[GR-62716] Use smaller dacapo archive whenever possible.

PullRequest: graal/21844

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()

sdk/mx.sdk/suite.py

@@ -315,6 +315,12 @@
       "packedResource": True,
     },
 
+    "DACAPO_23.11_MR2_chopin_minimal" : {
+      "urls" : ["https://lafo.ssw.uni-linz.ac.at/pub/graal-external-deps/dacapo-23.11-MR2-chopin-minimal.zip"],
+      "digest" : "sha512:51c04f81564f758e12c43dc44fc70fe2de15c5ea368e759b898a41bac3a458a421c1dbf557676533ba97c82283055fe73637cdc8c100b82384e5af9380113d40",
+      "packedResource": True,
+    },
+
     "DACAPO_SCALA" : {
       "urls" : ["https://lafo.ssw.uni-linz.ac.at/pub/graal-external-deps/dacapo-scala-0.1.0-20120216.jar"],
       "digest" : "sha512:9a3908f3a0f2937bbc7edcd78f7e7e229bea8dee566d5f2d922bd4dc3c5b02cf97de294e479200372ff90bfbdf80e88dc7fc9fda9cb294088346e4a5ff28893e",

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"]