6 Pull requests merged by 4 people

• Add dummy implementations of make_complex_(float/double) to simplify Windows DLL linking

  #5178 merged Mar 13, 2025

• Fix omission of LAPACKE interfaces for cgesvdq,strsyl3 and deprecated functions in CMAKE builds

  #5177 merged Mar 13, 2025

• Improving Load Imbalance in Thread-Parallel GEMM

  #5173 merged Mar 12, 2025

• Optimize gemv_n_sve kernel

  #5157 merged Mar 12, 2025

• Fix aarch64 sbgemv_t compilation error for GCC < 13

  #5176 merged Mar 12, 2025

• Improve thread safety of pthreads builds that rely on C11 atomic operations for locking

  #5170 merged Mar 7, 2025

4 Pull requests opened by 4 people

• Add thread throttling profile for DGEMV on NEOVERSEV1

  #5175 opened Mar 12, 2025

• CMake: Pass `OpenMP` compiler and linker flags through CMake targets

  #5180 opened Mar 13, 2025

• replace customize bf16_to_fp32 with arm neon vcvtah_f32_bf16

  #5181 opened Mar 13, 2025

• Optimize aarch64 sgemm_ncopy

  #5182 opened Mar 13, 2025

3 Issues closed by 1 person

• The performance curve of parallel GEMM with many cores shows significant up-down

  #5172 closed Mar 12, 2025

• [regression] 0.3.29 build/tests fail on Sandy Bridge x86_64 machine

  #5147 closed Mar 7, 2025

• [regression] 0.3.29 test failure on ARMV6

  #5104 closed Mar 7, 2025

1 Issue opened by 1 person

• Kaldi Makefile:154: *** OpenBLAS: Detecting CPU failed. Please set TARGET explicitly

  #5179 opened Mar 13, 2025

2 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• perf report shows most cycles spent in blas_thread_server

  #5171 commented on Mar 7, 2025 • 0 new comments

• Prefix Support Fails During OpenBLAS Build on Windows (WoA)

  #5154 commented on Mar 12, 2025 • 0 new comments