diff --git a/tests/benchmarks/sched/README.rst b/tests/benchmarks/sched/README.rst
index 89164b061be..90141b558b0 100644
--- a/tests/benchmarks/sched/README.rst
+++ b/tests/benchmarks/sched/README.rst
@@ -19,14 +19,3 @@ _pend_curr_irqlock().  From this initial state:
 It then iterates this many times, reporting timestamp latencies
 between each numbered step and for the whole cycle, and a running
 average for all cycles run.
-
-Note that because this involves no timer interaction (except, on some
-architectures, k_cycle_get_32()), it works correctly when run in QEMU
-using the -icount argument, which can produce 100% deterministic
-behavior (not cycle-exact hardware simulation, but exactly N
-instructions per simulated nanosecond).  You can enable this using an
-environment variable (set at cmake time -- it's not enough to do this
-for the subsequent make/ninja invocation, cmake needs to see the
-variable itself):
-
-    export QEMU_EXTRA_FLAGS="-icount shift=0,align=off,sleep=off"
diff --git a/tests/benchmarks/sched/src/main.c b/tests/benchmarks/sched/src/main.c
index a4497756411..fe354522070 100644
--- a/tests/benchmarks/sched/src/main.c
+++ b/tests/benchmarks/sched/src/main.c
@@ -27,15 +27,6 @@
  * It then iterates this many times, reporting timestamp latencies
  * between each numbered step and for the whole cycle, and a running
  * average for all cycles run.
- *
- * Note that because this involves no timer interaction (except, on
- * some architectures, k_cycle_get_32()), it works correctly when run
- * in qemu using the -icount argument, which can produce 100%
- * deterministic behavior (not cycle-exact hardware simulation, but
- * exactly N instructions per simulated nanosecond).  You can enable
- * for "make run" using an environment variable:
- *
- * export QEMU_EXTRA_FLAGS="-icount shift=0,align=off,sleep=off"
  */
 
 #define N_RUNS 1000