ${FLAG} AND NOT _requires_arg)
                 target_flags(_pflags2 ${FLAG})
                 string(APPEND ${FLAGS} " ${_pflags2}")
             else()
                 string(APPEND ${FLAGS} " ${PREFIX}${FLAG}")
             endif()
+            flags_requires_arg(_requires_arg "${FLAG}")
+        endif()
     endforeach()
 endmacro()
 
 macro(append_link_flags FLAGS TARGET PROPERTY)
     get_target_property2(_pflags ${TARGET} ${PROPERTY})
+    set(_requires_arg 0)
     foreach(FLAG ${_pflags})
-        if(TARGET ${FLAG})
+        string(STRIP "${FLAG}" FLAG)
+        if(FLAG)
+            if(TARGET ${FLAG} AND NOT _requires_arg)
                 target_flags(_pflags2 ${FLAG})
                 string(APPEND ${FLAGS} " ${_pflags2}")
             elseif(FLAG MATCHES "^-.*")
@@ -34,6 +67,8 @@ macro(append_link_flags FLAGS TARGET PROPERTY)
             else()
                 string(APPEND ${FLAGS} " -l${FLAG}")
             endif()
+            flags_requires_arg(_requires_arg "${FLAG}")
+        endif()
     endforeach()
 endmacro()
 

cppcheck.rules

@@ -152,6 +152,24 @@
         <summary>Else statement is not necessary.</summary>
     </message>
 </rule>
+<rule>
+    <tokenlist>normal</tokenlist>
+    <pattern><![CDATA[((?:(?:\w+|<|>|::) )*(?:\w+|>)(?: &|\*)*) (\w+) ; \2 = static_cast < \1 > (\([^()]*(?-1)*[^()]*\)) ;]]></pattern>
+    <message>
+        <id>RedundantCast</id>
+        <severity>style</severity>
+        <summary>Static cast is redundant.</summary>
+    </message>
+</rule>
+<rule>
+    <tokenlist>normal</tokenlist>
+    <pattern><![CDATA[auto (\w+) ; \1 = static_cast < (?:(?:\w+|<|>|::) )*(?:\w+|>)(?: &|\*)* > (\([^()]*(?-1)*[^()]*\)) ;]]></pattern>
+    <message>
+        <id>RedundantCast</id>
+        <severity>style</severity>
+        <summary>Static cast is redundant.</summary>
+    </message>
+</rule>
 <rule>
     <tokenlist>normal</tokenlist>
     <pattern><![CDATA[\? (true|false) : (true|false)]]></pattern>

dev-requirements.txt

 pfultz2/rocm-recipes
 facebook/zstd@v1.4.5 -X subdir -DCMAKE_DIR=build/cmake
 ccache@v4.1
+pcre,pfultz2/pcre@8.45 -H sha256:d6f7182602a775a7d500a0cedca6449af0400c6493951513046d17615ed0bf11
 danmar/cppcheck@2.6 -DHAVE_RULES=1
 RadeonOpenCompute/rocm-cmake@ececd2eccae4d01e7ec154efe90ac43ebf4df317 --build
 -f requirements.txt

doc/requirements.txt

-sphinx==2.2.2
-breathe==4.13.1
+docutils==0.17.1
+sphinx==4.2.0
+breathe==4.31.0
+sphinx_rtd_theme==1.0.0
 # git+https://github.com/arximboldi/breathe@fix-node-parent

doc/src/conf.py

@@ -18,6 +18,8 @@
 #
 # import os
 # import sys
+from datetime import date
+import re
 # sys.path.insert(0, os.path.abspath('.'))
 
