0.3.33

Issues

• Work around miscompilation of the AVX512 ?GEMM kernels by Windows LLVM

  Unread
  OpenMathLib/OpenBLAS#5756·by martin-frbg was merged Apr 15, 2026

• Improve performance on edges of GEMM for RISC-V

  Unread
  OpenMathLib/OpenBLAS#5674·by ChipKerchner was merged Apr 11, 2026
  27

• Work around miscompilation of the ARM64 non-SVE DDOT kernel

  Unread
  OpenMathLib/OpenBLAS#5710·by martin-frbg was merged Mar 27, 2026

• c_check: loongarch64: Fix typo

  Unread
  OpenMathLib/OpenBLAS#5709·by iv-m was merged Mar 24, 2026
  2

• Fix incorrect cast from BF16 to FP32 in SBGEMM

  Unread
  OpenMathLib/OpenBLAS#5712·by murste01 was merged Apr 24, 2026
  9

• typedef the unsupported fp16 as bfloat16 on Loongarch64 too

  Unread
  OpenMathLib/OpenBLAS#5715·by martin-frbg was merged Mar 27, 2026
  1

• [ARM64] Add optimized fp16 shgemm kernels for Neoverse N2

  Unread
  OpenMathLib/OpenBLAS#5716·by yuanjia111 was merged Mar 27, 2026
  1

• Fix CMake DYNAMIC_ARCH builds under Windows on Arm

  Unread
  OpenMathLib/OpenBLAS#5718·by martin-frbg was merged Mar 27, 2026

• ARM64 DYNAMIC_ARCH: add CortexA75/76 and restore VORTEX for DYNAMIC_LIST

  Unread
  OpenMathLib/OpenBLAS#5719·by martin-frbg was merged Mar 30, 2026

• Implement ?LARF1F and ?ORM2R (Reference-LAPACK PRs 1019/1020/1196)

  Unread
  OpenMathLib/OpenBLAS#5721·by martin-frbg was merged Mar 30, 2026

• Change loop order in ?GETC2 (Reference-LAPACK PR 1023)

  Unread
  OpenMathLib/OpenBLAS#5722·by martin-frbg was merged Mar 31, 2026

• Change WORK dimension in deprecated ?GELQS/?GEQRS (Reference-LAPACK PR 1094)

  Unread
  OpenMathLib/OpenBLAS#5723·by martin-frbg was merged Mar 31, 2026

• Add NaN checks for input matrix A in ?GEEV (Reference-LAPACK PR 1136)

  Unread
  OpenMathLib/OpenBLAS#5724·by martin-frbg was merged Mar 31, 2026

• Fix support for jobu/v in LAPACKE_?GESVDQ_WORK (Reference-LAPACK PR 1146)

  Unread
  OpenMathLib/OpenBLAS#5725·by martin-frbg was merged Mar 31, 2026

• Fix display of version number in LAPACK tests (Reference-LAPACK PR 1149)

  Unread
  OpenMathLib/OpenBLAS#5726·by martin-frbg was merged Mar 31, 2026

• Fix DGGES test seed to avoid bad matrix (Reference-LAPACK PR 1187)

  Unread
  OpenMathLib/OpenBLAS#5727·by martin-frbg was merged Mar 31, 2026

• Fix truncation of large workspace values in ZHE routines (Reference-LAPACK PR 1195)

  Unread
  OpenMathLib/OpenBLAS#5729·by martin-frbg was merged Mar 31, 2026

• Fix overwriting of LDSWORK in ?TRSYL3 (Reference-LAPACK PR 1206)

  Unread
  OpenMathLib/OpenBLAS#5730·by martin-frbg was merged Mar 31, 2026

• Fix crossover of INFO variables in some EIG tests (Reference-LAPACK PR 1207)

  Unread
  OpenMathLib/OpenBLAS#5731·by martin-frbg was merged Mar 31, 2026

• Remove unused parameter in DORBDB3/ZUNBDB3 (Reference-LAPACK PR 1209)

  Unread
  OpenMathLib/OpenBLAS#5732·by martin-frbg was merged Mar 31, 2026

• Re-enable testing of the ?BB/?GG driver routines (Reference-LAPACK PR 1211)

  Unread
  OpenMathLib/OpenBLAS#5733·by martin-frbg was merged Mar 31, 2026

• Fix workspace size in ?TGSEN (Reference-LAPACK PR 774)

  Unread
  OpenMathLib/OpenBLAS#5734·by martin-frbg was merged Apr 1, 2026

• Follow-up on ?GESVDQ updates from PR1146 (Reference-LAPACK PR 1221)

  Unread
  OpenMathLib/OpenBLAS#5736·by martin-frbg was merged Apr 1, 2026

• Fix out-of-bounds access in the converted C version of the CBLAS tests

  Unread
  OpenMathLib/OpenBLAS#5740·by martin-frbg was merged Apr 2, 2026

• Add note on using an x86 OpenBLAS in Windows on Arm via Prism

  Unread
  OpenMathLib/OpenBLAS#5741·by martin-frbg was merged Apr 2, 2026