[CI, CK examples] Disable time_kernel for CI tests and examples (#3464)

johannes-graner · web-flow · commit 0a474aa62f6d · 2026-01-07T16:30:57.000+01:00
* Disable kernel timing in tests

* default time_kernel = false in old CK examples
diff --git a/example/02_gemm_bilinear/gemm_bilinear_wmma_fp16.cpp b/example/02_gemm_bilinear/gemm_bilinear_wmma_fp16.cpp
@@ -119,7 +119,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 3840;
diff --git a/example/02_gemm_bilinear/gemm_bilinear_wmma_int8.cpp b/example/02_gemm_bilinear/gemm_bilinear_wmma_int8.cpp
@@ -119,7 +119,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 3840;
diff --git a/example/12_reduce/reduce_blockwise.cpp b/example/12_reduce/reduce_blockwise.cpp
@@ -31,7 +31,7 @@ class SimpleAppArgs
     bool do_verification = true;
     int data_type        = 1;
     int init_method      = 2;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     public:
     void show_usage(const char* cmd)
diff --git a/example/12_reduce/reduce_multiblock_atomic_add.cpp b/example/12_reduce/reduce_multiblock_atomic_add.cpp
@@ -31,7 +31,7 @@ class SimpleAppArgs
     bool do_verification = true;
     int data_type        = 1;
     int init_method      = 2;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     public:
     void show_usage(const char* cmd)
diff --git a/example/12_reduce/reduce_threadwise_multi_d.cpp b/example/12_reduce/reduce_threadwise_multi_d.cpp
@@ -31,7 +31,7 @@ class SimpleAppArgs
     bool do_verification = true;
     int data_type        = 1;
     int init_method      = 2;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     public:
     void show_usage(const char* cmd)
diff --git a/example/13_pool2d_fwd/pool2d_fwd_fp16.cpp b/example/13_pool2d_fwd/pool2d_fwd_fp16.cpp
@@ -53,7 +53,7 @@ int main(int argc, char* argv[])
     {
         do_verification = true;
         init_method     = 1;
-        time_kernel     = true;
+        time_kernel     = false;
     }
     else if(argc == 4)
     {
diff --git a/example/15_grouped_gemm/grouped_gemm_multiple_d_splitk_xdl_fp16.cpp b/example/15_grouped_gemm/grouped_gemm_multiple_d_splitk_xdl_fp16.cpp
@@ -90,7 +90,7 @@ struct ExecutionConfig final
     bool do_verification = true;
     int init_method      = 1;
     int k_batch          = 128;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 };
 
 bool run_grouped_gemm(const ProblemSize& problem_size, const ExecutionConfig& config)
diff --git a/example/15_grouped_gemm/grouped_gemm_multiple_d_xdl_fp16.cpp b/example/15_grouped_gemm/grouped_gemm_multiple_d_xdl_fp16.cpp
@@ -89,7 +89,7 @@ struct ExecutionConfig final
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 };
 
 bool run_grouped_gemm(const ProblemSize& problem_size, const ExecutionConfig& config)
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_add_add_mean_meansquare_xdl_fp16.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_add_add_mean_meansquare_xdl_fp16.cpp
@@ -268,7 +268,7 @@ int main()
         pass &= ck::utils::check_err(r1_m, r1_m_host, "Error: Incorrect results d1", 1e-2, 1e-2);
     }
 
-    bool time_kernel = true;
+    bool time_kernel = false;
     if(time_kernel)
     {
         float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel});
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_add_addsquare_xdl_int8.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_add_addsquare_xdl_int8.cpp
@@ -302,7 +302,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_bf16.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_bf16.cpp
@@ -106,7 +106,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_fp16.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_fp16.cpp
@@ -106,7 +106,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_fp32.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_fp32.cpp
@@ -106,7 +106,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_int4.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_int4.cpp
@@ -108,7 +108,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_int8.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_max_xdl_int8.cpp
@@ -105,7 +105,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_mean_meansquare_xdl_bf16.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_mean_meansquare_xdl_bf16.cpp
@@ -112,7 +112,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_mean_meansquare_xdl_fp16.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_mean_meansquare_xdl_fp16.cpp
@@ -112,7 +112,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/16_gemm_multi_d_multi_reduces/gemm_mean_meansquare_xdl_fp32.cpp b/example/16_gemm_multi_d_multi_reduces/gemm_mean_meansquare_xdl_fp32.cpp
@@ -112,7 +112,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M = 1024;
diff --git a/example/22_cgemm/cgemm_xdl_int4.cpp b/example/22_cgemm/cgemm_xdl_int4.cpp
@@ -81,7 +81,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // CGEMM shape
     ck::index_t M = 1024;
diff --git a/example/23_softmax/softmax_blockwise.cpp b/example/23_softmax/softmax_blockwise.cpp
@@ -65,7 +65,7 @@ class SimpleAppArgs
 