 # -- General configuration ------------------------------------------------
@@ -29,7 +31,11 @@
 # Add any Sphinx extension module names here, as strings. They can be
 # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
 # ones.
-extensions = ['breathe', 'sphinx.ext.mathjax', 'sphinx.ext.viewcode']
+extensions = [
+    'breathe', 'sphinx.ext.mathjax', 'sphinx.ext.viewcode', 'sphinx_rtd_theme',
+    'sphinx.ext.autosectionlabel'
+]
+autosectionlabel_prefix_document = True
 
 # Add any paths that contain templates here, relative to this directory.
 templates_path = ['_templates']
@@ -45,7 +51,7 @@ master_doc = 'index'
 
 # General information about the project.
 project = u'MIGraphX'
-copyright = u'2018, AMD'
+copyright = u'2018-{}, AMD'.format(date.today().year)
 author = u'AMD'
 
 # The version info for the project you're documenting, acts as replacement for
@@ -53,9 +59,12 @@ author = u'AMD'
 # built documents.
 #
 # The short X.Y version.
-version = u'0.1'
+with open('../../CMakeLists.txt') as file:
+    version = next((re.findall('[0-9.]+', line)[0]
+                    for line in file.readlines()
+                    if 'rocm_setup_version' in line))
 # The full version, including alpha/beta/rc tags.
-release = u'0.1'
+release = version
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.
@@ -82,7 +91,7 @@ todo_include_todos = False
 # The theme to use for HTML and HTML Help pages.  See the documentation for
 # a list of builtin themes.
 #
-html_theme = 'alabaster'
+html_theme = 'sphinx_rtd_theme'
 
 # Theme options are theme-specific and customize the look and feel of a theme
 # further.  For a list of options available for each theme, see the

doc/src/dev/tools.rst

+Tools
+=====
+
+roctx.py
+--------
+MIGraphX driver provides `roctx` command which can be used with `rocprof` binary to get marker timing information for each MIGraphX operator.  
+In order to help user to process timing information, rocTX helper script is provided at `tools/roctx.py`.
+The `roctx.py` helper script provides two main functionality: `run` and `parse`. Available knobs and usage are given below:
+
+::
+
+    Usage: roctx.py [-h] [--json-path json_path] [--out out]
+    [--study-name study-name] [--repeat repeat] [--parse]
+    [--run run] [--debug]
+
+.. option::  --run
+
+Runs `migraphx-driver roctx` command and given `migraphx-driver` knobs, and then parses the results, providing GPU kernel timing information.
+MIGraphX knobs can be given via a string to `--run` knob. Please see the examples below.
+
+.. option::  --parse
+
+Given `--json-path`, parses JSON file and provides GPU kernel timing information.
+
+.. option::  --out
+
+Output folder
+
+.. option::  --study-name
+
+Optional. Allows user to name a study for easier interpretation. Defaults to timestamp.
+
+.. option::  --repeat
+
+Number of iterations. Set to **2** by default.
+
+.. option::  --debug
+
+Provides additional debug information related to data. Only use for debugging purposes.
+
+**Examples:**
+
+**Running inference with rocTX for a given ONNX file:**
+::
+    python roctx.py --run '--onnx --gpu fcn-resnet50-11.onnx' --out output_folder --repeat 5
+
+After a run, similar to output given below is expected at terminal. The output will provide `SUM`, `MIN`, `MAX` and `COUNT` information for each kernel executed for a given model.
+Average total time is also provided. There are three files provided for reference:
+
+1. `OUTPUT CSV FILE` provides a summary of the run, providing utilized MIGraphX knobs and related kernel timing information
+2. `KERNEL TIMING DETAILS` provides the hotspot kernel timing information
+3. This will provide all output data related to all iterations executed during a run.
+
+An example output:
+
+.. image:: ./roctx1.jpg
+
+Hotspot kerel timing information:
+
+.. image:: ./roctx2.jpg
+
+**Parsing an already existing JSON file:**
+::
+
+    python roctx.py --parse --json-path ../trace.json
\ No newline at end of file

doc/src/developer_guide.rst

@@ -13,3 +13,4 @@ Developer Guide
    dev/quantization
    dev/pass
    dev/matchers
+   dev/tools

