EXPECT_EQ(attr_found->second.type(), tensorflow::DataType::DT_FLOAT);
  attr_found = attr_values.find("Tidx");
  EXPECT_NE(attr_found, attr_values.cend());
  EXPECT_EQ(attr_found->second.type(), tensorflow::DataType::DT_INT32);

  TFE_TensorHandle* retvals[1] = {nullptr};
  int num_retvals = 1;
  TFE_Execute(minOp, &retvals[0], &num_retvals, status);
  EXPECT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);

                          std::vector<int64_t> shape,
                          std::vector<int64_t> strides) {
  size_t num_elements = 1;
  for (int i = 0; i < static_cast<int32_t>(shape.size()); ++i) {
    num_elements *= shape[i];
  }
  std::vector<float> data(num_elements);
  for (size_t j = 0; j < num_elements; ++j) {
    data[j] = j;
  }
  DLManagedTensor dlm_in = {};

index e54a067..215b0e0 100644
--- a/sysdeps/ieee754/flt-32/k_rem_pio2f.c
+++ b/sysdeps/ieee754/flt-32/k_rem_pio2f.c
@@ -65,7 +65,8 @@ int __kernel_rem_pio2f(float *x, float *y, int e0, int nx, int prec, const int32
 
     /* compute q[0],q[1],...q[jk] */
 	for (i=0;i<=jk;i++) {
-	    for(j=0,fw=0.0;j<=jx;j++) fw += x[j]*f[jx+i-j]; q[i] = fw;
+	    for(j=0,fw=0.0;j<=jx;j++) fw += x[j]*f[jx+i-j];
+	    q[i] = fw;
 	}
 

                       const int32_t* values) {
  int64_t num_values = 1;
  for (int i = 0; i < num_dims; ++i) {
    num_values *= dims[i];
  }
  TF_Tensor* t =
      TF_AllocateTensor(TF_INT32, dims, num_dims, sizeof(int32_t) * num_values);
  memcpy(TF_TensorData(t), values, sizeof(int32_t) * num_values);
  return t;
}

TF_Tensor* Int32Tensor(const std::vector<int32_t>& values) {
  int64_t dims = values.size();

==============================================================================*/

#include <limits.h>
#include <memory.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#include <unistd.h>

#ifdef _WIN32
#include <process.h>
#endif

#include "tensorflow/c/c_api.h"
#include "tensorflow/c/c_api_experimental.h"
#include "tensorflow/c/env.h"
#include "tensorflow/c/kernels.h"

  // will be blocked till the copy completes. This is the default placement
  // policy.
  TFE_DEVICE_PLACEMENT_SILENT = 2,
  // Placement policy which silently copies int32 tensors but not other dtypes.
  TFE_DEVICE_PLACEMENT_SILENT_FOR_INT32 = 3,
} TFE_ContextDevicePlacementPolicy;
// LINT.ThenChange(//tensorflow/c/eager/immediate_execution_context.h)

// Sets the default execution mode (sync/async). Note that this can be

  /// that there is no limit on the number of entries in the cache (however, if
  /// `max_age` is also 0, the cache will not be populated).
  ExpiringLRUCache(uint64_t max_age, size_t max_entries,
                   std::function<uint64_t()> timer_seconds = TF_NowSeconds)
      : max_age_(max_age),
        max_entries_(max_entries),
        timer_seconds_(timer_seconds) {}

typedef struct TP_OptimizerRegistrationParams {
  size_t struct_size;
  void* ext;  // reserved for future use

  // Graph C API version.
  int32_t major_version;
  int32_t minor_version;
  int32_t patch_version;

  // Backend device type supported by the optimizer.
  const char* device_type;
  TP_OptimizerConfigs* optimizer_configs;  // output, set by plugin

}

TF_CAPI_EXPORT extern uint64_t TF_NowNanos(void) {
  return ::tensorflow::Env::Default()->NowNanos();
}

// Returns the number of microseconds since the Unix epoch.
TF_CAPI_EXPORT extern uint64_t TF_NowMicros(void) {
  return ::tensorflow::Env::Default()->NowMicros();
}

// Returns the number of seconds since the Unix epoch.
TF_CAPI_EXPORT extern uint64_t TF_NowSeconds(void) {

  // will be blocked till the copy completes. This is the default policy.
  DEVICE_PLACEMENT_SILENT = 2,
  // Placement policy which silently copies int32 tensors but not other dtypes.
  DEVICE_PLACEMENT_SILENT_FOR_INT32 = 3,
};
// LINT.ThenChange(//tensorflow/c/eager/c_api.h)

// Abstract interface to a context.
//
// A context is responsible for creating key objects such as Tensors,