Attribute hipDevAttrConcurrentKernels for obtaining Device property concurrentKernels is added.

include/hip_runtime_api.h

@@ -149,6 +149,7 @@ typedef enum hipDeviceAttribute_t {
     hipDeviceAttributeMaxThreadsPerMultiProcessor,          ///< Maximum resident threads per multiprocessor.
     hipDeviceAttributeComputeCapabilityMajor,               ///< Major compute capability version number.
     hipDeviceAttributeComputeCapabilityMinor,               ///< Minor compute capability version number.
+    hipDevAttrConcurrentKernels,                            ///< Device can possibly execute multiple kernels concurrently.
     hipDeviceAttributePciBusId,                             ///< PCI Bus ID.
     hipDeviceAttributePciDeviceId,                          ///< PCI Device ID.
     hipDeviceAttributeMaxSharedMemoryPerMultiprocessor,     ///< Maximum Shared Memory Per Multiprocessor.

include/nvcc_detail/hip_runtime_api.h

@@ -252,6 +252,8 @@ inline static hipError_t hipDeviceGetAttribute(int* pi, hipDeviceAttribute_t att
         cdattr = cudaDevAttrMaxThreadsPerMultiProcessor; break;
     case hipDeviceAttributeComputeCapabilityMajor:
         cdattr = cudaDevAttrComputeCapabilityMajor; break;
+    case hipDevAttrConcurrentKernels:
+        cdattr = cudaDevAttrConcurrentKernels; break;
     case hipDeviceAttributePciBusId:
         cdattr = cudaDevAttrPciBusId; break;
     case hipDeviceAttributePciDeviceId:

src/hip_hcc.cpp

@@ -852,6 +852,8 @@ hipError_t hipDeviceGetAttribute(int* pi, hipDeviceAttribute_t attr, int device)
             *pi = prop->minor; break;
         case hipDeviceAttributePciBusId:
             *pi = prop->pciBusID; break;
+        case hipDevAttrConcurrentKernels:
+            *pi = prop->concurrentKernels; break;
         case hipDeviceAttributePciDeviceId:
             *pi = prop->pciDeviceID; break;
         case hipDeviceAttributeMaxSharedMemoryPerMultiprocessor:

tests/src/hipGetDeviceAttribute.cpp

@@ -73,6 +73,7 @@ int main(int argc, char *argv[])
     CHECK(test_hipDeviceGetAttribute(deviceId, hipDeviceAttributeMaxThreadsPerMultiProcessor, props.maxThreadsPerMultiProcessor));
     CHECK(test_hipDeviceGetAttribute(deviceId, hipDeviceAttributeComputeCapabilityMajor, props.major));
     CHECK(test_hipDeviceGetAttribute(deviceId, hipDeviceAttributeComputeCapabilityMinor, props.minor));
+    CHECK(test_hipDeviceGetAttribute(deviceId, hipDevAttrConcurrentKernels, props.concurrentKernels));
     CHECK(test_hipDeviceGetAttribute(deviceId, hipDeviceAttributePciBusId, props.pciBusID));
     CHECK(test_hipDeviceGetAttribute(deviceId, hipDeviceAttributePciDeviceId, props.pciDeviceID));
     CHECK(test_hipDeviceGetAttribute(deviceId, hipDeviceAttributeMaxSharedMemoryPerMultiprocessor, props.maxSharedMemoryPerMultiProcessor));