From c0b6f76561580414f08633a804fc548ccad65659 Mon Sep 17 00:00:00 2001
From: Georgios Pinitas <georgios.pinitas@arm.com>
Date: Mon, 2 Nov 2020 01:37:17 +0000
Subject: COMPMID-3776: Indirect GEMM

Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I51a1b0f098bc3a8c408c50c92221e4df3061e12c
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4343
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
---
 .../kernels/sve_hybrid_u8u32_dot_6x4VL.hpp         | 85 ++++++++++++++++++++++
 1 file changed, 85 insertions(+)
 create mode 100644 src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_u8u32_dot_6x4VL.hpp

(limited to 'src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_u8u32_dot_6x4VL.hpp')

diff --git a/src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_u8u32_dot_6x4VL.hpp b/src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_u8u32_dot_6x4VL.hpp
new file mode 100644
index 0000000000..af9de4a6eb
--- /dev/null
+++ b/src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_u8u32_dot_6x4VL.hpp
@@ -0,0 +1,85 @@
+/*
+ * Copyright (c) 2019-2020 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.
+ */
+#pragma once
+#ifdef __ARM_FEATURE_SVE
+
+#include "../std_transforms_sve.hpp"
+
+#define ARGLIST  \
+   unsigned int, const unsigned int *, \
+   IndirectInputArg<uint8_t>, \
+   size_t, size_t, \
+   const uint8_t *, \
+   IndirectOutputArg<uint32_t>, \
+   const uint32_t *, Activation, bool
+
+namespace arm_gemm
+{
+
+// Actual kernel implementations
+void sve_hybrid_u8u32_dot_6x4VL( ARGLIST );
+
+class cls_sve_hybrid_u8u32_dot_6x4VL
+{
+public:
+    typedef uint8_t operand_type;
+    typedef uint32_t result_type;
+
+    typedef void (*kern_type)( ARGLIST );
+
+    /* Kernel blocking parameters */
+    static constexpr unsigned int out_height()
+    {
+        return 6;
+    }
+
+    static unsigned int out_width()
+    {
+        return get_vector_length<uint32_t>() * 4;
+    }
+
+    static constexpr unsigned int k_unroll()
+    {
+        return 4;
+    }
+
+    static constexpr bool supports_accumulate()
+    {
+        return true;
+    }
+
+    StdTransformsSVE<operand_type, result_type, 6, 4, 4> transforms = {};
+
+    // Default to the generic kernel
+    kern_type kernel=sve_hybrid_u8u32_dot_6x4VL;
+
+    cls_sve_hybrid_u8u32_dot_6x4VL(const CPUInfo *)
+    {
+    }
+};
+
+} // namespace arm_gemm
+
+#undef ARGLIST
+#endif // __ARM_FEATURE_SVE
-- 
cgit v1.2.1