diff --git a/test/test_jit.py b/test/test_jit.py
index a7d4536cfd..bba8ebd54a 100644
--- a/test/test_jit.py
+++ b/test/test_jit.py
@@ -838,7 +838,7 @@ class TestJitRandom(unittest.TestCase):
     tst = {0:[], 1:[]}
     for r in [0,1]:
       Tensor.manual_seed(1337)
-      with Context(RANGEIFY=r):
+      with Context(JIT=r):
         _ = Tensor.randint(4, high=3)
         # this second one makes the behavior different
         _ = Tensor.randint(4, high=3)
diff --git a/test/test_multitensor.py b/test/test_multitensor.py
index 253cedec19..fdf07f3ca7 100644
--- a/test/test_multitensor.py
+++ b/test/test_multitensor.py
@@ -2,7 +2,7 @@ import unittest, functools, random
 from tinygrad import Tensor, Device, nn, GlobalCounters, TinyJit, dtypes, Variable
 from tinygrad.device import is_dtype_supported
 from tinygrad.uop.ops import Ops, UOp
-from tinygrad.helpers import CI, getenv, prod, Context, RANGEIFY
+from tinygrad.helpers import CI, getenv, prod, Context
 from tinygrad.nn.state import get_parameters, get_state_dict
 from tinygrad.engine.realize import lower_schedule, BufferCopy, CompiledRunner, run_schedule
 import numpy as np
@@ -390,7 +390,7 @@ class TestMultiTensor(unittest.TestCase):
 
   # NOTE: this is failing on LLVM CI, no idea why. Works locally.
   @unittest.skipIf(CI and REAL_DEV in ("CUDA", "NV", "CPU", "AMD"), "slow, and flaky on CPU")
-  @unittest.skipIf(RANGEIFY, "TODO: pm_rangeify hangs")
+  @unittest.skip("TODO: pm_rangeify hangs")
   def test_data_parallel_resnet(self):
     from extra.models.resnet import ResNet18
 
@@ -427,7 +427,7 @@ class TestMultiTensor(unittest.TestCase):
     np.testing.assert_allclose(grad, shard_grad, atol=1e-5, rtol=1e-5)
 
   @unittest.skipIf(CI and REAL_DEV in ("CUDA", "NV", "CPU", "AMD"), "slow, and flaky on CPU")
-  @unittest.skipIf(RANGEIFY, "TODO: pm_rangeify hangs")
+  @unittest.skip("TODO: pm_rangeify hangs")
   def test_data_parallel_resnet_train_step(self):
     from extra.models.resnet import ResNet18
     fake_image = Tensor.rand((2, 3, 224//16, 224//16))
@@ -435,7 +435,7 @@ class TestMultiTensor(unittest.TestCase):
     m = ResNet18()
     self._test_model_train_step(m, fake_image, labels)
 
-  @unittest.skipIf(RANGEIFY, "TODO: pm_rangeify hangs")
+  @unittest.skip("TODO: pm_rangeify hangs")
   def test_data_parallel_simple_train_step(self):
     class Model:
       def __init__(self): self.conv1 = nn.Linear(128,128)
@@ -800,7 +800,7 @@ class TestMultiTensor(unittest.TestCase):
     t = Tensor.rand(16, 16).shard(devices_2, axis=0)
     np.testing.assert_allclose(t.numpy(), t.clone().numpy())
 
-  @unittest.skipIf(RANGEIFY, "RANGEIFY doesn't support multi const folding")
+  @unittest.skip("RANGEIFY doesn't support multi const folding")
   def test_multi_const_folding(self):
     with Context(TRACK_MATCH_STATS=0):
       a = Tensor.arange(3).realize()
diff --git a/test/test_rangeify.py b/test/test_rangeify.py
index 4f22a1dcc8..144375c6f4 100644
--- a/test/test_rangeify.py
+++ b/test/test_rangeify.py
@@ -70,6 +70,7 @@ class TestRangeify(unittest.TestCase):
     ret = A.sum(axis=2).contiguous(arg=(1,)).sum(axis=1)
     ret.realize()
 
+  @unittest.skip("RANGEIFY=0 does nothing")
   def test_double_gemm_real(self):
     def go():
       with Context(DEBUG=0):
@@ -199,6 +200,7 @@ class TestRangeify(unittest.TestCase):
     out = blk._feed_forward(x)
     out.realize()
 
+  @unittest.skip("RANGEIFY=0 does nothing")
   def test_flash_attention(self):
     BS, HEADS, SEQLEN, EMB = 4, 2, 16, 8
 
diff --git a/test/test_schedule.py b/test/test_schedule.py
index 68fe09b0f7..4817272317 100644
--- a/test/test_schedule.py
+++ b/test/test_schedule.py
@@ -344,7 +344,7 @@ class TestSchedule(unittest.TestCase):
     out1 = r1 + y
     schedule = check_schedule([out0, out1], 2)
     reduceops = [x for si in schedule for x in si.ast.toposort() if x.op in {Ops.REDUCE_AXIS, Ops.REDUCE}]
-    assert len(reduceops) in [2,3]  # why is RANGEIFY different?
+    self.assertEqual(len(reduceops), 2) # why is RANGEIFY different?
 
   def test_div_collapse_buffer(self):
     a = Tensor.full((4,), 4.0).contiguous().realize()
diff --git a/tinygrad/helpers.py b/tinygrad/helpers.py
index fd13c8b166..a7a0357403 100644
--- a/tinygrad/helpers.py
+++ b/tinygrad/helpers.py
@@ -141,7 +141,7 @@ DONT_REALIZE_EXPAND, DONT_GROUP_REDUCES = ContextVar("DONT_REALIZE_EXPAND", 0),
 QUANTIZE, VALIDATE_WITH_CPU, DISABLE_FAST_IDIV = ContextVar("QUANTIZE", 0), ContextVar("VALIDATE_WITH_CPU", 0), ContextVar("DISABLE_FAST_IDIV", 0)
 CORRECT_DIVMOD_FOLDING, FUSE_OPTIM = ContextVar("CORRECT_DIVMOD_FOLDING", 0), ContextVar("FUSE_OPTIM", 0)
 ALLOW_DEVICE_USAGE, MAX_BUFFER_SIZE = ContextVar("ALLOW_DEVICE_USAGE", 1), ContextVar("MAX_BUFFER_SIZE", 0)
-RANGEIFY, FUSE_ATTENTION = ContextVar("RANGEIFY", 1), ContextVar("FUSE_ATTENTION", 0)
+FUSE_ATTENTION = ContextVar("FUSE_ATTENTION", 0)
 EMULATE = ContextVar("EMULATE", "")
 CPU_COUNT = ContextVar("CPU_COUNT", max(1, (os.cpu_count() or 1) // (4 if ARCH_X86 else 2))) # take 1/2 of the cores, accounting HT
 CPU_LLVM, AMD_LLVM = ContextVar("CPU_LLVM", 0), ContextVar("AMD_LLVM", 1)