removing more unused commented code

llvm/projects/hpvm-tensor-rt/tensor_runtime/src/approx_simulation.cu

@@ -471,33 +471,15 @@ void sampleFilter2(Tensor* newFilter, Tensor* filter,
   int N = n * c * h * w;
 
   float mul_factor;
-  
-  /*if (interpolation_id == 0){
-    mul_factor = (skip_rate * 1.0) / (skip_rate - 1); 
-  }
-  else if (interpolation_id == 1){
-    mul_factor = (skip_rate * 1.0) / (skip_rate - 1);
-    mul_factor = (mul_factor + 1.0) / 2;
-  }
-  */
-
-  //printf ("interpolation_rate = %f \n", interpolation_rate);
-
   mul_factor = (skip_rate * 1.0) / (skip_rate - 1);
   mul_factor = 1 + (interpolation_rate * (mul_factor - 1.0));
- 
-
-  printf ("mul_factor = %f \n", mul_factor);
-
   DEBUG ("mul_factor = %f \n", mul_factor);
-
   
   sampleFilterElems<<<numBlocks,128>>>(N,
 				       n, c, h, w,
 				       (float *) filter->gpu_data,
 				       skip_rate, skip_offset, mul_factor,
 				       (float *) newFilter->gpu_data);
-
 }
 
 
@@ -943,6 +925,8 @@ int getSwing(int swing){
 
 void initializeAutotuner(){
 
+  printf("initializing tuner .... \n");
+  
   sampParamSet = new SampParamSet;
   perfParamSet = new PerfParamSet;  
 }

llvm/projects/hpvm-tensor-rt/tensor_runtime/src/fp16_gemm.cu

@@ -240,42 +240,5 @@ void hgemm(const float * const __restrict__ af,
 }
 
 
-/*int main() {
-    const unsigned num_rows_a = 5000 * 14 * 14;
-    const unsigned num_cols_a = 800;
-    const unsigned num_rows_b = num_cols_a;
-    const unsigned num_cols_b = 64;
-
-    const unsigned num_elements_a = num_rows_a * num_cols_a;
-    const unsigned num_elements_b = num_rows_b * num_cols_b;
-    const unsigned num_elements_c = num_rows_a * num_cols_b;
-
-    float *a;
-    float *b;
-    float *c;
-
-    checkCuda(cudaMallocManaged(&a, sizeof(float) * num_elements_a));
-    checkCuda(cudaMallocManaged(&b, sizeof(float) * num_elements_b));
-    checkCuda(cudaMallocManaged(&c, sizeof(float) * num_elements_c));
-
-    init(a, num_elements_a);
-    init(b, num_elements_b);
-    init(c, num_elements_c);
-
-    // FP32
-    sgemm(a, num_rows_a, num_cols_a, b, num_rows_b, num_cols_b, c);
-    printArray(c, 16);
-
-    // FP16
-    hgemm(a, num_rows_a, num_cols_a, b, num_rows_b, num_cols_b, c);
-    printArray(c, 16);
-
-    checkCuda(cudaFree(a));
-    checkCuda(cudaFree(b));
-    checkCuda(cudaFree(c));
-
-    return 0;
-}
-*/
-
+ 
 #endif