Remove .travis, add workflow

.github/workflows/run_tests.yaml

@@ -0,0 +1,53 @@
+name: Run benchmark tests
+on:
+  pull_request:
+  push:
+    branches: [master]
+
+jobs:
+  run_test:
+    name: Run optimization bench on intel_dev
+    runs-on: ubuntu-latest
+
+    strategy:
+      matrix:
+        python: ['3.10']
+
+    steps:
+      - name: Cancel Previous Runs
+        uses: styfle/cancel-workflow-action@0.11.0
+        with:
+          access_token: ${{ github.token }}
+
+      - uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Set pkgs_dirs
+        run: |
+          echo "pkgs_dirs: [~/.conda/pkgs]" >> ~/.condarc
+      
+      - name: Add conda to system path
+        run: echo $CONDA/bin >> $GITHUB_PATH
+
+      - name: Create Intel test environment
+        run: conda env create -f environments/intel.yaml
+
+      - name: Create stock test environment
+        run: conda env create -f environments/stock.yaml
+
+      - name: Run tests in Intel environment
+        run: |
+          . $CONDA/etc/profile.d/conda.sh
+          conda activate intel_env || exit 1
+          # This is toy run for GH action,
+          # The arguments are really bad for real perf testing, use default arguments instead
+          python numpy/umath/umath_mem_bench.py -v --size 10 --goal-time 0.01 --repeats 1 || exit 1
+
+      - name: Run tests in stock environment
+        run: |
+          . $CONDA/etc/profile.d/conda.sh
+          conda activate stock_env || exit 1
+          # This is toy run for GH action,
+          # The arguments are really bad for real perf testing, use default arguments instead
+          python numpy/umath/umath_mem_bench.py -v --size 10 --goal-time 0.01 --repeats 1 || exit 1

README.md

@@ -1,24 +1,25 @@
-[![Build Status](https://travis-ci.org/IntelPython/optimizations_bench.svg?branch=master)](https://travis-ci.org/IntelPython/optimizations_bench)
+[![Run benchmark tests](https://github.com/IntelPython/optimizations_bench/actions/workflows/run_tests.yaml/badge.svg)](https://github.com/IntelPython/optimizations_bench/actions/workflows/run_tests.yaml)
 
 # Optimization Benchmarks
 Collection of performance benchmarks used to present optimizations implemented for Intel(R) Distribution for Python*
 
 ## Environment Setup
 To install Python environments from Intel channel along with pip-installed packages
 
-- `python3 create_python_envs.py`
-
+- `conda env create -f environments/intel.yaml`
+- `conda activate intel_env`
+
 ## Run tests
-- `python3 run_tests.py`
+- `python numpy/umath/umath_mem_bench.py -v --size 10 --goal-time 0.01 --repeats 1`
 
 ## Run benchmarks
 ### umath
 - To run python benchmarks: `python numpy/umath/umath_mem_bench.py`
-- To compile and run native benchmarks (requires icc): `make -C numpy/umath`
+- To compile and run native benchmarks (requires `icx`): `make -C numpy/umath`
 
 ### Random number generation
 - To run python benchmarks: `python numpy/random/rng.py`
-- To compile and run native benchmarks (requires icc): `make -C numpy/random`
+- To compile and run native benchmarks (requires `icx`): `make -C numpy/random`
 
 ## See also
 "[Accelerating Scientific Python with Intel Optimizations](http://conference.scipy.org/proceedings/scipy2017/pdfs/oleksandr_pavlyk.pdf)" by Oleksandr Pavlyk, Denis Nagorny, Andres Guzman-Ballen, Anton Malakhov, Hai Liu, Ehsan Totoni, Todd A. Anderson, Sergey Maidanov. Proceedings of the 16th Python in Science Conference (SciPy 2017), July 10 - July 16, Austin, Texas

environments/intel.yaml

@@ -0,0 +1,15 @@
+name: intel_env
+channels:
+  - intel
+  - conda-forge
+  - nodefaults
+dependencies:
+  - numpy
+  - conda-forge::numexpr=*=mkl*
+  - numba
+  - scikit-learn
+  - cython
+  - pip
+  - pip:
+    - dask
+#    - rdtsc

environments/stock.yaml

@@ -0,0 +1,17 @@
+name: stock_env
+channels:
+  - intel
+  - conda-forge
+  - nodefaults
+dependencies:
+  - llvmlite
+  - cython
+  - pip
+  - pip:
+    - numpy
+    - scikit-learn
+    - toolz
+    - numexpr
+#    - rdtsc
+    - numba
+    - dask