diff --git a/kernel/sched/walt/walt_cfs.c b/kernel/sched/walt/walt_cfs.c
index 989ac84ad934..b6378331fd61 100644
--- a/kernel/sched/walt/walt_cfs.c
+++ b/kernel/sched/walt/walt_cfs.c
@@ -219,6 +219,7 @@ enum fastpaths {
 	NONE = 0,
 	SYNC_WAKEUP,
 	PREV_CPU_FASTPATH,
+	CLUSTER_PACKING_FASTPATH,
 };
 
 static inline bool is_complex_sibling_idle(int cpu)
@@ -251,6 +252,7 @@ static void walt_find_best_target(struct sched_domain *sd,
 	bool rtg_high_prio_task = task_rtg_high_prio(p);
 	cpumask_t visit_cpus;
 	struct walt_task_struct *wts = (struct walt_task_struct *) p->android_vendor_data1;
+	int packing_cpu;
 	struct walt_rq *prev_wrq = (struct walt_rq *) cpu_rq(prev_cpu)->android_vendor_data1;
 	struct walt_rq *start_wrq;
 
@@ -272,6 +274,14 @@ static void walt_find_best_target(struct sched_domain *sd,
 		most_spare_wake_cap = LONG_MIN;
 	}
 
+	/* fast path for packing_cpu */
+	packing_cpu = walt_find_cluster_packing_cpu(start_cpu);
+	if (walt_choose_packing_cpu(packing_cpu, p)) {
+		fbt_env->fastpath = CLUSTER_PACKING_FASTPATH;
+		cpumask_set_cpu(packing_cpu, candidates);
+			goto out;
+	}
+
 	/* fast path for prev_cpu */
 	if (((prev_wrq->cluster->id == start_wrq->cluster->id) ||
 				asym_cap_siblings(prev_cpu, start_cpu)) &&
@@ -842,6 +852,12 @@ int walt_find_energy_efficient_cpu(struct task_struct *p, int prev_cpu,
 		goto unlock;
 
 	first_cpu = cpumask_first(candidates);
+
+	if (fbt_env.fastpath == CLUSTER_PACKING_FASTPATH) {
+		best_energy_cpu = first_cpu;
+		goto unlock;
+	}
+
 	if (weight == 1) {
 		if (available_idle_cpu(first_cpu) || first_cpu == prev_cpu) {
 			best_energy_cpu = first_cpu;
diff --git a/kernel/sched/walt/walt_rt.c b/kernel/sched/walt/walt_rt.c
index 8952f2504f8c..02ba1d2200c1 100644
--- a/kernel/sched/walt/walt_rt.c
+++ b/kernel/sched/walt/walt_rt.c
@@ -234,6 +234,7 @@ static void walt_select_task_rq_rt(void *unused, struct task_struct *task, int c
 	bool sync = !!(wake_flags & WF_SYNC);
 	int ret, target = -1, this_cpu;
 	struct cpumask *lowest_mask;
+	int packing_cpu;
 
 	if (unlikely(walt_disabled))
 		return;
@@ -288,6 +289,13 @@ static void walt_select_task_rq_rt(void *unused, struct task_struct *task, int c
 	ret = cpupri_find_fitness(&task_rq(task)->rd->cpupri, task,
 				lowest_mask, walt_rt_task_fits_capacity);
 
+	/* create a fastpath for finding a packing cpu */
+	packing_cpu = walt_find_cluster_packing_cpu(task_cpu(task));
+	if (walt_choose_packing_cpu(packing_cpu, task)) {
+		*new_cpu = packing_cpu;
+		goto unlock;
+	}
+
 	walt_rt_energy_aware_wake_cpu(task, lowest_mask, ret, &target);
 
 	/*
@@ -310,7 +318,7 @@ static void walt_select_task_rq_rt(void *unused, struct task_struct *task, int c
 		if (target < nr_cpu_ids)
 			*new_cpu = target;
 	}
-
+unlock:
 	rcu_read_unlock();
 }
 
@@ -319,6 +327,15 @@ static void walt_rt_find_lowest_rq(void *unused, struct task_struct *task,
 				   struct cpumask *lowest_mask, int ret, int *best_cpu)
 
 {
+	int packing_cpu;
+
+	/* create a fastpath for finding a packing cpu */
+	packing_cpu = walt_find_cluster_packing_cpu(task_cpu(task));
+	if (walt_choose_packing_cpu(packing_cpu, task)) {
+		*best_cpu = packing_cpu;
+		return;
+	}
+
 	walt_rt_energy_aware_wake_cpu(task, lowest_mask, ret, best_cpu);
 
 	/*