From 19bd412fd044197726dbd8c756dbd74a9e33fd2b Mon Sep 17 00:00:00 2001 From: Michalis Spyrou Date: Wed, 22 Jan 2020 10:27:06 +0000 Subject: COMPMID-3022 Fix NEON QASYMM8 Reduction Operation Use proper offsets in our tests and subtract the accumulated offsets when we do a SUM operation. Also change the reference MEAN_SUM impementation to use floats. As a result the tolerance is increased to 2. Change-Id: Icac8b84680bd880fd30dc1dd82d19cce43921eb5 Signed-off-by: Michalis Spyrou Reviewed-on: https://review.mlplatform.org/c/2616 Comments-Addressed: Arm Jenkins Tested-by: Arm Jenkins Reviewed-by: Michele Di Giorgio --- .../core/NEON/wrapper/intrinsics/intrinsics.h | 1 + arm_compute/core/NEON/wrapper/intrinsics/qmovun.h | 46 ++++++++++++++++++++++ 2 files changed, 47 insertions(+) create mode 100644 arm_compute/core/NEON/wrapper/intrinsics/qmovun.h (limited to 'arm_compute') diff --git a/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h b/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h index 3d674757e8..26b4322fa4 100644 --- a/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h +++ b/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h @@ -57,6 +57,7 @@ #include "arm_compute/core/NEON/wrapper/intrinsics/pmax.h" #include "arm_compute/core/NEON/wrapper/intrinsics/pmin.h" #include "arm_compute/core/NEON/wrapper/intrinsics/pow.h" +#include "arm_compute/core/NEON/wrapper/intrinsics/qmovun.h" #include "arm_compute/core/NEON/wrapper/intrinsics/reinterpret.h" #include "arm_compute/core/NEON/wrapper/intrinsics/rev64.h" #include "arm_compute/core/NEON/wrapper/intrinsics/round.h" diff --git a/arm_compute/core/NEON/wrapper/intrinsics/qmovun.h b/arm_compute/core/NEON/wrapper/intrinsics/qmovun.h new file mode 100644 index 0000000000..a0347020db --- /dev/null +++ b/arm_compute/core/NEON/wrapper/intrinsics/qmovun.h @@ -0,0 +1,46 @@ +/* + * Copyright (c) 2020 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef ARM_COMPUTE_WRAPPER_QMOVUN_H +#define ARM_COMPUTE_WRAPPER_QMOVUN_H + +#include + +namespace arm_compute +{ +namespace wrapper +{ +#define VQMOVUN_IMPL(dtype, vtype, prefix, postfix) \ + inline dtype vqmovun(const vtype &a) \ + { \ + return prefix##_##postfix(a); \ + } + +VQMOVUN_IMPL(uint32x2_t, int64x2_t, vqmovun, s64) +VQMOVUN_IMPL(uint16x4_t, int32x4_t, vqmovun, s32) +VQMOVUN_IMPL(uint8x8_t, int16x8_t, vqmovun, s16) + +#undef VQMOVUN_IMPL +} // namespace wrapper +} // namespace arm_compute +#endif /* ARM_COMPUTE_WRAPPER_QMOVUN_H */ -- cgit v1.2.1