Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Prebuild the benchmark binary with -O3 on Android #532

Merged
merged 2 commits into from
Oct 14, 2020
Merged

Prebuild the benchmark binary with -O3 on Android #532

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
590f8ef
Prebuild the benchmark binary with -O3 on Android
lgeiger Oct 14, 2020
97d3db1
Remove failing debug check
lgeiger Oct 14, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions .bazelrc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,11 +23,11 @@ build:windows --host_cxxopt=/std:c++14
# These can be activated using --config=rpi3 and --config=aarch64
build:rpi3 --crosstool_top=@local_config_arm_compiler//:toolchain
build:rpi3 --cpu=armeabi
build:rpi3 -c opt --copt=-march=armv7-a --copt=-mfpu=neon-vfpv4 --copt=-std=gnu++11 --copt=-DS_IREAD=S_IRUSR --copt=-DS_IWRITE=S_IWUSR --copt=-O3 --copt=-fno-tree-pre --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_1 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_2 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 --define=raspberry_pi_with_neon=true --define=framework_shared_object=false --copt=-funsafe-math-optimizations --copt=-ftree-vectorize --copt=-fomit-frame-pointer --verbose_failures
build:rpi3 --copt=-march=armv7-a --copt=-mfpu=neon-vfpv4 --copt=-std=gnu++11 --copt=-DS_IREAD=S_IRUSR --copt=-DS_IWRITE=S_IWUSR --copt=-fno-tree-pre --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_1 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_2 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 --define=raspberry_pi_with_neon=true --define=framework_shared_object=false --copt=-funsafe-math-optimizations --copt=-ftree-vectorize --copt=-fomit-frame-pointer --verbose_failures

build:aarch64 --crosstool_top=@local_config_arm_compiler//:toolchain
build:aarch64 --cpu=aarch64
build:aarch64 -c opt --copt=-march=armv8-a --copt=-std=gnu++11 --copt=-DS_IREAD=S_IRUSR --copt=-DS_IWRITE=S_IWUSR --copt=-O3 --copt=-fno-tree-pre --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_1 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_2 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 --define=framework_shared_object=false --copt=-funsafe-math-optimizations --copt=-ftree-vectorize --copt=-fomit-frame-pointer --verbose_failures
build:aarch64 --copt=-march=armv8-a --copt=-std=gnu++11 --copt=-DS_IREAD=S_IRUSR --copt=-DS_IWRITE=S_IWUSR --copt=-fno-tree-pre --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_1 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_2 --copt=-U__GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 --define=framework_shared_object=false --copt=-funsafe-math-optimizations --copt=-ftree-vectorize --copt=-fomit-frame-pointer --verbose_failures

# Options to build TensorFlow 1.x or 2.x.
build:v1 --define=tf_api_version=1
Expand Down
6 changes: 3 additions & 3 deletions .github/workflows/release.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,15 +37,15 @@ jobs:
- run: mkdir benchmark-binaries
- name: Build Benchmark utility for AArch64
run: |
bazelisk build //larq_compute_engine/tflite/benchmark:lce_benchmark_model -c opt --config=aarch64
bazelisk build //larq_compute_engine/tflite/benchmark:lce_benchmark_model --config=aarch64 -c opt --copt=-O3
cp bazel-bin/larq_compute_engine/tflite/benchmark/lce_benchmark_model benchmark-binaries/lce_benchmark_model_aarch64
- name: Build Benchmark utility for AArch32
run: |
bazelisk build //larq_compute_engine/tflite/benchmark:lce_benchmark_model -c opt --config=rpi3
bazelisk build //larq_compute_engine/tflite/benchmark:lce_benchmark_model --config=rpi3 -c opt --copt=-O3
cp bazel-bin/larq_compute_engine/tflite/benchmark/lce_benchmark_model benchmark-binaries/lce_benchmark_model_aarch32
- name: Build Benchmark utility for Android
run: |
bazelisk build //larq_compute_engine/tflite/benchmark:lce_benchmark_model -c opt --config=android_arm64
bazelisk build //larq_compute_engine/tflite/benchmark:lce_benchmark_model --config=android_arm64 -c opt --copt=-O3
cp bazel-bin/larq_compute_engine/tflite/benchmark/lce_benchmark_model benchmark-binaries/lce_benchmark_model_android_arm64
- uses: actions/upload-artifact@v2.2.0
with:
Expand Down
4 changes: 0 additions & 4 deletions larq_compute_engine/core/bgemm/ruy_pack.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,10 +45,6 @@ struct LceRuyPackImpl<ThePath, FixedKernelLayout<Order::kColMajor, 4, 4>,
PMat<TBitpacked>* packed_matrix, int start_col, int end_col) {
profiler::ScopeLabel label("Pack (ColMajor, 4x4)");

// Ruy supports collecting column sums during the packing process, which we
// have no need for.
RUY_DCHECK_EQ(packed_matrix->sums, nullptr);

// Likewise, Ruy supports arbitrary zero points, but we only use true-zero.
RUY_DCHECK_EQ(src_matrix.zero_point, 0);

Expand Down