Use minimal dacapo archive when using small or default workload sizes

Author: farquet

sdk/mx.sdk/mx_sdk_benchmark.py

@@ -2055,9 +2055,15 @@ def jarPath(self, benchmark):
     def dataLocation(self):
         if self.version() == "23.11-MR2-chopin":
             basePath = self.daCapoPath()
-            return os.path.join(basePath, "dacapo-23.11-MR2-chopin")
+            subdir = "dacapo-23.11-MR2-chopin"
+            if self.minimalArchive():
+                subdir += "-minimal"
+            return os.path.join(basePath, subdir)
         else:
-            raise "data location is only supported for version 23.11-MR2-chopin"
+            raise f"data location is not supported for suite version '{self.version()}'"
+
+    def minimalArchive(self):
+        return False
 
     def createCommandLineArgs(self, benchmarks, bmSuiteArgs):
         if benchmarks is None:
@@ -2227,17 +2233,24 @@ def workloadSize(self):
     def daCapoClasspathEnvVarName(self):
         return "DACAPO_CP"
 
+    def minimalArchive(self):
+        # DaCapo Chopin archive is huge. A stripped version without large and huge sizes exists
+        # See dacapobench/dacapobench issue #345 on GitHub
+        return self.version() in ["23.11-MR2-chopin"] and self.workloadSize() in ["default", "tiny", "small"]
+
     def daCapoLibraryName(self):
+        library = None
         if self.version() == "9.12-bach":  # 2009 release
-            return "DACAPO"
+            library = "DACAPO"
         elif self.version() == "9.12-MR1-bach":  # 2018 maintenance release (January 2018)
-            return "DACAPO_MR1_BACH"
+            library = "DACAPO_MR1_BACH"
         elif self.version() == "9.12-MR1-git+2baec49":  # commit from July 2018
-            return "DACAPO_MR1_2baec49"
+            library = "DACAPO_MR1_2baec49"
         elif self.version() == "23.11-MR2-chopin":
-            return "DACAPO_23.11_MR2_chopin"
-        else:
-            return None
+            library = "DACAPO_23.11_MR2_chopin"
+        if library and self.minimalArchive():
+            library += "_minimal"
+        return library
 
     def daCapoIterations(self):
         iterations = _daCapoIterations.copy()

substratevm/mx.substratevm/mx_substratevm_benchmark.py

@@ -672,12 +672,6 @@ def name(self):
     def benchSuiteName(self, bmSuiteArgs=None):
         return 'dacapo'
 
-    def daCapoPath(self):
-        lib = mx.library(self.daCapoLibraryName(), False)
-        if lib:
-            return lib.get_path(True)
-        return None
-
     def availableSuiteVersions(self):
         # The version 9.12-MR1-git+2baec49 also ships a custom harness class to allow native image to find the entry point in the nested jar
         return ["9.12-MR1-git+2baec49", "23.11-MR2-chopin"]