Fix/minor

.github/SETUP.md

@@ -1,10 +1,9 @@
 # Expected github runners setup
 
-## Job configuration
-A worflow job may be configured to run within a docker machine or directly in the host environment, in the latter case be aware that Bambu distributions coming from other jobs as artifacts may need the _APPDIR_ variable to be set to the new install location.
+Github runners may be configured to run within a docker container or directly in the host environment. In the former case there may be some issues during the execution of external vendor tools which may not support a containerized environment.
 
 ## Runner labels
-Runners may expose many different labels based on what tools are available on the host machine.
+Runners may expose many different labels based on what tools are available on the host machine. Comments about docker volumes are intender for a containerized environment only.
 
 - **altera**: Altera synthesis tools are available on the host machine and a docker volume named altera-tools exposes Altera tools install directories (e.g. Quartus, QuestaSim, ...) 
 - **intel**: Intel synthesis tools are available on the host machine and a docker volume named intel-tools exposes modern Intel FPGA tools install directories (e.g. Quartus Prime, QuestaSim, ...)
@@ -13,20 +12,28 @@ Runners may expose many different labels based on what tools are available on th
 - **nanoxplore**: NanoXplore synthesis tools are available on the host machine and a docker volume named nanoxplore-tools exposes NanoXplore tools install directories
 - **xilinx**: Xilinx synthesis tools are available on the host machine and a docker volume named xilinx-tools exposes Xilinx tools install directories (e.g. Vivado, Vitis HLS, ...)
 
-Directories containing license files should be copied or linked in container user home. Volumes may be defined as read-only to avoid issues.
+If working with docker containers, directories containing license files should be copied or linked in container user home. Volumes may be defined as read-only to avoid issues.
 
-A **licenses-home** volume is always expected when at least one of the above is defined. It should contain all necessary license files for available tools. Workflow job is expected to copy the volume content into the home directory of current user when running a container.
-
-To create such volumes the following may be useful:
-
-```
-docker volume create --driver local --opt o=bind,ro --opt type=none --opt device=/path/to/dir vendor-tools
-```
-
-## Runner environment
+## Environment variables
 Some environment variables are expected to be set by each runner host:
 
 - **J**: number of maximum parallel jobs handled by the runner
 - **LM_LICENSE_FILE**: license file path for simulation/synthesis tools
+- **NXLMD_LICENSE_FILE**: NanoXplore license file path (needed only if different from LM_LICENSE_FILE)
 - **NANOXPLORE_BYPASS**: NanoXplore bypass setting
 - **LIBRARY_PATH**: necessary to support older gcc compilers (set to: /usr/lib/x86_64-linux-gnu)
+
+## Python support
+Current CI implementation requires Python 3.6.15 to be available in the runner environment. Pyenv is recommended to provide the support.
+Furthermore pip packages from `etc/scripts/requirements.txt` are required to run python scripts correctly.
+
+Use the following to install the required Python version through PyEnv and set it as global default.
+
+```
+CONFIGURE_OPTS="--enable-shared" pyenv install 3.6.15
+pyenv global 3.6.15
+pip install -r /path/to/repo/et/scripts/requirements.txt
+```
+
+Note that Github Runners are launched as systemd services, thus `~/.bashrc` or `~/.profile` are not loaded.
+A `.path` file should be added in the runner directory containing standard _PATH_ variable prepended with PyEnv shims and bin paths.

.github/actions/generate-appimage/entrypoint.sh

@@ -148,6 +148,8 @@ EOF
 cat > $dist_dir/usr/bin/tool_select.sh << EOF
 #!/bin/bash
 export LC_ALL="C"
+unset PYTHONHOME # Python is not bundled with this AppImage
+unset PYTHONPATH
 BINARY_NAME=\$(basename "\$ARGV0")
 BINARY_PATH="\$APPDIR/usr/bin/\$BINARY_NAME"
 if [ "\$BINARY_NAME" == "debug_terminal" ]; then