     bool do_verification = true;
     int init_method      = 2;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     public:
     void show_usage(const char* cmd)
diff --git a/example/24_batched_gemm/run_batched_gemm_example_fp16int4_b_scale.inc b/example/24_batched_gemm/run_batched_gemm_example_fp16int4_b_scale.inc
@@ -27,7 +27,7 @@ struct ExecutionConfig final
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 };
 
 template <typename DataType>
diff --git a/example/29_batched_gemm_bias_e_permute/batched_gemm_bias_e_permute_wmma_fp16.cpp b/example/29_batched_gemm_bias_e_permute/batched_gemm_bias_e_permute_wmma_fp16.cpp
@@ -248,7 +248,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     ck::index_t G0 = 1;
     ck::index_t G1 = 2;
diff --git a/example/30_grouped_conv_fwd_multiple_d/common.hpp b/example/30_grouped_conv_fwd_multiple_d/common.hpp
@@ -92,7 +92,7 @@ struct ExecutionConfig final
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 };
 
 #define DefaultConvParam                                                       \
diff --git a/example/30_grouped_conv_fwd_multiple_d/common_wmma.hpp b/example/30_grouped_conv_fwd_multiple_d/common_wmma.hpp
@@ -92,7 +92,7 @@ struct ExecutionConfig final
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 };
 
 #define DefaultConvParam                                                       \
diff --git a/example/33_multiple_reduce/dual_reduce_common.hpp b/example/33_multiple_reduce/dual_reduce_common.hpp
@@ -40,7 +40,7 @@ class SimpleAppArgs
 
     bool do_verification = true;
     int init_method      = 2;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     public:
     SimpleAppArgs()
diff --git a/example/35_splitK_gemm/common.hpp b/example/35_splitK_gemm/common.hpp
@@ -44,7 +44,7 @@ struct ExecutionConfig final
 {
     bool do_verification = true;
     int init_method      = 2;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 };
 
 template <ck::index_t... Is>
diff --git a/example/36_sparse_embedding/sparse_embedding3_forward_layernorm.cpp b/example/36_sparse_embedding/sparse_embedding3_forward_layernorm.cpp
@@ -56,7 +56,7 @@ template<> struct emb_kernel<ck::half_t, 8192> { using kernel_type = DeviceInsta
 
 int main(int argc, char* argv[])
 {
-    bool time_kernel = true;
+    bool time_kernel = false;
 
     ck::index_t num_rows          = 65536;
     constexpr auto dims           = ck::Sequence<256, 512, 768, 1024, 1536, 2048, 4096, 8192>{};
diff --git a/example/37_batched_gemm_add_add_relu_gemm_add/batched_gemm_add_add_relu_gemm_add_xdl_fp16.cpp b/example/37_batched_gemm_add_add_relu_gemm_add/batched_gemm_add_add_relu_gemm_add_xdl_fp16.cpp
@@ -195,7 +195,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     // GEMM shape
     ck::index_t M              = 1024;
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp
@@ -86,7 +86,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp
@@ -84,7 +84,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp
@@ -87,7 +87,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp
@@ -84,7 +84,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perchannel_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_dl_perchannel_quantization_int8.cpp
@@ -84,7 +84,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_xdl_bias_relu_perchannel_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_xdl_bias_relu_perchannel_quantization_int8.cpp
@@ -90,7 +90,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_xdl_bias_relu_perlayer_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_xdl_bias_relu_perlayer_quantization_int8.cpp
@@ -88,7 +88,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/40_conv2d_fwd_quantization/conv2d_fwd_xdl_perchannel_quantization_int8.cpp b/example/40_conv2d_fwd_quantization/conv2d_fwd_xdl_perchannel_quantization_int8.cpp
@@ -88,7 +88,7 @@ using DeviceGroupedConvNDFwdInstance =
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/42_groupnorm_fwd/run_groupnorm_fwd_example.inc b/example/42_groupnorm_fwd/run_groupnorm_fwd_example.inc
@@ -12,7 +12,7 @@ int run_groupnorm_fwd_example(int argc, char* argv[])
     ck::index_t C = 128;
 
