Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

CUDA 11 Test: `cupy_tests.fft_tests.test_fft.TestMultiGpu`

See original GitHub issue

FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_2_{n=None, norm=None, shape=(64,)}::test_fft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_2_{n=None, norm=None, shape=(64,)}::test_ifft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_6_{n=None, norm='ortho', shape=(64,)}::test_fft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_6_{n=None, norm='ortho', shape=(64,)}::test_ifft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_24_{n=64, norm=None, shape=(0,)}::test_fft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_24_{n=64, norm=None, shape=(0,)}::test_ifft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_26_{n=64, norm=None, shape=(64,)}::test_fft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_26_{n=64, norm=None, shape=(64,)}::test_ifft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_28_{n=64, norm='ortho', shape=(0,)}::test_fft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_28_{n=64, norm='ortho', shape=(0,)}::test_ifft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_30_{n=64, norm='ortho', shape=(64,)}::test_fft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuFft_param_30_{n=64, norm='ortho', shape=(64,)}::test_ifft - AssertionError: Parameterized test failed.
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_0_{n=None, norm=None, shape=(64,)}::test_fft - AssertionError: Parameterized test f...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_0_{n=None, norm=None, shape=(64,)}::test_ifft - AssertionError: Parameterized test ...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_1_{n=None, norm=None, shape=(128,)}::test_fft - AssertionError: Parameterized test ...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_1_{n=None, norm=None, shape=(128,)}::test_ifft - AssertionError: Parameterized test...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_2_{n=None, norm='ortho', shape=(64,)}::test_fft - AssertionError: Parameterized tes...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_2_{n=None, norm='ortho', shape=(64,)}::test_ifft - AssertionError: Parameterized te...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_3_{n=None, norm='ortho', shape=(128,)}::test_fft - AssertionError: Parameterized te...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_3_{n=None, norm='ortho', shape=(128,)}::test_ifft - AssertionError: Parameterized t...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_4_{n=64, norm=None, shape=(64,)}::test_fft - AssertionError: Parameterized test fai...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_4_{n=64, norm=None, shape=(64,)}::test_ifft - AssertionError: Parameterized test fa...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_5_{n=64, norm=None, shape=(128,)}::test_fft - AssertionError: Parameterized test fa...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_5_{n=64, norm=None, shape=(128,)}::test_ifft - AssertionError: Parameterized test f...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_6_{n=64, norm='ortho', shape=(64,)}::test_fft - AssertionError: Parameterized test ...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_6_{n=64, norm='ortho', shape=(64,)}::test_ifft - AssertionError: Parameterized test...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_7_{n=64, norm='ortho', shape=(128,)}::test_fft - AssertionError: Parameterized test...
FAILED tests/cupy_tests/fft_tests/test_fft.py::TestMultiGpuPlanCtxManagerFft_param_7_{n=64, norm='ortho', shape=(128,)}::test_ifft - AssertionError: Parameterized tes...

Issue Analytics

State:
Created 3 years ago
Reactions:1
Comments:20 (20 by maintainers)

Top GitHub Comments

3reactions

kmaehashicommented, Aug 11, 2020

Multi-gpu cuFFT failing with CUFFT_EXEC_FAILED for complex64.

cc/ @pentschev

2reactions

pentschevcommented, Oct 27, 2020

For future reference: I was provided with a preview build of cuFFT 11.2 and was able to verify that the issue is indeed fixed in that version, the condition in https://github.com/cupy/cupy/pull/3775/files#diff-ccb51faf00a37e7d1236eefcdef3d78aaba921baeca21e7405169a4c43d9398bR169 should then be fine and no other action should be needed on CuPy’s end. Thanks everyone!

Top Results From Across the Web

CUDA Demo Suite - NVIDIA Documentation Center

The CUDA Demo Suite contains pre-built applications which use CUDA. ... Test the bandwidth for device to host, host to device, and device...

Request for testing: CUDA from BinaryBuilder - Julia Discourse

Hi all, We are looking into using BinaryBuilder for installing CUDA when you use the Julia/CUDA stack, i.e. CuArrays or CUDAnative. The support...

Getting Started - RAPIDS.ai

RAPIDS uses optimized NVIDIA CUDA® primitives and high-bandwidth GPU memory to ... a basic XGBoost model training and testing workflow runs with RAPIDS....

Amber (PMEMD) GPU Support

Installation and Testing. The single GPU version of PMEMD is called pmemd.cuda while the multi-GPU version is called pmemd.cuda.MPI. These are built ...

Enabling GPU acceleration on Ubuntu on WSL2 with the ...

Enabling GPU acceleration on Ubuntu on WSL2 with the NVIDIA CUDA Platform ... support for GPU-accelerated AI/ML training and the ability to develop...