cpu: conv: int8: depthwise: use bool for true/false flagging

larroy · Dec 30, 2018 · f79efef · f79efef
1 parent 0ae350c
commit f79efef
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 10 deletions.
diff --git a/src/cpu/jit_avx512_core_x8s8s32x_conv_kernel.cpp b/src/cpu/jit_avx512_core_x8s8s32x_conv_kernel.cpp
@@ -107,9 +107,8 @@ void jit_avx512_core_x8s8s32x_fwd_kernel::compute_eltwise(int ur_w) {
                 k * jcp.ur_w + ur_w);
 }
 
-void jit_avx512_core_x8s8s32x_fwd_kernel::store_output(int ur_w,
-        int last_oc_block_flag)
-{
+void jit_avx512_core_x8s8s32x_fwd_kernel::store_output(
+        int ur_w, bool last_oc_block_flag) {
     int nb_oc_block
             = jcp.is_depthwise ? jcp.nb_ch_blocking : jcp.nb_oc_blocking;
     int oc_block = jcp.is_depthwise ? jcp.ch_block : jcp.oc_block;
@@ -134,7 +133,7 @@ void jit_avx512_core_x8s8s32x_fwd_kernel::store_output(int ur_w,
     }
 
     for (int k = 0; k < nb_oc_block; k++) {
-        const bool mask_flag = last_oc_block_flag == 1 && k == nb_oc_block - 1;
+        const bool mask_flag = last_oc_block_flag && k == nb_oc_block - 1;
         int scale_offset = jcp.is_oc_scale * (sizeof(float) * k * oc_block);
         if (jcp.with_bias) {
             int bias_offset = jcp.typesize_bia * k * oc_block;
@@ -169,7 +168,7 @@ void jit_avx512_core_x8s8s32x_fwd_kernel::store_output(int ur_w,
     if (maybe_eltwise(0)) compute_eltwise(ur_w);
     if (p_sum_scale) { // post_op: sum
         for (int k = 0; k < nb_oc_block; k++) {
-            const bool mask_flag = last_oc_block_flag == 1 && k == nb_oc_block - 1;
+            const bool mask_flag = last_oc_block_flag && k == nb_oc_block - 1;
             for (int j = 0; j < ur_w; j++) {
                 int aux_output_offset
                         = jcp.typesize_out
@@ -188,7 +187,7 @@ void jit_avx512_core_x8s8s32x_fwd_kernel::store_output(int ur_w,
     if (maybe_eltwise(1)) compute_eltwise(ur_w);
 
     for (int k = 0; k < nb_oc_block; k++) {
-        const bool mask_flag = last_oc_block_flag == 1 && k == nb_oc_block - 1;
+        const bool mask_flag = last_oc_block_flag && k == nb_oc_block - 1;
         for (int j = 0; j < ur_w; j++) {
             Zmm zmm = zmm_out(j, k);
             if (jcp.dst_dt == data_type::u8) {
@@ -477,15 +476,15 @@ void jit_avx512_core_x8s8s32x_fwd_kernel::icb_loop(
 
         jne(common_store, T_NEAR);
 
-        store_output(ur_w, 1);
+        store_output(ur_w, true); // last oc block
         jmp(end_store, T_NEAR);
 
         L(common_store);
-        store_output(ur_w, 0);
+        store_output(ur_w, false);
 
         L(end_store);
     } else {
-        store_output(ur_w, 0);
+        store_output(ur_w, false);
     }
 }
 

diff --git a/src/cpu/jit_avx512_core_x8s8s32x_conv_kernel.hpp b/src/cpu/jit_avx512_core_x8s8s32x_conv_kernel.hpp
@@ -164,7 +164,7 @@ struct jit_avx512_core_x8s8s32x_fwd_kernel : public jit_generator {
 
     bool maybe_eltwise(int position);
     void prepare_output(int ur_w);
-    void store_output(int ur_w, int last_oc_block_flag);
+    void store_output(int ur_w, bool last_oc_block_flag);
     void compute_ker_dw(
             int ur_w, int pad_l, int pad_r, ic_block_t last_ic_block_flag);
     void compute_ker(int ur_w, int pad_l, int pad_r,