1 # In earlier versions of isl, this test case would take an inordinate
2 # amount of time (in the order of 30s versus 0.1s in later versions)
3 # due to a call to isl_basic_set_coefficients.
4 # Check that this no longer happens in the sense that this test case
5 # would stand out if it were to take that long again.
6 # The actual schedule that is produced is not that important,
7 # but it is different depending on whether the whole-component scheduler
8 # is being used, so pick a particular setting.
9 # OPTIONS: --no-schedule-whole-component
10 domain: { S_3[0:223, 0:223, 0:15, 0:3, 0:6, 0:6, 0:15]; group0[0:3,
12 validity: { group0[i0 = 0:3, i1 = 0:223, i2 = 0:223, i3 = 0:15]
13 -> S_3[h = 0:223, w = 0:223, c0 = 0:15, kc1 = i0, kh = 3 + i1 - h,
14 kw = 3 + i2 - w, kc0 = i3] : -3 + i1 <= h <= 3 + i1 and -3 + i2 <=
15 w <= 3 + i2; group0[i0 = 0:3, i1 = 0:223, i2 = 0:223, i3 = 0:15] ->
16 group0[i0, i1, i2, i3] : (i1) mod 2 = 0 and (i2) mod 2 = 0; S_3[h =
17 0:223, w = 0:223, c0 = 0:15, kc1 = 0:3, kh = 0:6, kw = 0:6, kc0 = 0:14]
18 -> S_3[h' = h, w' = w, c0' = c0, kc1' = kc1, kh' = kh, kw' = kw, kc0'
19 = 1 + kc0]; S_3[h = 0:223, w = 0:223, c0 = 0:15, kc1 = 0:3, kh = 0:6,
20 kw = 0:5, kc0 = 15] -> S_3[h' = h, w' = w, c0' = c0, kc1' = kc1, kh'
21 = kh, kw' = 1 + kw, kc0' = 0]; S_3[h = 0:223, w = 0:223, c0 = 0:15,
22 kc1 = 0:3, kh = 0:5, kw = 6, kc0 = 15] -> S_3[h' = h, w' = w, c0' =
23 c0, kc1' = kc1, kh' = 1 + kh, kw' = 0, kc0' = 0]; S_3[h = 0:223, w =
24 0:223, c0 = 0:15, kc1 = 0:2, kh = 6, kw = 6, kc0 = 15] -> S_3[h' =
25 h, w' = w, c0' = c0, kc1' = 1 + kc1, kh' = 0, kw' = 0, kc0' = 0] }
26 proximity: { group0[i0 = 0:3, i1 = 0:223, i2 = 0:223, i3 = 0:15] ->
27 S_3[h = 0:223, w = 0:223, c0 = 0:15, kc1 = i0, kh = 3 + i1 - h, kw =
28 3 + i2 - w, kc0 = i3] : -3 + i1 <= h <= 3 + i1 and -3 + i2 <= w <=