From 07e4604b1961a32bbe21841a1e97fc274b50c443 Mon Sep 17 00:00:00 2001
From: Alejandro Acosta <alejandro@magicpony.technology>
Date: Mon, 29 May 2023 17:23:58 +0100
Subject: [PATCH] Replace usage of CudaStreamDevice with GpuStreamDevice in
 tensor benchmarks GPU

---
 bench/tensors/tensor_benchmarks_fp16_gpu.cu | 6 +++---
 bench/tensors/tensor_benchmarks_gpu.cu      | 6 +++---
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/bench/tensors/tensor_benchmarks_fp16_gpu.cu b/bench/tensors/tensor_benchmarks_fp16_gpu.cu
index 65784d0d6..d63ff8bb2 100644
--- a/bench/tensors/tensor_benchmarks_fp16_gpu.cu
+++ b/bench/tensors/tensor_benchmarks_fp16_gpu.cu
@@ -10,7 +10,7 @@
 #define BM_FuncGPU(FUNC)                                                       \
   static void BM_##FUNC(int iters, int N) {                                    \
     StopBenchmarkTiming();                                                     \
-    Eigen::CudaStreamDevice stream;                                            \
+    Eigen::GpuStreamDevice stream;                                             \
     Eigen::GpuDevice device(&stream);                                          \
     BenchmarkSuite<Eigen::GpuDevice, Eigen::half> suite(device, N);            \
     cudaDeviceSynchronize();                                                   \
@@ -40,7 +40,7 @@ BM_FuncGPU(fullReduction);
 #define BM_FuncWithInputDimsGPU(FUNC, D1, D2, D3)                              \
   static void BM_##FUNC##_##D1##x##D2##x##D3(int iters, int N) {               \
     StopBenchmarkTiming();                                                     \
-    Eigen::CudaStreamDevice stream;                                            \
+    Eigen::GpuStreamDevice stream;                                             \
     Eigen::GpuDevice device(&stream);                                          \
     BenchmarkSuite<Eigen::GpuDevice, Eigen::half> suite(device, D1, D2, D3);   \
     cudaDeviceSynchronize();                                                   \
@@ -59,7 +59,7 @@ BM_FuncWithInputDimsGPU(contraction, N, N, 64);
 #define BM_FuncWithKernelDimsGPU(FUNC, DIM1, DIM2)                             \
   static void BM_##FUNC##_##DIM1##x##DIM2(int iters, int N) {                  \
     StopBenchmarkTiming();                                                     \
-    Eigen::CudaStreamDevice stream;                                            \
+    Eigen::GpuStreamDevice stream;                                             \
     Eigen::GpuDevice device(&stream);                                          \
     BenchmarkSuite<Eigen::GpuDevice, Eigen::half> suite(device, N);            \
     cudaDeviceSynchronize();                                                   \
diff --git a/bench/tensors/tensor_benchmarks_gpu.cu b/bench/tensors/tensor_benchmarks_gpu.cu
index 76d68c5c1..c77810264 100644
--- a/bench/tensors/tensor_benchmarks_gpu.cu
+++ b/bench/tensors/tensor_benchmarks_gpu.cu
@@ -10,7 +10,7 @@
 #define BM_FuncGPU(FUNC)                                                       \
   static void BM_##FUNC(int iters, int N) {                                    \
     StopBenchmarkTiming();                                                     \
-    Eigen::CudaStreamDevice stream;                                            \
+    Eigen::GpuStreamDevice stream;                                             \
     Eigen::GpuDevice device(&stream);                                          \
     BenchmarkSuite<Eigen::GpuDevice, float> suite(device, N);                  \
     cudaDeviceSynchronize();                                                   \
@@ -40,7 +40,7 @@ BM_FuncGPU(fullReduction);
 #define BM_FuncWithInputDimsGPU(FUNC, D1, D2, D3)                              \
   static void BM_##FUNC##_##D1##x##D2##x##D3(int iters, int N) {               \
     StopBenchmarkTiming();                                                     \
-    Eigen::CudaStreamDevice stream;                                            \
+    Eigen::GpuStreamDevice stream;                                             \
     Eigen::GpuDevice device(&stream);                                          \
     BenchmarkSuite<Eigen::GpuDevice, float> suite(device, D1, D2, D3);         \
     cudaDeviceSynchronize();                                                   \
@@ -59,7 +59,7 @@ BM_FuncWithInputDimsGPU(contraction, N, N, 64);
 #define BM_FuncWithKernelDimsGPU(FUNC, DIM1, DIM2)                             \
   static void BM_##FUNC##_##DIM1##x##DIM2(int iters, int N) {                  \
     StopBenchmarkTiming();                                                     \
-    Eigen::CudaStreamDevice stream;                                            \
+    Eigen::GpuStreamDevice stream;                                             \
     Eigen::GpuDevice device(&stream);                                          \
     BenchmarkSuite<Eigen::GpuDevice, float> suite(device, N);                  \
     cudaDeviceSynchronize();                                                   \