[SYCL] Add a test for in-order queue event dependencies (#79)

Author: sergey-semenov

SYCL/Scheduler/InOrderQueueDeps.cpp

@@ -0,0 +1,72 @@
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
+// RUN: env %HOST_RUN_PLACEHOLDER %t.out
+// RUN: env SYCL_PI_TRACE=2 %CPU_RUN_PLACEHOLDER %t.out 2>&1 %CPU_CHECK_PLACEHOLDER
+// RUN: env SYCL_PI_TRACE=2 %GPU_RUN_PLACEHOLDER %t.out 2>&1 %GPU_CHECK_PLACEHOLDER
+// RUN: env SYCL_PI_TRACE=2 %ACC_RUN_PLACEHOLDER %t.out 2>&1 %ACC_CHECK_PLACEHOLDER
+
+//==----------------------- InOrderQueueDeps.cpp ---------------------------==//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include <CL/sycl.hpp>
+
+class KernelA;
+
+using namespace cl;
+
+void submitKernel(sycl::queue &Queue, sycl::buffer<int, 1> &Buf) {
+  Queue.submit([&](sycl::handler &Cgh) {
+    auto BufAcc = Buf.get_access<sycl::access::mode::read_write>(Cgh);
+    Cgh.single_task<KernelA>([=]() { (void)BufAcc[0]; });
+  });
+}
+
+int main() {
+  int val;
+  sycl::buffer<int, 1> Buf{&val, sycl::range<1>(1)};
+
+  sycl::default_selector DeviceSelector;
+  sycl::device Dev = DeviceSelector.select_device();
+  sycl::context Ctx{Dev};
+
+  sycl::queue InOrderQueueA{Ctx, Dev, sycl::property::queue::in_order()};
+  sycl::queue InOrderQueueB{Ctx, Dev, sycl::property::queue::in_order()};
+
+  // Sequential submissions to the same in-order queue should not result in any
+  // event dependencies.
+  // CHECK: piEnqueueKernelLaunch
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: : 0
+  submitKernel(InOrderQueueA, Buf);
+  // CHECK: piEnqueueKernelLaunch
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: : 0
+  submitKernel(InOrderQueueA, Buf);
+  // Submisssion to a different in-order queue should explicitly depend on the
+  // previous command group.
+  // CHECK: piEnqueueKernelLaunch
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: :
+  // CHECK-NEXT: : 1
+  submitKernel(InOrderQueueB, Buf);
+
+  return 0;
+}