Read from dataclass fields

Summary: We can remove hardcoding the builtin metrics set by reading from the dataclass fields.

Reviewed By: int3

Differential Revision: D59989273

fbshipit-source-id: 76963617cc58395cdd853468970a96f101f86288

torchbenchmark/util/triton_op.py

@@ -43,20 +43,6 @@
 REGISTERED_METRICS: Dict[str, List[str]] = {}
 REGISTERED_X_VALS: Dict[str, str] = {}
 BASELINE_BENCHMARKS: Dict[str, str] = {}
-BUILTIN_METRICS = [
-    "latency",
-    "tflops",
-    "speedup",
-    "accuracy",
-    "compile_time",
-    "ncu_trace",
-    "ncu_rep",
-    "kineto_trace",
-    "cpu_peak_mem",
-    "gpu_peak_mem",
-    "hw_roofline",
-    "best_config",
-]
 BASELINE_SKIP_METRICS = set(["speedup", "accuracy"])
 X_ONLY_METRICS = set(["hw_roofline"])
 PRECISION_DTYPE_MAPPING = {
@@ -202,6 +188,7 @@ class BenchmarkOperatorMetrics:
     # extra metrics
     extra_metrics: Optional[Dict[str, float]] = None
 
+BUILTIN_METRICS =  set(map(lambda x: x.name, fields(BenchmarkOperatorMetrics))) - {"extra_metrics"}
 
 @dataclass
 class BenchmarkOperatorResult: