diff options
author | Pablo Marquez Tello <pablo.tello@arm.com> | 2022-07-19 12:19:46 +0100 |
---|---|---|
committer | Pablo Marquez Tello <pablo.tello@arm.com> | 2022-07-25 15:56:00 +0000 |
commit | d208f4f1c2da9e407f86022959c32e8ab9a4aa6e (patch) | |
tree | f1dc538edb68fa6e266ad8890582f179b961bfae /src/cpu/kernels | |
parent | 962531886eee29527bf027107124a27ce94047eb (diff) | |
download | ComputeLibrary-d208f4f1c2da9e407f86022959c32e8ab9a4aa6e.tar.gz |
Enable march=armv8.6-a in non multi-isa builds
* scons arch=armv8.6-a translates to -march=armv8.6-a
* scons arch=armv8.6-a-sve translates to -march=armv8.6-a+sve
* scons arch=armv8.6-a-sve2 translates to -march=armv8.6-a+sve2
* Resolves COMPMID-5408
Change-Id: I0901e1de864d00109759509af7cc2b5c9ae1cd75
Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7943
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu/kernels')
-rw-r--r-- | src/cpu/kernels/CpuIm2ColKernel.cpp | 10 | ||||
-rw-r--r-- | src/cpu/kernels/cast/generic/neon/bfloat16.cpp | 4 |
2 files changed, 7 insertions, 7 deletions
diff --git a/src/cpu/kernels/CpuIm2ColKernel.cpp b/src/cpu/kernels/CpuIm2ColKernel.cpp index 875d66594f..25ff6c291c 100644 --- a/src/cpu/kernels/CpuIm2ColKernel.cpp +++ b/src/cpu/kernels/CpuIm2ColKernel.cpp @@ -359,11 +359,11 @@ void CpuIm2ColKernel::configure(const ITensorInfo *src, ITensorInfo *dst, const case DataType::F32: _func = (!conv_info.has_padding()) ? &CpuIm2ColKernel::run_im2col<float, false, true> : &CpuIm2ColKernel::run_im2col<float, true, true>; break; -#if defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) +#if defined(ARM_COMPUTE_ENABLE_BF16) case DataType::BFLOAT16: _func = (!conv_info.has_padding()) ? &CpuIm2ColKernel::run_im2col<bfloat16, false, true> : &CpuIm2ColKernel::run_im2col<bfloat16, true, true>; break; -#endif /* defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) */ +#endif /* defined(ARM_COMPUTE_ENABLE_BF16) */ #ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC case DataType::F16: _func = (!conv_info.has_padding()) ? &CpuIm2ColKernel::run_im2col<float16_t, false, true> : &CpuIm2ColKernel::run_im2col<float16_t, true, true>; @@ -385,11 +385,11 @@ void CpuIm2ColKernel::configure(const ITensorInfo *src, ITensorInfo *dst, const case DataType::F32: _func = (!conv_info.has_padding()) ? &CpuIm2ColKernel::run_im2col<float, false, false> : &CpuIm2ColKernel::run_im2col<float, true, false>; break; -#if defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) +#if defined(ARM_COMPUTE_ENABLE_BF16) case DataType::BFLOAT16: _func = (!conv_info.has_padding()) ? &CpuIm2ColKernel::run_im2col<bfloat16, false, false> : &CpuIm2ColKernel::run_im2col<bfloat16, true, false>; break; -#endif /* defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) */ +#endif /* defined(ARM_COMPUTE_ENABLE_BF16) */ #ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC case DataType::F16: _func = (!conv_info.has_padding()) ? &CpuIm2ColKernel::run_im2col<float16_t, false, false> : &CpuIm2ColKernel::run_im2col<float16_t, true, false>; @@ -453,4 +453,4 @@ size_t CpuIm2ColKernel::get_mws(const CPUInfo &platform, size_t thread_count) co } } // namespace kernels } // namespace cpu -} // namespace arm_compute
\ No newline at end of file +} // namespace arm_compute diff --git a/src/cpu/kernels/cast/generic/neon/bfloat16.cpp b/src/cpu/kernels/cast/generic/neon/bfloat16.cpp index aac4ef4ca0..eed537039f 100644 --- a/src/cpu/kernels/cast/generic/neon/bfloat16.cpp +++ b/src/cpu/kernels/cast/generic/neon/bfloat16.cpp @@ -21,7 +21,7 @@ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE * SOFTWARE. */ -#if defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) +#if defined(ARM_COMPUTE_ENABLE_BF16) #include "arm_compute/core/TensorInfo.h" #include "src/core/NEON/wrapper/wrapper.h" @@ -142,4 +142,4 @@ void neon_bfloat16_to_fp32_cast(const ITensor *_src, ITensor *_dst, const Thread } // namespace cpu } // namespace arm_compute -#endif /* defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) */ +#endif /* defined(ARM_COMPUTE_ENABLE_BF16) */ |