COMPMID-2577: Fuse bias addition and activation in gemm assembly kernels

Change-Id: I7f52112d2d05b1ea3d3f3d4b19b8eafab05d6c44
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-on: https://review.mlplatform.org/c/2141
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Pablo Marquez <pablo.tello@arm.com>

1 files changed, 9 insertions, 10 deletions

diff --git a/src/core/NEON/kernels/arm_gemm/gemm_hybrid_quantized.hpp b/src/core/NEON/kernels/arm_gemm/gemm_hybrid_quantized.hpp
index b4edece8d5..574ecef5b2 100644
--- a/src/core/NEON/kernels/arm_gemm/gemm_hybrid_quantized.hpp
+++ b/src/core/NEON/kernels/arm_gemm/gemm_hybrid_quantized.hpp
@@ -58,8 +58,6 @@ class GemmHybridQuantized : public GemmCommon<To, Tr> {
 
     const bool _trB;
 
-    const Tr _beta;
-
     /* Blocking info */
     const unsigned int _k_block;
     const unsigned int _n_block;
@@ -82,7 +80,7 @@ class GemmHybridQuantized : public GemmCommon<To, Tr> {
         return _Nsize * _nmulti * sizeof(int32_t);
     }
 
-    static unsigned int compute_k_block(const GemmArgs<Tr> &args) {
+    static unsigned int compute_k_block(const GemmArgs &args) {
         // We don't support K blocks as we only temporarily store 32 bit results.
         return args._Ksize;
 
@@ -112,7 +110,7 @@ class GemmHybridQuantized : public GemmCommon<To, Tr> {
         return k_block;
     }
 
-    static unsigned int compute_n_block(const GemmArgs<Tr> &args) {
+    static unsigned int compute_n_block(const GemmArgs &args) {
         if (args._cfg && args._cfg->outer_block_size) {
             return args._cfg->outer_block_size;
         }
@@ -142,9 +140,9 @@ public:
     GemmHybridQuantized & operator= (GemmHybridQuantized &) = delete;
 
     /* Constructor */
-    GemmHybridQuantized(const GemmArgs<Tr> &args, const ARequantizeLayer32 &qp)
+    GemmHybridQuantized(const GemmArgs &args, const ARequantizeLayer32 &qp)
               : _ci(args._ci), _Msize(args._Msize), _Nsize(args._Nsize), _Ksize(args._Ksize),
-                _nbatches(args._nbatches), _nmulti(args._nmulti), _trB(args._trB), _beta(args._beta),
+                _nbatches(args._nbatches), _nmulti(args._nmulti), _trB(args._trB),
                 _k_block(compute_k_block(args)), _n_block(compute_n_block(args)),
                 _Mround(roundup(args._Msize, strategy::out_height())),
                 _window_range(iceildiv(args._Msize, strategy::out_height()), _nbatches, iceildiv(_Nsize, _n_block), _nmulti),
@@ -210,8 +208,8 @@ public:
                     strat.kernel(this->_Aptr + (multi * this->_A_multi_stride) + (batch * this->_A_batch_stride) + (m_start * this->_lda) + k0, this->_lda,
                                  b_panel,
                                  result_buffer, (nmax-n0),
-                                 (k0 == 0) ? _beta : static_cast<Tr>(1),
-                                 (m_end - m_start), (nmax - n0), kern_k);
+                                 (m_end - m_start), (nmax - n0), kern_k,
+                                 nullptr, Activation(), false);
                 }
 
                 {
@@ -262,7 +260,7 @@ public:
         col_bias = reinterpret_cast<int32_t *>(in_buffer);
 
         for (unsigned int i=0; i<_nmulti; i++) {
-            compute_col_sums(_qp, _Nsize, _Ksize, B + (i * B_multi_stride), ldb, col_bias + (i * _Nsize),  _Ksize, 0);
+            compute_col_sums(_qp, _Nsize, _Ksize, B + (i * B_multi_stride), ldb, col_bias + (i * _Nsize),  _Ksize, i, 0);
         }
 
         uintptr_t buffer_int = reinterpret_cast<uintptr_t>(in_buffer);
@@ -295,8 +293,9 @@ public:
         col_bias = reinterpret_cast<int32_t *>(in_buffer);
     }
 
-    void set_quantized_bias(const int32_t *bias) override {
+    void set_quantized_bias(const int32_t *bias, size_t bias_multi_stride) override {
         _qp.bias = bias;
+        _qp.bias_multi_stride = bias_multi_stride;
     }
 };