Function BM_Sum2DRowCumsumGPU

tensorflow/core/kernels/scan_ops_test.cc:111–113

BENCHMARK(BM_OneDCumsumGPUHalf)->Range(1, 1 << 21);

static void BM_Sum2DRowCumsumGPU(int iters, int num_x, int num_y) {
  DoRowCumsum(iters, "gpu", num_x, num_y);
}
BENCHMARK(BM_Sum2DRowCumsumGPU)->RangePair(1, 8192, 1, 8192);

static void BM_Sum2DColumnCumsumGPU(int iters, int num_x, int num_y) {