From 55035780824b5d07502c280e59f9ac9c8ad3b4b4 Mon Sep 17 00:00:00 2001 From: Isuru Fernando Date: Thu, 31 Jul 2025 07:58:20 -0500 Subject: [PATCH 1/3] Add new accelerate --- docs/maintainer/knowledge_base.md | 18 +++++++++++++----- news/2025-07-31-new-accelerate-macos.md | 16 ++++++++++++++++ 2 files changed, 29 insertions(+), 5 deletions(-) create mode 100644 news/2025-07-31-new-accelerate-macos.md diff --git a/docs/maintainer/knowledge_base.md b/docs/maintainer/knowledge_base.md index 5c8a1afaaf..88c41174fd 100644 --- a/docs/maintainer/knowledge_base.md +++ b/docs/maintainer/knowledge_base.md @@ -1295,16 +1295,23 @@ requirements: You can switch your BLAS implementation by doing, ```bash -conda install "libblas=*=*mkl" -conda install "libblas=*=*openblas" -conda install "libblas=*=*blis" -conda install "libblas=*=*accelerate" -conda install "libblas=*=*netlib" +conda install "libblas=*=*_mkl" +conda install "libblas=*=*_openblas" +conda install "libblas=*=*_blis" +conda install "libblas=*=*_accelerate" +conda install "libblas=*=*_newaccelerate" +conda install "libblas=*=*_netlib" ``` This would change the BLAS implementation without changing the conda packages depending on BLAS. +:::note + +For macOS 13.3+, you can use `newaccelerate` to use the new BLAS/LAPACK +support in Accelerate. + + The following legacy commands are also supported as well. ```bash @@ -1312,6 +1319,7 @@ conda install "blas=*=mkl" conda install "blas=*=openblas" conda install "blas=*=blis" conda install "blas=*=accelerate" +conda install "blas=*=newaccelerate" conda install "blas=*=netlib" ``` diff --git a/news/2025-07-31-new-accelerate-macos.md b/news/2025-07-31-new-accelerate-macos.md new file mode 100644 index 0000000000..42f085c31e --- /dev/null +++ b/news/2025-07-31-new-accelerate-macos.md @@ -0,0 +1,16 @@ +# New Accelerate support for macOS 13.3+ + +conda-forge by default uses `OpenBLAS` as its BLAS and LAPACK +provider on macOS as it was updated regularly and is the least +buggiest performant BLAS/LAPACK implementation. + +macOS 13.3 updated the Accelerate framework after a long time with +improved support for LAPACK APIs and has long-time known bugs in +its BLAS and LAPACK APIs. conda-forge has added support for this +new Accelerate framework by using a shim library to expose its +functionality to most conda-forge packages including `numpy`, +`scipy` and `pytorch`. + +You can use it by doing + + conda install libblas=*=*_newaccelerate From 73571c3168bb2fd99ad2659208deca18f0adfd6e Mon Sep 17 00:00:00 2001 From: "pre-commit-ci[bot]" <66853113+pre-commit-ci[bot]@users.noreply.github.com> Date: Thu, 31 Jul 2025 13:19:39 +0000 Subject: [PATCH 2/3] [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --- docs/maintainer/knowledge_base.md | 1 - 1 file changed, 1 deletion(-) diff --git a/docs/maintainer/knowledge_base.md b/docs/maintainer/knowledge_base.md index 88c41174fd..de31af4539 100644 --- a/docs/maintainer/knowledge_base.md +++ b/docs/maintainer/knowledge_base.md @@ -1311,7 +1311,6 @@ on BLAS. For macOS 13.3+, you can use `newaccelerate` to use the new BLAS/LAPACK support in Accelerate. - The following legacy commands are also supported as well. ```bash From a51fcfb6129e5a0128f0903cd533b042660170f1 Mon Sep 17 00:00:00 2001 From: Isuru Fernando Date: Tue, 5 Aug 2025 11:32:13 -0500 Subject: [PATCH 3/3] Update 2025-07-31-new-accelerate-macos.md --- news/2025-07-31-new-accelerate-macos.md | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/news/2025-07-31-new-accelerate-macos.md b/news/2025-07-31-new-accelerate-macos.md index 42f085c31e..9e7892ad70 100644 --- a/news/2025-07-31-new-accelerate-macos.md +++ b/news/2025-07-31-new-accelerate-macos.md @@ -5,8 +5,9 @@ provider on macOS as it was updated regularly and is the least buggiest performant BLAS/LAPACK implementation. macOS 13.3 updated the Accelerate framework after a long time with -improved support for LAPACK APIs and has long-time known bugs in -its BLAS and LAPACK APIs. conda-forge has added support for this +improved support for LAPACK APIs and has fixes for long-time known +bugs in the older Accelerate's BLAS and LAPACK APIs. +conda-forge has added support for this new Accelerate framework by using a shim library to expose its functionality to most conda-forge packages including `numpy`, `scipy` and `pytorch`.