[arm][kernel][test] test: fix equality chk in elementwise_op test test=develop

lite/kernels/arm/elementwise_compute_test.cc

@@ -268,6 +268,14 @@ template void elementwise_imod_compute_ref<int32_t>(
 template void elementwise_imod_compute_ref<int64_t>(
     const operators::ElementwiseParam& param, const std::string act_type);
 
+template <class T>
+bool is_fp_close(T v1, T v2, T rel_tol = 1e-4, T abs_tol = 1e-5) {
+  bool abs_chk = std::abs(v1 - v2) < abs_tol;
+  bool rel_chk =
+      (std::abs(v1 - v2) / std::min(std::abs(v1), std::abs(v2))) < rel_tol;
+  return abs_chk || rel_chk;
+}
+
 template <template <class, PrecisionType> class ElementWiseComputeTemplate,
           typename T,
           PrecisionType PType>
@@ -345,13 +353,14 @@ void do_elementwise_compute(const char* op_type_str) {
               elementwise_compute_ref<T>(param, op_type_str, "");
               if (std::is_floating_point<T>::value) {
                 for (int i = 0; i < output.dims().production(); i++) {
-                  ASSERT_NEAR(output_data[i], output_ref_data[i], 1e-5)
-                      << "Value differ at index " << i;
+                  ASSERT_EQ(is_fp_close(output_data[i], output_ref_data[i]),
+                            true)
+                      << op_type_str << "Value differ at index " << i;
                 }
               } else {
                 for (int i = 0; i < output.dims().production(); i++) {
                   ASSERT_EQ(output_data[i], output_ref_data[i])
-                      << "Value differ at index " << i;
+                      << op_type_str << "Value differ at index " << i;
                 }
               }
             }
@@ -362,7 +371,7 @@ void do_elementwise_compute(const char* op_type_str) {
   }
 }
 
-TEST(elementwise_add, compute_fp32) {
+TEST(elementwise_op, compute_fp32) {
   do_elementwise_compute<ElementwiseAddCompute, float, PRECISION(kFloat)>(
       "add");
   do_elementwise_compute<ElementwiseSubCompute, float, PRECISION(kFloat)>(
@@ -376,7 +385,7 @@ TEST(elementwise_add, compute_fp32) {
   }
 }
 
-TEST(elementwise_add, compute_i32) {
+TEST(elementwise_op, compute_i32) {
   do_elementwise_compute<ElementwiseAddCompute, int32_t, PRECISION(kInt32)>(
       "add");
   do_elementwise_compute<ElementwiseSubCompute, int32_t, PRECISION(kInt32)>(
@@ -390,7 +399,7 @@ TEST(elementwise_add, compute_i32) {
   }
 }
 
-TEST(elementwise_add, compute_i64) {
+TEST(elementwise_op, compute_i64) {
   do_elementwise_compute<ElementwiseAddCompute, int64_t, PRECISION(kInt64)>(
       "add");
   do_elementwise_compute<ElementwiseSubCompute, int64_t, PRECISION(kInt64)>(