cpu: x64: rnn: multiplier is not required for gru

The multiplier (2) is only required when rnn.brgemm_fwd_iter_layer_fuse_possible
is true and is only used in kernel_fused_iter_layer.

Author: densamoilov

src/cpu/x64/rnn/brgemm_cell_common_fwd.cpp

@@ -1,5 +1,5 @@
 /*******************************************************************************
-* Copyright 2021-2024 Intel Corporation
+* Copyright 2021-2025 Intel Corporation
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
@@ -688,9 +688,8 @@ void brgemm_gru_t<src_t, weights_t, scratch_t, gemm_acc_t>::kernel(
     gemm_acc_t *const amx_buffer = is_amx
             ? amx_scratchpad_ + rnn_.m_block * rnn_.n_block * ithr
             : nullptr;
-    const int max_K_Block = 2
-            * nstl::max(rnn_.KB1_blocks + 1,
-                    nstl::max(rnn_.KBproj_blocks + 1, rnn_.KB2_blocks + 1));
+    const int max_K_Block = nstl::max(rnn_.KB1_blocks + 1,
+            nstl::max(rnn_.KBproj_blocks + 1, rnn_.KB2_blocks + 1));
     brgemm_batch_element_t *const addr_batch
             = addr_batch_global_ + ithr * max_K_Block;