Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Feb 15, 2025

• Merge pull request #41 from markushuelsbusch/master

  Show description for 3830ec9
  vincefn

  authored
  3830ec9

  Copy full SHA for 3830ec9

• Update CHANGELOG.rst

  vincefn

  authored
  1b92be6

  Copy full SHA for 1b92be6

Commits on Feb 13, 2025

• change numpy dtype specifier to be compatible with width of c_long on all platforms

  markushuelsbusch

  committed
  6818b29

  Copy full SHA for 6818b29

Commits on Oct 14, 2024

• Further fixes for pyvkfft-benchmark: handle importing sql files from older versions and plotting for cupy/skcuda/gpyfft tests

  vincefn

  committed
  f828848

  Copy full SHA for f828848

Commits on Oct 11, 2024

• Include tuning parameters in benchmark sql filename.

  vincefn

  committed
  5dfc5f2

  Copy full SHA for 5dfc5f2

• Include transform in benchmark legend and output filename

  vincefn

  committed
  6b57e0b

  Copy full SHA for 6b57e0b

Commits on Sep 30, 2024

• Bump VkFFT version

  vincefn

  committed
  2e954f6

  Copy full SHA for 2e954f6

Commits on Sep 26, 2024

• Remove numexpr import (??)

  vincefn

  committed
  8e604e5

  Copy full SHA for 8e604e5

• Update benchmark to allow testing pyfftw and scipy. Bump version to 2024.2b0

  vincefn

  committed
  8c8adda

  Copy full SHA for 8c8adda

• Update VkFFT for pre-1.3.5 version (#39), with updated VkFFTLaunchParams

  vincefn

  committed
  63c25f3

  Copy full SHA for 63c25f3

Commits on Jul 23, 2024

• Version 2024.1.4

  vincefn

  committed
  fd91c92

  Copy full SHA for fd91c92

Commits on Jul 11, 2024

• Re-activate dct an dst tests

  vincefn

  committed
  6c41dd0

  Copy full SHA for 6c41dd0

• Add (some) support for numpy 2.0 (#35)

  Show description for fa9691d
  alexfikl

  authored
  fa9691d

  Copy full SHA for fa9691d

Commits on May 1, 2024

• Handle arrays where axes have a length equal to 1, notably handling the complicated cases of R2C where the fast axis must be known, even when only one axis has a length>1 [https://github.com/vincef…

  vincefn

  committed
  8d5b584

  Copy full SHA for 8d5b584

Commits on Mar 19, 2024

• Fix nvcc search in setup.py without a CUDA_HOME type environment variable under linux.

  vincefn

  committed
  6f5cac1

  Copy full SHA for 6f5cac1

Commits on Feb 16, 2024

• Vresion 2024.1.2

  vincefn

  committed
  e310e16

  Copy full SHA for e310e16

• Version 2024.1.1 ; Update documentation regarding conda installation using `cuda-version` to select the cuda nvrtc or cuda-toolkit version

  vincefn

  committed
  65458e0

  Copy full SHA for 65458e0

• Update tests for cuda and opencl libx

  vincefn

  committed
  bd0497f

  Copy full SHA for bd0497f

• Add basic import tests for the cuda and opencl compiled librairies, for tests during builds

  vincefn

  committed
  7166226

  Copy full SHA for 7166226

Commits on Feb 13, 2024

• *correctly* fix the default cuda context init in pycuda accuracy tests...

  vincefn

  committed
  5d4b515

  Copy full SHA for 5d4b515

Commits on Feb 12, 2024

• Fix pycuda initialisation during accuracy tests (pyvkff-test)

  vincefn

  committed
  ac4b0b4

  Copy full SHA for ac4b0b4

Commits on Feb 6, 2024

• Update VkFFT subproject version

  vincefn

  committed
  614032b

  Copy full SHA for 614032b

• pyvkfft release 2024.1

  vincefn

  committed
  35de35c

  Copy full SHA for 35de35c

Commits on Feb 3, 2024

• update install doc formatting

  vincefn

  committed
  479947c

  Copy full SHA for 479947c

• Update setup.py for conda environments (e.g. windows). Update install doc.

  vincefn

  committed
  430a33b

  Copy full SHA for 430a33b

Commits on Jan 29, 2024

• Update VkFFT version

  vincefn

  committed
  d79a784

  Copy full SHA for d79a784

Commits on Jan 28, 2024

• Benchmark: correct nloop

  vincefn

  committed
  8259548

  Copy full SHA for 8259548

• Accuray: use pycuda's device.retain_primary_context() rather than create a new context

  vincefn

  committed
  85bb7fb

  Copy full SHA for 85bb7fb

• Benchmark script: add --nbatch parameter, be verbose by default

  vincefn

  committed
  2ddbecf

  Copy full SHA for 2ddbecf

Commits on Jan 27, 2024

• Update benchmark plot, correctly differentiating radix, rader and Bluestein

  vincefn

  committed
  22aa886

  Copy full SHA for 22aa886

Commits on Jan 24, 2024

• tune_vkfft: correct dest shape when auto-allocated during out-of-place tuning

  vincefn

  committed
  8a76138

  Copy full SHA for 8a76138

• Increase timeout for benchmark (useful when tuning)

  vincefn

  committed
  9c3ed43

  Copy full SHA for 9c3ed43

• Merge branch 'master' of github.com:vincefn/pyvkfft

  Show description for 4a8c62d
  vincefn

  committed
  4a8c62d

  Copy full SHA for 4a8c62d

• Correct _bench_pyvkfft_cuda

  vincefn

  committed
  df04521

  Copy full SHA for df04521

• Correct skcuda benchmark

  vincefn

  committed
  9b01f51

  Copy full SHA for 9b01f51

Pagination

Previous

Next