diff --git a/benchmarks/scripts/cub/bench/config.py b/benchmarks/scripts/cub/bench/config.py
index ee393ff12..eea9515ea 100644
--- a/benchmarks/scripts/cub/bench/config.py
+++ b/benchmarks/scripts/cub/bench/config.py
@@ -90,6 +90,21 @@ def __new__(cls, *args, **kwargs):
             cls._instance = super().__new__(cls, *args, **kwargs)
             cls._instance.ctk, cls._instance.cub, cls._instance.benchmarks = parse_meta()
         return cls._instance
+    
+    def label_to_variant_point(self, algname, label):
+        if label == "base":
+            return BasePoint()
+
+        label_to_definition = {}
+        for param_space in self.benchmarks[algname]:
+            label_to_definition[param_space.label] = param_space.definition
+        
+        points = []
+        for point in label.split('.'):
+            label, value = point.split('_')
+            points.append(RangePoint(label_to_definition[label], label, int(value)))
+        
+        return VariantPoint(points)
 
     def variant_space(self, algname):
         variants = []
diff --git a/benchmarks/scripts/verify.py b/benchmarks/scripts/verify.py
new file mode 100755
index 000000000..f653fa4aa
--- /dev/null
+++ b/benchmarks/scripts/verify.py
@@ -0,0 +1,47 @@
+#!/usr/bin/env python3
+
+import cub.bench 
+
+
+def workload_header(ct_workload_space, rt_workload_space):
+    for ct_workload in ct_workload_space:
+        for rt_workload in rt_workload_space:
+            workload_point = ct_workload + rt_workload
+            return ", ".join([x.split('=')[0] for x in workload_point])
+
+
+def workload_entry(ct_workload, rt_workload):
+    workload_point = ct_workload + rt_workload
+    return ", ".join([x.split('=')[1] for x in workload_point])
+
+
+class VerifySeeker:
+    def __init__(self):
+        self.estimator = cub.bench.MedianCenterEstimator()
+
+    def __call__(self, algname, ct_workload_space, rt_workload_space):
+        label = 'trp_0.ld_0.ipt_20.tpb_448.l2b_485.l2w_710'
+        variant_point = cub.bench.Config().label_to_variant_point(algname, label)
+
+        print("{}, MinS, MedianS, MaxS".format(workload_header(ct_workload_space, rt_workload_space)))
+        for ct_workload in ct_workload_space:
+            bench = cub.bench.Bench(algname, variant_point, list(ct_workload))
+            if bench.build():
+                base = bench.get_base()
+                for rt_workload in rt_workload_space:
+                    workload_point = ct_workload + rt_workload
+                    base_samples, base_elapsed = base.do_run(workload_point, None)
+                    variant_samples, _ = bench.do_run(workload_point, base_elapsed * 10)
+                    min_speedup = min(base_samples) / min(variant_samples)
+                    median_speedup = self.estimator(base_samples) / self.estimator(variant_samples)
+                    max_speedup = max(base_samples) / max(variant_samples)
+                    point_str = workload_entry(ct_workload, rt_workload)
+                    print("{}, {}, {}, {}".format(point_str, min_speedup, median_speedup, max_speedup))
+
+
+def main():
+    cub.bench.search(VerifySeeker())
+
+
+if __name__ == "__main__":
+    main()