doc/src/driver.rst

@@ -61,3 +61,21 @@ Verify each instruction
 .. option::  -r, --reduce
 
 Reduce program and verify
+
+roctx
+----
+
+.. program:: migraphx-driver roctx
+
+Provides marker information for each operation, allowing MIGraphX to be used with `rocprof <https://rocmdocs.amd.com/en/latest/ROCm_Tools/ROCm-Tools.html>`_ for performance analysis.
+This allows user to get GPU-level kernel timing information.
+An example command line combined with rocprof for tracing purposes is given below:
+
+.. code-block:: bash
+
+    /opt/rocm/bin/rocprof --hip-trace --roctx-trace --flush-rate 1ms --timestamp on -d <OUTPUT_PATH> --obj-tracking on /opt/rocm/bin/migraphx-driver roctx <ONNX_FILE> <MIGRAPHX_OPTIONS>
+
+After `rocprof` is run, the output directory will contain trace information for HIP, HCC and ROCTX in seperate `.txt` files.
+To understand the interactions between API calls, it is recommended to utilize `roctx.py` helper script as desribed in :ref:`dev/tools:rocTX` section. 
+
+.. include:: ./driver/compile.rst
\ No newline at end of file

examples/migraphx/cpp_parse_load_save/parse_load_save.cpp

@@ -24,7 +24,6 @@ int main(int argc, char** argv)
         return 0;
     }
 
-    char* parse_arg        = getCmdOption(argv + 2, argv + argc, "--parse");
     char* load_arg         = getCmdOption(argv + 2, argv + argc, "--load");
     char* save_arg         = getCmdOption(argv + 2, argv + argc, "--save");
     const char* input_file = argv[1];

examples/nlp/python_bert_squad/requirements_bertsquad.txt

-tensorflow==2.5.1
+tensorflow==2.5.2
 onnxruntime
 tokenizers
\ No newline at end of file

examples/vision/python_3dunet/3dunet_inference.ipynb

@@ -10,6 +10,16 @@
     "https://github.com/naomifridman/Unet_Brain_tumor_segmentation"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "09ceec31",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!pip install SimpleITK matplotlib scikit-image"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,

examples/vision/python_nfnet/nfnet_inference.ipynb

@@ -17,7 +17,9 @@
     "- How to optimize NFNet ONNX model with AMD MIGraphX.\n",
     "- How to run inference on AMD GPU with the optimized ONNX model.\n",
     "\n",
-    "The NFNet utilized in this example is the smallest NFNet version, F0: 71.5M parameters (83.6% top-1 accuracy on ImageNet)"
+    "The NFNet utilized in this example is the smallest NFNet version, F0: 71.5M parameters (83.6% top-1 accuracy on ImageNet)\n",
+    "\n",
+    "Please make sure MIGraphX Python API is installed following the instructions at Github page."
    ]
   },
   {
@@ -107,7 +109,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "with open('../python_api_inference/imagenet_simple_labels.json') as json_data:\n",
+    "with open('../python_resnet50/imagenet_simple_labels.json') as json_data:\n",
     "    labels = json.load(json_data)"
    ]
   },

examples/vision/python_nfnet/requirements_nfnet.txt

 opencv-python
 onnxruntime
+image
\ No newline at end of file

hip-clang.docker

@@ -6,7 +6,7 @@ ARG PREFIX=/usr/local
 RUN dpkg --add-architecture i386
 
 # Add rocm repository
-RUN sh -c 'echo deb [arch=amd64 trusted=yes] http://repo.radeon.com/rocm/apt/4.2/ xenial main > /etc/apt/sources.list.d/rocm.list'
+RUN sh -c 'echo deb [arch=amd64 trusted=yes] http://repo.radeon.com/rocm/apt/4.5/ ubuntu main > /etc/apt/sources.list.d/rocm.list'
 
 # Install dependencies
 RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \
@@ -29,6 +29,8 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-
     software-properties-common \
     wget \
     rocm-device-libs \
+    hip-base \
+    libnuma-dev \
     miopen-hip \
     rocblas \
     zlib1g-dev && \