ttones=tt.ones(10,400)
tt_test=ttones*0.
for i in range(50):
tt_test=tt_test+ttones
tt_test=tt_test.round(1e-3)
print(tt_test.r[10])
It seems like tt.round is not really controlling the rank growth for the above example (d=400). If one uses, say, ttones=tt.ones(10,200) instead, the loop finishes instantly with a rank-1 final TT array.
ttones=tt.ones(10,400) tt_test=ttones*0. for i in range(50): tt_test=tt_test+ttones tt_test=tt_test.round(1e-3) print(tt_test.r[10])
It seems like tt.round is not really controlling the rank growth for the above example (d=400). If one uses, say, ttones=tt.ones(10,200) instead, the loop finishes instantly with a rank-1 final TT array.