diff options
| author | Pablo Marquez Tello <pablo.tello@arm.com> | 2023-12-12 10:09:59 +0000 |
|---|---|---|
| committer | Pablo Marquez Tello <pablo.tello@arm.com> | 2023-12-12 14:36:16 +0000 |
| commit | b526431b1079a86b30a1bf6e905435e7121584f9 (patch) | |
| tree | 09956b6447fac37957fe491e78dd4e8d190246ab /src/cpu | |
| parent | 391f872c9fd9831cc2298a2a27eccfc21d727cb8 (diff) | |
| download | ComputeLibrary-b526431b1079a86b30a1bf6e905435e7121584f9.tar.gz | |
Winograd changes to enable fp16 in armv8a multi_isa builds
* Changes in filelist.json: moved fp16 code from common to fp16
* Replaced the guard __ARM_FEATURE_FP16_VECTOR_ARITHMETIC
with ENABLE_FP16_KERNELS.
* Resolves COMPMID-6755
Change-Id: I4da1c53d3f9e4734e5e67125265ab4e3fc0dcbe4
Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/10865
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu')
| -rw-r--r-- | src/cpu/operators/CpuWinogradConv2d.cpp | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/src/cpu/operators/CpuWinogradConv2d.cpp b/src/cpu/operators/CpuWinogradConv2d.cpp index e4bcdc0b64..7d81aee0e9 100644 --- a/src/cpu/operators/CpuWinogradConv2d.cpp +++ b/src/cpu/operators/CpuWinogradConv2d.cpp @@ -122,13 +122,13 @@ bool get_winograd_kernel_implementation(const ITensorInfo success = arm_conv::winograd::get_implementation<float>(*winograd_impl, &CPUInfo::get(), *conv_args, nthreads, enable_fast_math, &winograd_cfg, nullptr); } -#if defined(__aarch64__) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) +#if defined(__aarch64__) && defined(ENABLE_FP16_KERNELS) else if (data_type == DataType::F16) { success = arm_conv::winograd::get_implementation<__fp16>(*winograd_impl, &CPUInfo::get(), *conv_args, nthreads, enable_fast_math, &winograd_cfg, nullptr); } -#endif // defined(__aarch64__) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) +#endif // defined(__aarch64__) && defined(ENABLE_FP16_KERNELS) else { success = false; |