Update GEMM assembly kernels

- Introduce Fp32 kernels with internal calculations in Bfloat16 when
fast_mode is enabled
- Improve kernel selection heuristics

Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I68a9e7e862b6fd2721b46e0d7cc791091c4ab279
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5965
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 13 insertions, 1 deletions

diff --git a/src/core/NEON/kernels/arm_gemm/gemv_batched.hpp b/src/core/NEON/kernels/arm_gemm/gemv_batched.hpp
index 12216009d2..4fc9b3456a 100644
--- a/src/core/NEON/kernels/arm_gemm/gemv_batched.hpp
+++ b/src/core/NEON/kernels/arm_gemm/gemv_batched.hpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2017-2020 Arm Limited.
+ * Copyright (c) 2017-2021 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -95,6 +95,18 @@ public:
     void set_pretransposed_B_data(void *buffer) override {
         _subgemm->set_pretransposed_B_data(buffer);
     }
+
+    GemmConfig get_config() override {
+        GemmConfig c = _subgemm->get_config();
+
+        std::string new_filter = "gemv_batched[";
+        new_filter.append(c.filter);
+        new_filter.append("]");
+
+        c.filter = new_filter;
+
+        return c;
+    }
 };
 
 } // namespace arm_gemm