Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Inactive

Commits : Listings

Analyzed 5 days ago. based on code collected 5 days ago.
Mar 30, 2025 — Mar 30, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
changed fftconvolve.py to set DATA_W, DATA_H from input data trivial changes to fftlab.py so nico can play around
Justin Riley
More... almost 17 years ago
Merge branch 'master' of [email protected]:npinto/python-cuda
thor
More... almost 17 years ago
Merge branch 'master' of [email protected]:npinto/python-cuda
thor
More... almost 17 years ago
Merge branch 'master' of [email protected]:npinto/python-cuda
thor
More... almost 17 years ago
cleaning up
thor
More... almost 17 years ago
cleaning up
thor
More... almost 17 years ago
cleaning up
thor
More... almost 17 years ago
cleaning up
thor
More... almost 17 years ago
added fftconvolve to cuda/sugar/fft/__init__.py trivial changes to conv_gold.py (this file will eventually go away) modified fftconvolve.py main() method to be fftconvolve2d() method fftlab now visualizes input data, kernel, cpu convolution, and gpu convolution
Justin Riley
More... almost 17 years ago
added fftconvolve to cuda/sugar/fft/__init__.py trivial changes to conv_gold.py (this file will eventually go away) modified fftconvolve.py main() method to be fftconvolve2d() method fftlab now visualizes input data, kernel, cpu convolution, and gpu convolution
Justin Riley
More... almost 17 years ago
modified setup.py to reflect latest package list added visualization tool fftlab.py for cuda.cufft/cuda.sugar.fft in tests/cufft for now
Justin Riley
More... almost 17 years ago
modified setup.py to reflect latest package list added visualization tool fftlab.py for cuda.cufft/cuda.sugar.fft in tests/cufft for now
Justin Riley
More... almost 17 years ago
fftconvolve.py is now a fully working python-cuda equivalent of CUDA SDK convolutionFFT2D the cuda sdk cpu reference calculation is used for now, need to switch this to scipy.signal.fftconvolve still need to clean all of the fft code and wrap a decent python/numpy API around it trivial edits to conv_gold.py - this file will disappear once I've figured out how to compare between scipy's fftconvolve results and CUDA SDK fftconvolve
Justin Riley
More... almost 17 years ago
fftconvolve.py is now a fully working python-cuda equivalent of CUDA SDK convolutionFFT2D the cuda sdk cpu reference calculation is used for now, need to switch this to scipy.signal.fftconvolve still need to clean all of the fft code and wrap a decent python/numpy API around it trivial edits to conv_gold.py - this file will disappear once I've figured out how to compare between scipy's fftconvolve results and CUDA SDK fftconvolve
Justin Riley
More... almost 17 years ago
started moving reference CPU methods for convFFT2D to conv_gold.py
Justin Riley
More... almost 17 years ago
started moving reference CPU methods for convFFT2D to conv_gold.py
Justin Riley
More... almost 17 years ago
temporarily added conv_gold.py for testing SDK CPU results to scipy/numpy minor cleanup/improvements to fftconvolve, still not producing correct results yet
Justin Riley
More... almost 17 years ago
temporarily added conv_gold.py for testing SDK CPU results to scipy/numpy minor cleanup/improvements to fftconvolve, still not producing correct results yet
Justin Riley
More... almost 17 years ago
added vim swp files to gitignore
Justin Riley
More... almost 17 years ago
added vim swp files to gitignore
Justin Riley
More... almost 17 years ago
ITS ALIVE! fftconvolve.py now gives results, still need to verify results the errors I was getting when calling cufftExecC2C were related to texture binding turns out I was calling cudaBindTextureToArray with a null channel descriptor needed to get channelDescriptor corresponding to a particular cudaArray by using cudaGetChannelDesc minor edits to fft.py
Justin Riley
More... almost 17 years ago
ITS ALIVE! fftconvolve.py now gives results, still need to verify results the errors I was getting when calling cufftExecC2C were related to texture binding turns out I was calling cudaBindTextureToArray with a null channel descriptor needed to get channelDescriptor corresponding to a particular cudaArray by using cudaGetChannelDesc minor edits to fft.py
Justin Riley
More... almost 17 years ago
removed unnecessary init methods for complex arrays. Now uses np.ctypes.data
Justin Riley
More... almost 17 years ago
removed unnecessary init methods for complex arrays. Now uses np.ctypes.data
Justin Riley
More... almost 17 years ago
removed unnecessary 1d/2d/3d initialization functions for complex arrays (use np.ctypes.data instead) added sugar to cuda/__init__.py imports
Justin Riley
More... almost 17 years ago
removed unnecessary 1d/2d/3d initialization functions for complex arrays (use np.ctypes.data instead) added sugar to cuda/__init__.py imports
Justin Riley
More... almost 17 years ago
fftconvolve.py now finishes but with errors...more investigating to come
Justin Riley
More... almost 17 years ago
fftconvolve.py now finishes but with errors...more investigating to come
Justin Riley
More... almost 17 years ago
Switched commands.getoutput back to subprocess since subprocess is the new module for running system commands and has a nice cwd keyword to specify the process to run in a separate directory fftconvolve.py is now fully python and im in a run/debug/fix cycle until it works removed all files in testbed since kernel loading is mostly taken care of. imported fftconvolve2d_kernel.cu into cuda/sugar/fft package for fftconvolve.py
Justin Riley
More... almost 17 years ago
Switched commands.getoutput back to subprocess since subprocess is the new module for running system commands and has a nice cwd keyword to specify the process to run in a separate directory fftconvolve.py is now fully python and im in a run/debug/fix cycle until it works removed all files in testbed since kernel loading is mostly taken care of. imported fftconvolve2d_kernel.cu into cuda/sugar/fft package for fftconvolve.py
Justin Riley
More... almost 17 years ago