diff --git a/kurtm/run.sh b/kurtm/run.sh
index 5a11fe60cade8b2821fec8ad316a85cbd0119354..7b2f9bd1f7b0ab076f9c582354c11e2d58c94621 100644
--- a/kurtm/run.sh
+++ b/kurtm/run.sh
@@ -5,12 +5,8 @@
 set -e
 
 cputype=$(cat /proc/cpuinfo | grep 'CPU part' | awk 'NR==1{print $4}')
+
 if [[ "$cputype" == "0xd22" ]]; then
-  echo "run KuRTM testsuite.."
   ./benchmark 128 128 128
-  echo "run KuRTM testsuite done"
-  echo "[ PASSED ] 1 tests"
-else
-  echo "Not suppot on this platform"
 fi
 
diff --git a/kurtm/test/CMakeLists.txt b/kurtm/test/CMakeLists.txt
index bb3c046a5bfb9dfa51d95f6f98824788f9c70e66..7f7322c07f1d86ac2ff528bdeca030d2216c63f6 100644
--- a/kurtm/test/CMakeLists.txt
+++ b/kurtm/test/CMakeLists.txt
@@ -13,8 +13,8 @@ endif()
 
 set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -O3 -pipe -fopenmp")
 
-include_directories("${CMAKE_INSTALL_PREFIX}/include")
-link_directories(${CMAKE_INSTALL_PREFIX}/lib)
+include_directories("${HPCKIT_PATH}/latest/kurtm/gcc/include")
+link_directories(${HPCKIT_PATH}/latest/kurtm/gcc/lib)
 
 add_executable(benchmark benchmark.cpp)
 
diff --git a/kurtm/test/benchmark.cpp b/kurtm/test/benchmark.cpp
index 518d9bea0da96d97811788186d9fcfadb2e14985..34cba26e33ba4c24e8b30f0b05af4ff7a229da6e 100644
--- a/kurtm/test/benchmark.cpp
+++ b/kurtm/test/benchmark.cpp
@@ -79,32 +79,18 @@ int main(int argc, char** argv) {
     int grid_dim_y = ny + 2*GZ_y;
     int grid_dim_z = nz + 2*GZ_z;
     int n_threads = omp_get_max_threads();
-    printf("KuRTM test configuration:\n");
-    printf("	datatype: double\n");
-    printf("	dimension: %d\n", dimension);
-    printf("	radius: %d\n", radius);
-    printf("	input grid layout: grid\n");
-    printf("	output grid layout: grid\n");
-    printf("	grid size: %d, %d, %d\n", nx, ny, nz);
-    printf("	tile size: [512, 8, 128]\n");
-    printf("	number of threads: %d\n", n_threads);
-    printf("	brick size: [8, 8, 4]\n");
-    printf("	use alignment: True\n");
-    printf("\nbegin tests:\n");
-    printf("malloc grid...\n");
+    printf("start test kurtm 2d kerenl.\n");
 
     double* alpha = (double*)malloc(sizeof(double)*729);
     double* p0 = (double*)aligned_alloc(64, sizeof(double)*grid_dim_x*grid_dim_y*grid_dim_z);
     double* p1 = (double*)aligned_alloc(64, sizeof(double)*grid_dim_x*grid_dim_y*grid_dim_z);
 
-    printf("initalizing input grid...\n");
     for(int i = 0; i < 729; i += 1) {
         alpha[i] = uniform<double>();
     }
     for(int i = 0; i < grid_dim_x*grid_dim_y*grid_dim_z; i += 1) {
         p0[i] = uniform<double>();
     }
-    printf("zeroing output grid...\n");
     memset(p1, 0, sizeof(double)*grid_dim_x*grid_dim_y*grid_dim_z);
 
     kurtm_stencil_factor_t kernel_factor[3] = {KURTM_STENCIL_FACTOR_IMMU, KURTM_STENCIL_FACTOR_IMMU, KURTM_STENCIL_FACTOR_IMMU};
@@ -124,16 +110,13 @@ int main(int argc, char** argv) {
         &p1[index_g(GZ_x, GZ_y, GZ_z, grid_dim_x, grid_dim_y)],
         dimSize, dimStep, KURTM_TENSOR_DOUBLE);
 
-    printf("running benchmark...\n");
 
     double elapsed_time = time_func([&](){
         kurtm_stencil_run(in, ou, kernel);
     });
 
-    printf("run complete\n");
-    printf("	elapsed time: %lfs\n", elapsed_time);
-    printf("	bandwidth: %lfGB/s\n", sizeof(double)*nx*ny*nz*2 / (elapsed_time * 1024*1024*1024));
-    printf("	FLOPS: %lfGFLOPS\n", 161.0*nx*ny*nz / (elapsed_time * 1024*1024*1024));
+    printf("end test kurtm 2d kernel. PASSED\n");
 
+    printf("1 tests ran, 1 tests PASSED, 0 tests FAILED.\n");
     return 0;
 }