diff --git a/src/cli/shared.rs b/src/cli/shared.rs
index 14f3bfda..1fe13474 100644
--- a/src/cli/shared.rs
+++ b/src/cli/shared.rs
@@ -136,12 +136,7 @@ pub struct ExecAndRunSharedArgs {
     pub exclude_allocations: bool,
 
     /// Measure the subprocesses spawned by the benchmarked process in simulation mode.
-    #[arg(
-        long,
-        env = "CODSPEED_SIMULATION_TRACK_SUBPROCESS",
-        default_value_t = false,
-        action = clap::ArgAction::Set
-    )]
+    #[arg(long, env = "CODSPEED_SIMULATION_TRACK_SUBPROCESS")]
     pub simulation_track_subprocess: bool,
 
     #[command(flatten)]
diff --git a/src/executor/valgrind/measure.rs b/src/executor/valgrind/measure.rs
index 54bddf1a..62807b92 100644
--- a/src/executor/valgrind/measure.rs
+++ b/src/executor/valgrind/measure.rs
@@ -33,19 +33,18 @@ fn get_valgrind_args(tool: &SimulationTool, config: &ExecutorConfig) -> Vec<Stri
     .map(|x| x.to_string())
     .collect();
 
-    args.push(if config.simulation_track_subprocess {
-        "--instr-atstart=inherit".to_string()
+    if config.simulation_track_subprocess {
+        args.push("--instr-atstart=inherit".to_string());
+        args.push("--separate-threads=yes".to_string());
     } else {
-        "--instr-atstart=no".to_string()
-    });
+        args.push("--instr-atstart=no".to_string());
+        args.push("--separate-threads=no".to_string());
+    }
 
     match tool {
         SimulationTool::Callgrind => {
             if config.cycle_estimation {
                 args.push("--cycle-estimation=yes".to_string());
-                args.push("--separate-threads=yes".to_string());
-            } else {
-                args.push("--separate-threads=no".to_string());
             }
 
             args.push("--tool=callgrind".to_string());