Internal error during partition loops

[TH3CHARLie]

A simple repro

    Func input("input");
    Func local_sum("local_sum");
    Func blurry("blurry");
    Var x("x"), y("y");
    input(x, y) = 2 * x + 5 * y;
    RDom r(-2, 5, -2, 5);
    local_sum(x, y) = 0;
    local_sum(x, y) += input(x + r.x, y + r.y);
    blurry(x, y) = cast<int32_t>(local_sum(x, y) / 25);
    Var yo("yo"), yi("yi"), xo("xo"), xi("xi"), yo_x_f("yo_x_f"), yo_x_fo("yo_x_fo"), yo_x_fi("yo_x_fi");
    blurry.split(y,yo, yi, 2, TailStrategy::RoundUp).fuse(yo, x, yo_x_f).vectorize(yi).split(yo_x_f, yo_x_fo, yo_x_fi, 2, TailStrategy::Predicate).reorder(yo_x_fo, yo_x_fi, yi);
    input.split(y, yo, yi, 2, TailStrategy::PredicateStores).fuse(yo, x, yo_x_f).vectorize(yi).split(yo_x_f, yo_x_fo, yo_x_fi, 2, TailStrategy::Predicate).reorder(yo_x_fo, yo_x_fi, yi);
    blurry.store_root();
    input.compute_at(blurry, yi);
    Pipeline p({blurry});
    Buffer<int> buf = p.realize({32, 32});

Triggers:

Internal Error at /home/xuanda/dev/Serializer/Halide/src/PartitionLoops.cpp:276 triggered by user code at : Condition failed: s.condition.type().is_scalar(): (ramp((((((input.s0.y.yo.yo_x_f.yo_x_fo*2) + input.s0.y.yo.yo_x_f.yo_x_fi) % ((blurry.extent.1 + 5)/2))*2) + blurry.min.1) + -4, 1, 2) <= x2((((blurry.extent.1 + -1)/2)*2) + blurry.min.1))