.github/actions/perf-check/action.yml

@@ -29,10 +29,7 @@ inputs:
 outputs:
  compare-csv:
  description: "Compare output csv file"
- value: ${{ steps.cmp.outputs.out-csv }}
- is-better:
- description: "True if new results are better than main, false else"
- value: ${{ steps.cmp.outputs.better }}
+ value: ${{ steps.cmp.outputs.outcsv }}
 runs:
  using: "composite"
  steps:
@@ -52,42 +49,33 @@ runs:
  echo "Path '${{inputs.new-results}}' does not name a file or directory"
  exit -1;
  fi
- - name: Download test results from main
- id: download
+ - name: Compare against main
+ id: cmp
  shell: bash
  run: |
  set -e
  PKG_URL="${{inputs.storage-url}}/results/main/${{inputs.test-id}}.csv"
+ COMMIT_SHA="${{github.sha}}"
+ COMPARE_CSV="${{inputs.test-id}}.$(echo '${{github.ref}}' | sed 's/\//_/g')_${COMMIT_SHA:0:7}.csv"
+ score=""
+ outcsv=""
  CURL_CMD=""
  if [[ ! -z '${{inputs.username}}' ]]; then
  CURL_CMD+='-u ${{inputs.username}}:${{inputs.password}} '
  fi
  mkdir -p main-reports
- curl -s -L --fail $CURL_CMD $PKG_URL -o main-reports/${{inputs.test-id}}.csv
- if [[ $? -eq 0 ]]; then
- echo "::set-output name=found::1"
+ if curl -s -L --fail $CURL_CMD $PKG_URL -o main-reports/${{inputs.test-id}}.csv; then
+ if [[ ! -z "${{ inputs.score }}" ]]; then
+ score="-s ${{ inputs.score }}"
+ fi
+ worse_count="0"
+ python ./etc/scripts/perf_diff.py -d ${{inputs.datapoints}} $score -o $COMPARE_CSV --returnfail "main-reports/${{inputs.test-id}}.csv" "${{inputs.new-results}}" || worse_count=$?
+ if [[ "$worse_count" != "0" ]]; then
+ outcsv="$COMPARE_CSV"
+ echo "::warning:: Performance are worse than main for ${{inputs.test-id}}: $worse_count benchmarks have degraded performance"
+ fi
  else
  echo "::warning:: No previous benchmark result found to compare"
  fi
- exit 0
- - name: Compare performance results
- if: ${{steps.download.outputs.found}}
- id: cmp
- shell: bash
- run: |
- COMMIT_SHA="${{github.sha}}"
- COMPARE_CSV="${{inputs.test-id}}.$(echo '${{github.ref}}' | sed 's/\//_/g')_${COMMIT_SHA:0:7}.csv"
- score=""
- if [[ ! -z "${{ inputs.score }}" ]]; then
- score="-s ${{ inputs.score }}"
- fi
- worse_count="0"
- ./etc/scripts/perf_diff.py -d ${{inputs.datapoints}} $score -o $COMPARE_CSV --returnfail "main-reports/${{inputs.test-id}}.csv" "${{inputs.new-results}}" || worse_count=$?
- if [[ "$worse_count" != "0" ]]; then
- echo "::set-output name=out-csv::$COMPARE_CSV"
- echo "::set-output name=better::0"
- echo "::warning:: Performance are worse than main for ${{inputs.test-id}}: $worse_count benchmarks have degraded performance"
- else
- echo "::set-output name=better::1"
- fi
+ echo "::set-output name=outcsv::$outcsv"
  exit 0

.github/workflows/minimal.yml

@@ -96,7 +96,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -173,7 +173,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -228,7 +228,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -283,7 +283,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -395,7 +395,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -487,7 +487,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -541,7 +541,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -714,7 +714,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}
@@ -739,6 +739,8 @@ jobs:
  datapoints: PERIOD,AREA,REGISTERS,DSPS,BRAMS
  - test: lattice
  datapoints: CYCLES,PERIOD,AREA,REGISTERS,DSPS,BRAMS
+ - test: nanoxplore
+ datapoints: CYCLES,PERIOD,AREA,REGISTERS,DSPS,BRAMS
  - test: sim
  datapoints: CYCLES
  - test: xilinx_ise
@@ -784,7 +786,7 @@ jobs:
  username: "${{env.storage-user}}"
  password: "${{env.storage-token}}"
  - name: Add performance results
- if: ${{ steps.perf-check.outputs.is-better == 0 }}
+ if: ${{ steps.perf-check.outputs.compare-csv != '' }}
  shell: bash
  run: |
  mv ${{ steps.perf-check.outputs.compare-csv }} ${{env.report-dir}}

documentation/tutorial_date_2022/03-optimizations/Exercise1/solution/synthesize.sh

