Welcome to mirror list, hosted at ThFree Co, Russian Federation.

github.com/marian-nmt/marian.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorNikolay Bogoychev <nheart@gmail.com>2021-12-16 18:07:34 +0300
committerGitHub <noreply@github.com>2021-12-16 18:07:34 +0300
commite26e5b6faf7c826868efe12316e0b02499d4edcf (patch)
treedf0201120d217e8bd7098796a4bce348a4261298
parente8a1a2530fb84cbff7383302ebca393e5875c441 (diff)
Use apple accelerate on MacOs by default (#897)
-rw-r--r--.github/workflows/macos.yml11
-rw-r--r--CHANGELOG.md1
-rw-r--r--CMakeLists.txt6
3 files changed, 8 insertions, 10 deletions
diff --git a/.github/workflows/macos.yml b/.github/workflows/macos.yml
index 5e3e57c6..20907d9b 100644
--- a/.github/workflows/macos.yml
+++ b/.github/workflows/macos.yml
@@ -18,16 +18,10 @@ jobs:
submodules: recursive
- name: Install dependencies
- run: brew install boost openblas openssl protobuf
+ run: brew install boost openssl protobuf
- # Openblas location is exported explicitly because openblas is keg-only,
- # which means it was not symlinked into /usr/local/.
- # CMake cannot find BLAS on GitHub runners if Marian is being compiled
- # statically, hence USE_STATIC_LIBS=off
- name: Configure CMake
run: |
- export LDFLAGS="-L/usr/local/opt/openblas/lib"
- export CPPFLAGS="-I/usr/local/opt/openblas/include"
mkdir -p build
cd build
cmake .. \
@@ -37,8 +31,7 @@ jobs:
-DCOMPILE_SERVER=on \
-DCOMPILE_TESTS=on \
-DUSE_FBGEMM=on \
- -DUSE_SENTENCEPIECE=on \
- -DUSE_STATIC_LIBS=off
+ -DUSE_SENTENCEPIECE=on
- name: Compile
working-directory: build
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 80715ad3..a5dd305f 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -54,6 +54,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
- Enforce validation for the task alias in training mode.
### Changed
+- MacOS marian uses Apple Accelerate framework by default, as opposed to openblas/mkl.
- Optimize LSH for speed by treating is as a shortlist generator. No option changes in decoder
- Set REQUIRED_BIAS_ALIGNMENT = 16 in tensors/gpu/prod.cpp to avoid memory-misalignment on certain Ampere GPUs.
- For BUILD_ARCH != native enable all intrinsics types by default, can be disabled like this: -DCOMPILE_AVX512=off
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 4e6f24c7..eb6ca97b 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -16,7 +16,11 @@ option(COMPILE_CUDA "Compile GPU version" ON)
option(COMPILE_EXAMPLES "Compile examples" OFF)
option(COMPILE_SERVER "Compile marian-server" OFF)
option(COMPILE_TESTS "Compile tests" OFF)
-option(USE_APPLE_ACCELERATE "Compile with Apple Accelerate" OFF)
+if(APPLE)
+ option(USE_APPLE_ACCELERATE "Compile with Apple Accelerate" ON)
+else(APPLE)
+ option(USE_APPLE_ACCELERATE "Compile with Apple Accelerate" OFF)
+endif(APPLE)
option(USE_CCACHE "Use ccache compiler cache (https://ccache.dev)" OFF)
option(USE_CUDNN "Use CUDNN library" OFF)
option(USE_DOXYGEN "Build documentation with Doxygen" ON)