250731.200131.PDT try starting a parpool with the number of cores before profiling

Author: zaikunzhang

.github/workflows/profile_cobyla_small.yml

@@ -41,6 +41,18 @@ jobs:
       - name: Get the solver name
         run:  echo "SOLNAME=$(echo ${{ matrix.solver }} | cut -c1-6)" >> $GITHUB_ENV
 
+      - name: Free disk space
+        uses: jlumbroso/free-disk-space@main
+        if: startsWith(matrix.os, 'ubuntu')
+        with:
+          # all of these default to true, but feel free to set to "false" if necessary for your workflow
+          android: true
+          dotnet: true
+          haskell: true
+          large-packages: true
+          docker-images: true
+          swap-storage: false  # Important, or the runner may be shut down due to memory starvation.
+
       - name: Clone Repository (Latest)
         uses: actions/[email protected]
         if: github.event.inputs.git-ref == ''
@@ -109,6 +121,8 @@ jobs:
         with:
           command: |
             ver;
+            feature('numCores')  % Check the number of available cores
+            parpool(feature('numCores'));  % Start a parallel pool with the number of available cores
             root_dir = pwd();
 
             cd(fullfile(root_dir, 'matcutest')); install(); which macup

.github/workflows/profile_lincoa_small.yml

@@ -115,6 +115,8 @@ jobs:
         with:
           command: |
             ver;
+            feature('numCores')  % Check the number of available cores
+            parpool(feature('numCores'));  % Start a parallel pool with the number of available cores
             root_dir = pwd();
 
             cd(fullfile(root_dir, 'matcutest')); install(); which macup