libint (package): add tuning options for MOLGW (#17329)

var/spack/repos/builtin/packages/libint/package.py

@@ -13,6 +13,10 @@
     'cp2k-lmax-5',
     'cp2k-lmax-6',
     'cp2k-lmax-7',
+    'molgw-lmax-4',
+    'molgw-lmax-5',
+    'molgw-lmax-6',
+    'molgw-lmax-7',
 )
 
 
@@ -136,6 +140,25 @@ def configure_args(self):
                     '--enable-generic-code',
                     '--disable-unrolling',
                 ]
+            if tune_value.startswith('molgw'):
+                lmax = int(tune_value.split('-lmax-')[1])
+                config_args += [
+                    '--enable-1body=1',
+                    '--enable-eri=0',
+                    '--enable-eri2=0',
+                    '--enable-eri3=0',
+                    '--with-multipole-max-order=0',
+                    '--with-max-am={0}'.format(lmax),
+                    '--with-eri-max-am={0}'.format(lmax),
+                    '--with-eri2-max-am={0}'.format(lmax),
+                    '--with-eri3-max-am={0}'.format(lmax),
+                    '--with-opt-am=2',
+                    '--enable-contracted-ints',
+                    # keep code-size at an acceptable limit,
+                    # cf. https://github.com/evaleev/libint/wiki#program-specific-notes:
+                    '--enable-generic-code',
+                    '--disable-unrolling',
+                ]
 
         return config_args