Non-deterministic simulation results at rt.sionna

[KoutsonasVag]

_I want to take power delay profile results in a ray tracing scenario where an object a transmitter near of him moves deterministically. I want to run the same scenario twice and take the same results. So every every random parameter like orientation at transmitter and receiver and receivers position are fixed. Then, when i run the code below twice the snnin variable is not same in every run as it should be .

SEED=106_
def txnmove(txp1,tx1,x,y):
    txp1.position=mi.Point3f(x,
                             y,
                             txp1.position[2])
    tx1.position=[(x+MALE_Rx[0]), (y-MALE_Rx[1]), MALE_Rx[2]]
x_pos=np.arange(1,21,0.75)
y_pos=np.arange(1,21,0.75)

person_power=[]
t_power=[]
tx_position=[]
p_solver = PathSolver()
for i in tqdm(range(len(x_pos))):
    if i%2==0:
        for j in range(len(y_pos)):
            txnmove(txp1=txp, tx1=tx,x=float(x_pos[i]),y=float(y_pos[j]))
            paths = p_solver(scene, max_depth=2,synthetic_array=True,seed=SEED) 
            a,tau=paths.cir(out_type="numpy",normalize_delays=False)
            person_power.append(np.sum(np.abs(a),axis=(1,3,5)))
            t_power.append(tau)
            tx_position.append([x_pos[i],y_pos[j]])
    else:
        for j in range(len(y_pos)-1,-1,-1):
            txnmove(txp1=txp, tx1=tx,x=float(x_pos[i]),y=float(y_pos[j]))
            paths = p_solver(scene, max_depth=2,synthetic_array=True,seed=SEED) 
            a,tau=paths.cir(out_type="numpy",normalize_delays=False)
            person_power.append(np.sum(np.abs(a),axis=(1,3,5)))
            t_power.append(tau)
            tx_position.append([x_pos[i],y_pos[j]])
            
k=[]
for i in t_power:
    k.append(i.shape[2])
snnin=np.zeros((len(t_power),4*max(k)))
person_pos=np.zeros((len(tx_position),2))
for i in tqdm(range(len(person_power))):
    count=0
    for j in range(len(t_power[i][0,0])):
        if person_power[i][0,0,j]>=1e-6:
            snnin[i,count]=t_power[i][0,0,j]
            snnin[i,count+1]=person_power[i][0,0,j]
        if person_power[i][0,0,j]>=1e-6:
            snnin[i,count+2]=t_power[i][1,0,j]
            snnin[i,count+3]=person_power[i][1,0,j]
        count+=4
    person_pos[i,0]=tx_position[i][0]
    person_pos[i,1]=tx_position[i][1]

snnin[snnin<0]=0

[merlinND]

Hello @KoutsonasVag,

Please see: #851 (comment)