Revise how we set tiering and pgo options for spmi benchmark collections (#87292)

Don't set these in the BDN environment; set them via the BDN command line
so they only impact the process being benchmarked.

Fixes #86410

diff --git a/src/coreclr/scripts/superpmi_benchmarks.py b/src/coreclr/scripts/superpmi_benchmarks.py
index 8b78a00..a20d393 100644 (file)
--- a/src/coreclr/scripts/superpmi_benchmarks.py
+++ b/src/coreclr/scripts/superpmi_benchmarks.py
@@ -206,25 +206,33 @@ def build_and_run(coreclr_args, output_mch_name):
          "--framework", "net8.0", "--no-restore", "/p:NuGetPackageRoot=" + artifacts_packages_directory,
          "-o", artifacts_directory], _exit_on_fail=True)
 
+    # common BDN prefix
+    collection_command = f"{dotnet_exe} {benchmarks_dll} --corerun {os.path.join(core_root, corerun_exe)} "
+
+    # test specific filters
     if benchmark_binary.lower().startswith("microbenchmarks"):
-        # Disable ReadyToRun so we always JIT R2R methods and collect them
-        collection_command = f"{dotnet_exe} {benchmarks_dll}  --filter \"*\" --corerun {os.path.join(core_root, corerun_exe)} --partition-count {partition_count} " \
-                            f"--partition-index {partition_index} --envVars DOTNET_JitName:{shim_name} " \
-                            " DOTNET_ZapDisable:1  DOTNET_ReadyToRun:0 " \
-                            "--iterationCount 1 --warmupCount 0 --invocationCount 1 --unrollFactor 1 --strategy ColdStart --logBuildOutput"
+        collection_command += f"--filter \"*\" --partition-count {partition_count} --partition-index {partition_index} "
     elif benchmark_binary.lower().startswith("demobenchmarks"):
-        # Disable ReadyToRun so we always JIT R2R methods and collect them
-        collection_command = f"{dotnet_exe} {benchmarks_dll}  -f *CollisionBatcherTaskBenchmarks.* *GroupedCollisionTesterBenchmarks.* *GatherScatterBenchmarks.* " \
-                            " *OneBodyConstraintBenchmarks.* *TwoBodyConstraintBenchmarks.* *ThreeBodyConstraintBenchmarks.* *FourBodyConstraintBenchmarks.* " \
-                            " *SweepBenchmarks.* *ShapeRayBenchmarks.* *ShapePileBenchmark.* *RagdollTubeBenchmark.* " \
-                            f" --corerun {os.path.join(core_root, corerun_exe)} --envVars DOTNET_JitName:{shim_name} " \
-                            " DOTNET_ZapDisable:1  DOTNET_ReadyToRun:0 " \
-                            "--iterationCount 1 --warmupCount 0 --invocationCount 1 --unrollFactor 1 --strategy ColdStart --logBuildOutput"
+        collection_command += "-f *CollisionBatcherTaskBenchmarks.* *GroupedCollisionTesterBenchmarks.* *GatherScatterBenchmarks.* " \
+                              " *OneBodyConstraintBenchmarks.* *TwoBodyConstraintBenchmarks.* *ThreeBodyConstraintBenchmarks.* *FourBodyConstraintBenchmarks.* " \
+                              " *SweepBenchmarks.* *ShapeRayBenchmarks.* *ShapePileBenchmark.* *RagdollTubeBenchmark.* "
+    else:
+        collection_command += "--filter \"*\" "
+
+    # common BDN arguments
+    collection_command += "--iterationCount 1 --warmupCount 0 --invocationCount 1 --unrollFactor 1 --strategy ColdStart --logBuildOutput "
+
+    # common BDN environment var settings
+    # Disable ReadyToRun so we always JIT R2R methods and collect them
+    collection_command += f"--envVars DOTNET_JitName:{shim_name} DOTNET_ZapDisable:1 DOTNET_ReadyToRun:0 "
+
+    # custom BDN environment var settings
+    if coreclr_args.tiered_pgo:
+        collection_command += "DOTNET_TieredCompilation:1 DOTNET_TieredPGO:1"
+    elif coreclr_args.tiered_compilation:
+        collection_command += "DOTNET_TieredCompilation:1 DOTNET_TieredPGO:0"
     else:
-        # Disable ReadyToRun so we always JIT R2R methods and collect them
-        collection_command = f"{dotnet_exe} {benchmarks_dll}  --filter \"*\" --corerun {os.path.join(core_root, corerun_exe)} --envVars DOTNET_JitName:{shim_name} " \
-                            " DOTNET_ZapDisable:1  DOTNET_ReadyToRun:0 " \
-                            "--iterationCount 1 --warmupCount 0 --invocationCount 1 --unrollFactor 1 --strategy ColdStart --logBuildOutput"
+        collection_command += "DOTNET_TieredCompilation:0"
 
     # Generate the execution script in Temp location
     with TempDir() as temp_location:
@@ -232,13 +240,18 @@ def build_and_run(coreclr_args, output_mch_name):
 
         contents = []
         # Unset the JitName so dotnet process will not fail
+        # Unset TieredCompilation and TieredPGO so the parent BDN process is just running with defaults
         if is_windows:
             contents.append("set JitName=%DOTNET_JitName%")
             contents.append("set DOTNET_JitName=")
+            contents.append("set DOTNET_TieredCompilation=")
+            contents.append("set DOTNET_TieredPGO=")
         else:
             contents.append("#!/bin/bash")
             contents.append("export JitName=$DOTNET_JitName")
             contents.append("unset DOTNET_JitName")
+            contents.append("unset DOTNET_TieredCompilation")
+            contents.append("unset DOTNET_TieredPGO")
         contents.append(f"pushd {performance_directory}")
         contents.append(collection_command)
 
@@ -262,12 +275,7 @@ def build_and_run(coreclr_args, output_mch_name):
                        "-output_mch_path", output_mch_name,
                        "-log_level", "debug"]
 
-        if coreclr_args.tiered_compilation:
-            script_args.append("--tiered_compilation");
-        elif coreclr_args.tiered_pgo:
-            script_args.append("--tiered_pgo");
-
-        script_args.append(script_name);
+        script_args.append(script_name)
 
         run_command(script_args, _exit_on_fail=True)