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
     bool log_kernel      = true;
 
     if(argc == 1)
diff --git a/example/44_elementwise_permute/elementwise_binary_4D_fp16.cpp b/example/44_elementwise_permute/elementwise_binary_4D_fp16.cpp
@@ -53,7 +53,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     std::vector<std::size_t> nchw = {16, 128, 32, 64};
 
diff --git a/example/44_elementwise_permute/elementwise_permute_4D_fp16.cpp b/example/44_elementwise_permute/elementwise_permute_4D_fp16.cpp
@@ -46,7 +46,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/44_elementwise_permute/elementwise_permute_4D_fp16_col.cpp b/example/44_elementwise_permute/elementwise_permute_4D_fp16_col.cpp
@@ -50,7 +50,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/44_elementwise_permute/elementwise_permute_4D_fp16_row.cpp b/example/44_elementwise_permute/elementwise_permute_4D_fp16_row.cpp
@@ -50,7 +50,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/44_elementwise_permute/elementwise_permute_4D_fp32_col.cpp b/example/44_elementwise_permute/elementwise_permute_4D_fp32_col.cpp
@@ -49,7 +49,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/44_elementwise_permute/elementwise_permute_4D_fp32_row.cpp b/example/44_elementwise_permute/elementwise_permute_4D_fp32_row.cpp
@@ -50,7 +50,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     if(argc == 1)
     {
diff --git a/example/44_elementwise_permute/elementwise_scale_permute_amax_2D_fp16_fp8.cpp b/example/44_elementwise_permute/elementwise_scale_permute_amax_2D_fp16_fp8.cpp
@@ -121,7 +121,7 @@ void reference_scale_permute_amax(Tensor<InputDataType>& input,
 int main(int argc, char* argv[])
 {
     bool do_verification = true;
-    bool time_kernel     = true;
+    bool time_kernel     = false;
 
     const float scale = 2.f;
 
