Merge r2c into master #61
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…e test test_shift_axis0. All 46 CPU tests pass. The shift_axis0_core and relative helper functions perform the vertical swap (along axis0) of a rectangular matrix of size (nx, nx/2), equivalent to np.fft.fftshift(data, axes=0). I added the test test_shift_axis0. I renamed the test test_shift into test_shift_axes01 to make explicit the fact that it performs a swap on both axes.
…nx, ny), nx even. All CPU tests passes.
Only keep atomic operations. The burden of maintaining combinations is too high
The updated test passes, 19/25 tests on CPU fail and require more updates to reflect the new data format. No work done on GPU yet
Many tests fail, interpolate leads to a segfault as it expect a square image python_package/tests/test_galario.py::test_uv_idx_R2C[SP] FAILED python_package/tests/test_galario.py::test_uv_idx_R2C[DP] FAILED python_package/tests/test_galario.py::test_sample_R2C[DP_par1] FAILED python_package/tests/test_galario.py::test_sample_R2C[DP_par2] FAILED python_package/tests/test_galario.py::test_sample_R2C[DP_par3] FAILED python_package/tests/test_galario.py::test_uv_idx[SP] PASSED python_package/tests/test_galario.py::test_uv_idx[DP] PASSED python_package/tests/test_galario.py::test_interpolate[SP] PASSED python_package/tests/test_galario.py::test_interpolate[DP] PASSED python_package/tests/test_galario.py::test_FFT[SP] PASSED python_package/tests/test_galario.py::test_FFT[DP] PASSED python_package/tests/test_galario.py::test_shift_axes01[SP] PASSED python_package/tests/test_galario.py::test_shift_axes01[DP] PASSED python_package/tests/test_galario.py::test_shift_axis0[SP] PASSED python_package/tests/test_galario.py::test_shift_axis0[DP] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[SP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[DP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[SP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[DP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[SP_par3] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[DP_par3] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[SP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[DP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[SP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[DP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[SP_par3] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[DP_par3] PASSED python_package/tests/test_galario.py::test_reduce_chi2[SP] PASSED python_package/tests/test_galario.py::test_reduce_chi2[DP] PASSED python_package/tests/test_galario.py::test_loss[SP_par1] FAILED python_package/tests/test_galario.py::test_loss[DP_par1] FAILED python_package/tests/test_galario.py::test_sample[SP_par1] python/py.test.sh: line 36: 30519 Segmentation fault (core dumped) /home/beaujean/.local/miniconda3/envs/galario3/bin/python -m pytest "$@"
…tions of a few functions; Py3 xrange The CUDA version of the new shift_axis0 were not tested yet, a few bugfixes necessary. The whole functions sample_d and copy_real_to_device were present twice. Removed the additional version which was causing 'aready defined' error. Converted a few xrange to range in utils.py 44 CPU and GPU tests pass. test_uv_idx_r2c SP and DP do not pass.
`_acc_` is an old relict, now it's gone everywhere
GPU mostly untouched. Most tests even on CPU fail python_package/tests/test_galario.py::test_uv_idx_R2C[SP] PASSED python_package/tests/test_galario.py::test_uv_idx_R2C[DP] PASSED python_package/tests/test_galario.py::test_sample_R2C[DP_par1] FAILED python_package/tests/test_galario.py::test_sample_R2C[DP_par2] FAILED python_package/tests/test_galario.py::test_sample_R2C[DP_par3] FAILED python_package/tests/test_galario.py::test_uv_idx[SP] PASSED python_package/tests/test_galario.py::test_uv_idx[DP] PASSED python_package/tests/test_galario.py::test_interpolate[SP] PASSED python_package/tests/test_galario.py::test_interpolate[DP] PASSED python_package/tests/test_galario.py::test_FFT[SP] PASSED python_package/tests/test_galario.py::test_FFT[DP] PASSED python_package/tests/test_galario.py::test_shift_axes01[SP] PASSED python_package/tests/test_galario.py::test_shift_axes01[DP] PASSED python_package/tests/test_galario.py::test_shift_axis0[SP] PASSED python_package/tests/test_galario.py::test_shift_axis0[DP] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[SP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[DP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[SP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[DP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[SP_par3] PASSED python_package/tests/test_galario.py::test_apply_phase_2d[DP_par3] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[SP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[DP_par1] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[SP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[DP_par2] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[SP_par3] PASSED python_package/tests/test_galario.py::test_apply_phase_sampled[DP_par3] PASSED python_package/tests/test_galario.py::test_reduce_chi2[SP] PASSED python_package/tests/test_galario.py::test_reduce_chi2[DP] PASSED python_package/tests/test_galario.py::test_loss[SP_par1] FAILED python_package/tests/test_galario.py::test_loss[DP_par1] FAILED python_package/tests/test_galario.py::test_sample[SP_par1] python/py.test.sh: line 36: 11053 Segmentation fault /usr/bin/python -m pytest "$@"
Some open issues remain, see #57
So far CPU code was build w/o optimization :(
It's an arbitrary matrix, this should help reduce the confusion of what `nx, ny` mean. In this case, it's just the size of the matrix, `ny` is not the size of the original real array
Relates to #54
Abstraction layer for whether memory allocated by `fftw_malloc or ordinary `malloc
* openmp included even with CUDACC: there still needs to be host code * fftw only w/o CUDACC * indent definitions for better readibility
Copies the real input to a complex buffer with padding for FFTW
Not completed on GPU yet
The phase is always applied to the sampled points, there is not anymore a use case for applying the phase via 2d product of two matrices.
left over from duv and nrow renamings
Introduced change as profiling should that 1000s of calls to cudaMemcpy were much slower than the computation
32^2=1024 was the maximum. The cuda manual recommends as rule of thumb to start with 16^2=256 threads and tune from there http://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#thread-hierarchy
|
I would merge r2c to master before implementing the new feature of computing the image on the GPU. |
added 5 commits
August 12, 2017 23:38
Profiling is done via `speed_benchmark.py` these days. The failing test should be displayed on travis CI
Now all tests should pass on travis
|
I unified the three |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We fully implemented usage of the Real to Complex (R2C) Fourier transform, fundamentally addressing issue #39.
In doing so, we also:
I think we also Fix Optimize memory creation and access on GPU #40, at least its main body with the usage of
cudaMemcpy2D.The relevant commits resolving the issues are in the last comment of the issue, which will be closed merging r2c to master.