Published July 7, 2026 | Version v2.7.0
Software Open

JuliaGPU/AMDGPU.jl: v2.7.0

Authors/Creators

  • 1. University of Lausanne
  • 2. Advanced Micro Devices
  • 3. CERMICS, ENPC
  • 4. @JuliaComputing
  • 5. @AMD
  • 6. Curtin Institute of Data Science
  • 7. RUB University
  • 8. @mit
  • 9. @atdepth
  • 10. Argonne National Laboratory
  • 11. @UCL-ARC
  • 12. @dwavesystems
  • 13. UIUC/Google Deepmind
  • 14. DLR

Description

AMDGPU v2.7.0

Diff since v2.6.0

Merged pull requests:

  • Improve discovery robustness (#934) (@luraess)
  • Add eigensolver support for Hermitian and Symmetric (#945) (@luraess)
  • Fix scalar multiplication for ROCSparseMatrix adjoint/transpose wrappers. (#946) (@luraess)
  • Add mixed real/complex sparse-dense matrix multiplication (#947) (@luraess)
  • Improve CSCS CI (#956) (@luraess)
  • Improve documentation (#959) (@luraess)
  • Add svd/svdvals support for ROCArray (#960) (@luraess)

Closed issues:

  • SVD not consistently working with ROCArrays (#837)
  • Scalar indexing with multiplication between a scalar and the adjoint of ROC sparse matrices (#854)
  • Generalized eigensolver for Hermitian matrices fails (#903)
  • ROCSparseMatrxCSC{<:Complex}*ROCMatrix{<:Real} triggers allowscalar error (#904)
  • Improve MI300 CI (#948)

Files

JuliaGPU/AMDGPU.jl-v2.7.0.zip

Files (855.0 kB)

Name Size Download all
md5:2ab33820e882afe5196e88bb8082a6b5
855.0 kB Preview Download

Additional details

Related works

Is supplement to
Software: https://github.com/JuliaGPU/AMDGPU.jl/tree/v2.7.0 (URL)

Software

Repository URL
https://github.com/JuliaGPU/AMDGPU.jl