diff --git a/example/44_elementwise_permute/elementwise_trinary_4D_fp16.cpp b/example/44_elementwise_permute/elementwise_trinary_4D_fp16.cpp
diff --git a/example/45_elementwise_normalization/elementwise_layernorm_blockwise.cpp b/example/45_elementwise_normalization/elementwise_layernorm_blockwise.cpp
diff --git a/example/65_gemm_multiply_multiply/moe_gemm1_xdl_fp8.cpp b/example/65_gemm_multiply_multiply/moe_gemm1_xdl_fp8.cpp
diff --git a/example/65_gemm_multiply_multiply/moe_gemm1_xdl_fp8_blockscale.cpp b/example/65_gemm_multiply_multiply/moe_gemm1_xdl_fp8_blockscale.cpp
diff --git a/example/65_gemm_multiply_multiply/moe_gemm1_xdl_pk_i4.cpp b/example/65_gemm_multiply_multiply/moe_gemm1_xdl_pk_i4.cpp
diff --git a/example/65_gemm_multiply_multiply/moe_gemm2_xdl_fp8.cpp b/example/65_gemm_multiply_multiply/moe_gemm2_xdl_fp8.cpp
diff --git a/example/65_gemm_multiply_multiply/moe_gemm2_xdl_fp8_blockscale.cpp b/example/65_gemm_multiply_multiply/moe_gemm2_xdl_fp8_blockscale.cpp
diff --git a/example/65_gemm_multiply_multiply/moe_gemm2_xdl_pk_i4.cpp b/example/65_gemm_multiply_multiply/moe_gemm2_xdl_pk_i4.cpp
diff --git a/example/67_gemm_microscaling/moe_gemm1_xdl_mx_fp4.cpp b/example/67_gemm_microscaling/moe_gemm1_xdl_mx_fp4.cpp
diff --git a/example/67_gemm_microscaling/moe_gemm1_xdl_mx_fp4_bns.cpp b/example/67_gemm_microscaling/moe_gemm1_xdl_mx_fp4_bns.cpp
diff --git a/example/67_gemm_microscaling/moe_gemm1_xdl_mx_fp4_bpreshuffle.cpp b/example/67_gemm_microscaling/moe_gemm1_xdl_mx_fp4_bpreshuffle.cpp
diff --git a/example/67_gemm_microscaling/moe_gemm2_xdl_mx_fp4.cpp b/example/67_gemm_microscaling/moe_gemm2_xdl_mx_fp4.cpp
diff --git a/example/67_gemm_microscaling/moe_gemm2_xdl_mx_fp4_bns.cpp b/example/67_gemm_microscaling/moe_gemm2_xdl_mx_fp4_bns.cpp
diff --git a/example/67_gemm_microscaling/moe_gemm2_xdl_mx_fp4_bpreshuffle.cpp b/example/67_gemm_microscaling/moe_gemm2_xdl_mx_fp4_bpreshuffle.cpp
diff --git a/test/batched_gemm_multi_d/test_batched_gemm_multi_d_dl.cpp b/test/batched_gemm_multi_d/test_batched_gemm_multi_d_dl.cpp
diff --git a/test/gemm/gemm_standalone_xdl_fp16.cpp b/test/gemm/gemm_standalone_xdl_fp16.cpp
diff --git a/test/wrapper/test_wrapper_gemm_xdl.cpp b/test/wrapper/test_wrapper_gemm_xdl.cpp

Original file line number	Diff line number	Diff line change
`@@ -119,7 +119,7 @@ int main(int argc, char* argv[])`
`119`	`119`	`{`
`120`	`120`	`bool do_verification = true;`
`121`	`121`	`int init_method = 1;`
`122`		`- bool time_kernel = true;`
	`122`	`+ bool time_kernel = false;`
`123`	`123`
`124`	`124`	`// GEMM shape`
`125`	`125`	`ck::index_t M = 3840;`
Original file line number	Diff line number	Diff line change
`@@ -53,7 +53,7 @@ int main(int argc, char* argv[])`
`53`	`53`	`{`
`54`	`54`	`do_verification = true;`
`55`	`55`	`init_method = 1;`
`56`		`- time_kernel = true;`
	`56`	`+ time_kernel = false;`
`57`	`57`	`}`
`58`	`58`	`else if(argc == 4)`
`59`	`59`	`{`
Original file line number	Diff line number	Diff line change
`@@ -89,7 +89,7 @@ struct ExecutionConfig final`
`89`	`89`	`{`
`90`	`90`	`bool do_verification = true;`
`91`	`91`	`int init_method = 1;`
`92`		`- bool time_kernel = true;`
	`92`	`+ bool time_kernel = false;`
`93`	`93`	`};`
`94`	`94`
`95`	`95`	`bool run_grouped_gemm(const ProblemSize& problem_size, const ExecutionConfig& config)`
Original file line number	Diff line number	Diff line change
`@@ -268,7 +268,7 @@ int main()`
`268`	`268`	`pass &= ck::utils::check_err(r1_m, r1_m_host, "Error: Incorrect results d1", 1e-2, 1e-2);`
`269`	`269`	`}`
`270`	`270`
`271`		`- bool time_kernel = true;`
	`271`	`+ bool time_kernel = false;`
`272`	`272`	`if(time_kernel)`
`273`	`273`	`{`
`274`	`274`	`float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel});`
Original file line number	Diff line number	Diff line change
`@@ -302,7 +302,7 @@ int main(int argc, char* argv[])`
`302`	`302`	`{`
`303`	`303`	`bool do_verification = true;`
`304`	`304`	`int init_method = 1;`
`305`		`- bool time_kernel = true;`
	`305`	`+ bool time_kernel = false;`
`306`	`306`
`307`	`307`	`// GEMM shape`
`308`	`308`	`ck::index_t M = 1024;`