remove realize from threefry (#5969)

test/test_jit.py

@@ -266,6 +266,47 @@ class TestJit(unittest.TestCase):
     assert len(res3) == 5, "All values should be different, rand works in jit."
     assert res3 != res2, "Jit rand is diff with diff seeds"
 
+  def test_jit_multiple_random_regen(self):
+    def f(a, b):
+      rn = Tensor.randn(*a.shape)
+      rn = rn * a
+      rn2 = Tensor.randn(*a.shape)
+      rn2 = rn2 * b
+      rn = rn + rn2
+      rn2 = rn2 + Tensor.randn(*a.shape)
+      return ((a+b)*rn).realize(), ((a+b)*rn2).realize()
+    a = Tensor.randn(10, 10).realize()  # realize these before resetting the random seed
+    b = Tensor.randn(10, 10).realize()
+
+    Tensor.manual_seed(1234)
+    jf = TinyJit(f)
+    res = set()
+    for _ in range(5):
+      o1, o2 = jf(a, b)
+      res.add(o1.numpy()[0][0])
+      res.add(o2.numpy()[0][0])
+    assert len(res) == 10, "All values should be different, rand works in jit."
+
+    Tensor.manual_seed(1234)
+    jf2 = TinyJit(f)
+    res2 = set()
+    for _ in range(5):
+      o1, o2 = jf2(a, b)
+      res2.add(o1.numpy()[0][0])
+      res2.add(o2.numpy()[0][0])
+    assert len(res2) == 10, "All values should be different, rand works in jit."
+    assert res == res2, "Jit rand is not reproducible with the same seed"
+
+    Tensor.manual_seed(3421)
+    jf3 = TinyJit(f)
+    res3 = set()
+    for _ in range(5):
+      o1, o2 = jf3(a, b)
+      res3.add(o1.numpy()[0][0])
+      res3.add(o2.numpy()[0][0])
+    assert len(res3) == 10, "All values should be different, rand works in jit."
+    assert res3 != res2, "Jit rand is diff with diff seeds"
+
   def test_jit_realization_and_sampling(self):
     w = Tensor.eye(5)
 

test/test_linearizer.py

@@ -854,9 +854,9 @@ class TestLinearizer(unittest.TestCase):
       lin = Kernel(sched[0].ast)
       assert sum(u.arg is UnaryOps.RECIP for u in lin.linearize().uops) == max_ops, msg
 
-    a = Tensor.rand((4,4))
-    b = Tensor.rand((4,4))
-    d = Tensor.rand((4,4))
+    a = Tensor.empty((4,4))
+    b = Tensor.empty((4,4))
+    d = Tensor.empty((4,4))
 
     c = (a*b)/b
     helper(c, "found UnaryOps.RECIP in (a*b)/b operation")

tinygrad/tensor.py

@@ -401,7 +401,7 @@ class Tensor:
     print(Tensor.rand(5).numpy())
     ```
     """
-    Tensor._seed, Tensor._rng_counter = seed, Tensor([0], dtype=dtypes.uint32, requires_grad=False)
+    Tensor._seed, Tensor._rng_counter = seed, None
 
   @staticmethod
   def rand(*shape, device:Optional[Union[Tuple[str, ...], str]]=None, dtype:Optional[DTypeLike]=None, **kwargs):
@@ -417,7 +417,8 @@ class Tensor:
     print(t.numpy())
     ```
     """
-    if Tensor._rng_counter is None: Tensor._rng_counter = Tensor([0], dtype=dtypes.uint32, requires_grad=False)
+    if (had_counter := Tensor._rng_counter is None): Tensor._rng_counter = Tensor([0], dtype=dtypes.uint32, requires_grad=False)
+    if not all(s >= 0 for s in argfix(*shape)): raise ValueError(f"cannot create tensor with negative dimension in {shape=}")
     if not THREEFRY.value:
       # for bfloat16, numpy rand passes buffer in float
       if to_dtype(dtype or dtypes.default_float) == dtypes.bfloat16:
@@ -426,9 +427,9 @@ class Tensor:
 
     # threefry
     if (num := prod((shape:=argfix(*shape)))) == 0: return Tensor.zeros(shape, device=device, dtype=dtype, **kwargs)
-    counts1 = (Tensor.arange(math.ceil(num / 2), device=device, dtype=dtypes.uint32, requires_grad=False)+Tensor._rng_counter.to(device)).realize()
+    if not had_counter: Tensor._rng_counter.assign(Tensor._rng_counter + num)
+    counts1 = (Tensor.arange(math.ceil(num / 2), device=device, dtype=dtypes.uint32, requires_grad=False)+Tensor._rng_counter.to(device))
     counts2 = counts1 + math.ceil(num / 2)
-    Tensor._rng_counter.assign(Tensor._rng_counter + num).realize()
 
     x = counts2.cast(dtypes.uint64) << 32 | counts1.cast(dtypes.uint64)
     x = F.Threefry.apply(*x._broadcasted(Tensor._seed))