MultipleDoubleBuffer_CUDA might fail sometimes

[zasdfgbnm]

Have no idea why, but I am seeing the following failure non-deterministically after we become our own repo.

[ RUN      ] LoopRotationTest.MultipleDoubleBuffer_CUDA
unknown file: Failure
C++ exception with description "aten_output_tensor.allclose( fusion_output_tensor.to(aten_output_tensor.dtype()), tolerance_values.second, tolerance_values.first, true) INTERNAL ASSERT FAILED at "/home/gaoxiang/Fuser/test/test_gpu_validator.h":400, please report a bug to PyTorch. 

Validation error in output 0 on line 712 in file /home/gaoxiang/Fuser/test/test_loop_rotation.cpp.
  Detected abs error of: 2.55172
    absolute tolerance was set to 1.68222e-06
    and relative tolerance set to 2.23704e-06
Exception raised from testValidate at /home/gaoxiang/Fuser/test/test_gpu_validator.h:400 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x5c (0x7f33003bccdc in /home/gaoxiang/pytorch-viable/build/lib/libc10.so)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) + 0x64 (0x7f33003865f6 in /home/gaoxiang/pytorch-viable/build/lib/libc10.so)
frame #2: c10::detail::torchInternalAssertFail(char const*, char const*, unsigned int, char const*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) + 0x4f (0x7f33003bacaf in /home/gaoxiang/pytorch-viable/build/lib/libc10.so)
frame #3: <unknown function> + 0x37c27d (0x560a19f3b27d in ./build/bin/nvfuser_tests)
frame #4: <unknown function> + 0x3813af (0x560a19f403af in ./build/bin/nvfuser_tests)
frame #5: void testing::internal::HandleExceptionsInMethodIfSupported<testing::Test, void>(testing::Test*, void (testing::Test::*)(), char const*) + 0x87 (0x560a1a096097 in ./build/bin/nvfuser_tests)
frame #6: testing::Test::Run() + 0xf6 (0x560a1a08a6b6 in ./build/bin/nvfuser_tests)
frame #7: <unknown function> + 0x4cb8b5 (0x560a1a08a8b5 in ./build/bin/nvfuser_tests)
frame #8: <unknown function> + 0x4cbfba (0x560a1a08afba in ./build/bin/nvfuser_tests)
frame #9: testing::internal::UnitTestImpl::RunAllTests() + 0x754 (0x560a1a08b9f4 in ./build/bin/nvfuser_tests)
frame #10: bool testing::internal::HandleExceptionsInMethodIfSupported<testing::internal::UnitTestImpl, bool>(testing::internal::UnitTestImpl*, bool (testing::internal::UnitTestImpl::*)(), char const*) + 0x87 (0x560a1a096607 in ./build/bin/nvfuser_tests)
frame #11: testing::UnitTest::Run() + 0x91 (0x560a1a08a9d1 in ./build/bin/nvfuser_tests)
frame #12: <unknown function> + 0x119882 (0x560a19cd8882 in ./build/bin/nvfuser_tests)
frame #13: <unknown function> + 0x23790 (0x7f32cec3c790 in /usr/lib/libc.so.6)
frame #14: __libc_start_main + 0x8a (0x7f32cec3c84a in /usr/lib/libc.so.6)
frame #15: _start + 0x25 (0x560a19d079a5 in ./build/bin/nvfuser_tests)
" thrown in the test body.
[  FAILED  ] LoopRotationTest.MultipleDoubleBuffer_CUDA (477 ms)