@@ -1,7 +1,7 @@
 #!/bin/bash
 abs_script=$(readlink -e $0)
 dir_script=$(dirname $abs_script)
-$dir_script/../../test_panda.py --tool=bambu --bambu=bambu --spider=spider \
+python3 $dir_script/../../test_panda.py --tool=bambu --bambu=bambu --spider=spider \
  --args="--configuration-name=GCC49 --compiler=I386_GCC49" \
  --args="--configuration-name=GCC7 --compiler=I386_GCC7" \
  --args="--configuration-name=CLANG6 --compiler=I386_CLANG6" \

documentation/tutorial_date_2022/03-optimizations/Exercise3/solution/synthesize.sh

@@ -1,6 +1,6 @@
 #!/bin/bash
 abs_script=$(readlink -e $0)
 dir_script=$(dirname $abs_script)
-$dir_script/../../test_panda.py --tool=bambu --bambu=bambu --spider=spider \
+python3 $dir_script/../../test_panda.py --tool=bambu --bambu=bambu --spider=spider \
  --args="--configuration-name=GCC49 --compiler=I386_GCC49" \
  -c=--simulate -b$dir_script -l$dir_script/list "$@"

documentation/tutorial_date_2022/03-optimizations/Exercise4/solution/synthesize.sh

@@ -1,7 +1,7 @@
 #!/bin/bash
 abs_script=$(readlink -e $0)
 dir_script=$(dirname $abs_script)
-$dir_script/../../test_panda.py --tool=bambu --bambu=bambu --spider=spider \
+python3 $dir_script/../../test_panda.py --tool=bambu --bambu=bambu --spider=spider \
  --args="--configuration-name=pow_square " \
  --args="--configuration-name=mult_square -DMULT_SQUARE" \
  --args="--configuration-name=single_pow_square -DFP_SINGLE" \

documentation/tutorial_date_2022/03-optimizations/test_panda.py

@@ -1,4 +1,4 @@
-#!/usr/bin/python
+#!python
 
 import argparse
 import datetime

documentation/tutorial_fpl_2017/lab2/solutions/2.7/synthesize.sh

@@ -2,7 +2,7 @@
 abs_script=$(readlink -e $0)
 dir_script=$(dirname $abs_script)
 rm -rf output
-$dir_script/test_panda.py --tool=bambu \
+python3 $dir_script/test_panda.py --tool=bambu \
  --args="--configuration-name=pow --top-fname=formula_pow" \
  --args="--configuration-name=mult --top-fname=formula_mult " \
  --args="--configuration-name=double_pow --top-fname=double_formula_pow" \

documentation/tutorial_fpl_2017/lab2/solutions/2.7/test_panda.py

@@ -1,4 +1,4 @@
-#!/usr/bin/python3
+#!python3
 
 import argparse
 import datetime

documentation/tutorial_fpt_2017/optimizations/fourth/solution/synthesize.sh

@@ -2,7 +2,7 @@
 abs_script=$(readlink -e $0)
 dir_script=$(dirname $abs_script)
 rm -rf output
-$dir_script/test_panda.py --tool=bambu \
+python3 $dir_script/test_panda.py --tool=bambu \
  --args="--configuration-name=pow --top-fname=formula_pow" \
  --args="--configuration-name=mult --top-fname=formula_mult " \
  --args="--configuration-name=double_pow --top-fname=double_formula_pow" \

documentation/tutorial_fpt_2017/optimizations/fourth/solution/test_panda.py

@@ -1,4 +1,4 @@
-#!/usr/bin/python3
+#!python3
 
 import argparse
 import datetime

documentation/tutorial_pact_2019/Optimization/fourth/solution/synthesize.sh

@@ -2,7 +2,7 @@
 abs_script=$(readlink -e $0)
 dir_script=$(dirname $abs_script)
 rm -rf output
-$dir_script/test_panda.py --tool=bambu --bambu=/opt/panda/bin/bambu --spider=/opt/panda/bin/spider \
+python3 $dir_script/test_panda.py --tool=bambu --bambu=/opt/panda/bin/bambu --spider=/opt/panda/bin/spider \
  --args="--configuration-name=pow --top-fname=formula_pow" \
  --args="--configuration-name=mult --top-fname=formula_mult " \
  --args="--configuration-name=double_pow --top-fname=double_formula_pow" \

documentation/tutorial_pact_2019/Optimization/fourth/solution/test_panda.py

@@ -1,4 +1,4 @@
-#!/usr/bin/python3
+#!python3
 
 import argparse
 import datetime