diff --git a/backend.c b/backend.c
index 399c299e14..cb4f40826a 100644
--- a/backend.c
+++ b/backend.c
@@ -2251,9 +2251,11 @@ static bool waitee_running(struct thread_data *me)
 static void run_threads(struct sk_out *sk_out)
 {
 	struct thread_data *td;
-	unsigned int i, todo, nr_running, nr_started;
+	struct timespec last_finish_time;
+	unsigned int i, todo, nr_running, nr_started, prev_nr_running;
 	uint64_t m_rate, t_rate;
 	uint64_t spent;
+	uint64_t per_job_spent;
 
 	if (fio_gtod_offload && fio_start_gtod_thread())
 		return;
@@ -2294,6 +2296,7 @@ static void run_threads(struct sk_out *sk_out)
 	todo = thread_number;
 	nr_running = 0;
 	nr_started = 0;
+	prev_nr_running = 0;
 	m_rate = t_rate = 0;
 
 	for_each_td(td, i) {
@@ -2338,6 +2341,7 @@ static void run_threads(struct sk_out *sk_out)
 	fio_idle_prof_start();
 
 	set_genesis_time();
+	fio_gettime(&last_finish_time, NULL);
 
 	while (todo) {
 		struct thread_data *map[REAL_MAX_JOBS];
@@ -2380,6 +2384,13 @@ static void run_threads(struct sk_out *sk_out)
 				continue;
 			}
 
+			if (td->o.stonewall && td->o.per_job_start_delay) {
+				per_job_spent = utime_since_now(&last_finish_time);
+
+				if (td->o.per_job_start_delay > per_job_spent)
+					continue;
+			}
+
 			init_disk_util(td);
 
 			td->rusage_sem = fio_sem_init(FIO_SEM_LOCKED);
@@ -2498,7 +2509,12 @@ static void run_threads(struct sk_out *sk_out)
 			fio_sem_up(td->sem);
 		}
 
+		prev_nr_running = nr_running;
 		reap_threads(&nr_running, &t_rate, &m_rate);
+		if (nr_running == prev_nr_running - 1) {
+				//sequential job has finished get new base time
+				fio_gettime(&last_finish_time, NULL);
+		}
 
 		if (todo)
 			do_usleep(100000);
diff --git a/options.c b/options.c
index ddabaa82d2..66130a429e 100644
--- a/options.c
+++ b/options.c
@@ -2744,6 +2744,19 @@ struct fio_option fio_options[FIO_MAX_OPTS] = {
 		.category = FIO_OPT_C_GENERAL,
 		.group	= FIO_OPT_G_RUNTIME,
 	},
+	{
+		.name	= "per_job_startdelay",
+		.lname	= "Per Job Start delay",
+		.type	= FIO_OPT_STR_VAL_TIME,
+		.off1	= offsetof(struct thread_options, per_job_start_delay),
+		.off2	= offsetof(struct thread_options, per_job_start_delay_high),
+		.help	= "Only start job when this period has passed",
+		.def	= "0",
+		.is_seconds = 1,
+		.is_time = 1,
+		.category = FIO_OPT_C_GENERAL,
+		.group	= FIO_OPT_G_RUNTIME,
+	},
 	{
 		.name	= "runtime",
 		.lname	= "Runtime",
diff --git a/thread_options.h b/thread_options.h
index 5ecc72d7b5..ab794b6311 100644
--- a/thread_options.h
+++ b/thread_options.h
@@ -187,6 +187,9 @@ struct thread_options {
 	unsigned long long start_delay;
 	unsigned long long start_delay_orig;
 	unsigned long long start_delay_high;
+	unsigned long long per_job_start_delay;
+	unsigned long long per_job_start_delay_orig;
+	unsigned long long per_job_start_delay_high;
 	unsigned long long timeout;
 	unsigned long long ramp_time;
 	unsigned int ss_state;