device_impl.cpp

//==----------------- device_impl.cpp - SYCL device ------------------------==//
//
// 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 <detail/device_impl.hpp>
#include <detail/device_info.hpp>
#include <detail/platform_impl.hpp>
#include <detail/ur_info_code.hpp>
#include <sycl/detail/ur.hpp>
#include <sycl/device.hpp>

#include <algorithm>

namespace sycl {
inline namespace _V1 {
namespace detail {

/// Constructs a SYCL device instance using the provided
/// UR device instance.
device_impl::device_impl(ur_device_handle_t Device, platform_impl &Platform)
    : MDevice(Device), MPlatform(&Platform),
      MDeviceHostBaseTime(std::make_pair(0, 0)) {
  const AdapterPtr &Adapter = Platform.getAdapter();

  // TODO catch an exception and put it to list of asynchronous exceptions
  Adapter->call<UrApiKind::urDeviceGetInfo>(
      MDevice, UR_DEVICE_INFO_TYPE, sizeof(ur_device_type_t), &MType, nullptr);

  // No need to set MRootDevice when MAlwaysRootDevice is true
  if (!Platform.MAlwaysRootDevice) {
    // TODO catch an exception and put it to list of asynchronous exceptions
    Adapter->call<UrApiKind::urDeviceGetInfo>(
        MDevice, UR_DEVICE_INFO_PARENT_DEVICE, sizeof(ur_device_handle_t),
        &MRootDevice, nullptr);
  }

  // TODO catch an exception and put it to list of asynchronous exceptions
  // Interoperability Constructor already calls DeviceRetain in
  // urDeviceCreateWithNativeHandle.
  Adapter->call<UrApiKind::urDeviceRetain>(MDevice);

  Adapter->call<UrApiKind::urDeviceGetInfo>(
      MDevice, UR_DEVICE_INFO_USE_NATIVE_ASSERT, sizeof(ur_bool_t),
      &MUseNativeAssert, nullptr);
}

device_impl::~device_impl() {
  try {
    // TODO catch an exception and put it to list of asynchronous exceptions
    const AdapterPtr &Adapter = getAdapter();
    ur_result_t Err =
        Adapter->call_nocheck<UrApiKind::urDeviceRelease>(MDevice);
    __SYCL_CHECK_UR_CODE_NO_EXC(Err);
  } catch (std::exception &e) {
    __SYCL_REPORT_EXCEPTION_TO_STREAM("exception in ~device_impl", e);
  }
}

bool device_impl::is_affinity_supported(
    info::partition_affinity_domain AffinityDomain) const {
  auto SupportedDomains = get_info<info::device::partition_affinity_domains>();
  return std::find(SupportedDomains.begin(), SupportedDomains.end(),
                   AffinityDomain) != SupportedDomains.end();
}

cl_device_id device_impl::get() const {
  // TODO catch an exception and put it to list of asynchronous exceptions
  __SYCL_OCL_CALL(clRetainDevice, ur::cast<cl_device_id>(getNative()));
  return ur::cast<cl_device_id>(getNative());
}

platform device_impl::get_platform() const {
  return createSyclObjFromImpl<platform>(*MPlatform);
}

template <typename Param>
typename Param::return_type device_impl::get_info() const {
  return get_device_info<Param>(*this);
}
// Explicitly instantiate all device info traits
#define __SYCL_PARAM_TRAITS_SPEC(DescType, Desc, ReturnT, PiCode)              \
  template ReturnT device_impl::get_info<info::device::Desc>() const;

#define __SYCL_PARAM_TRAITS_SPEC_SPECIALIZED(DescType, Desc, ReturnT, PiCode)  \
  template ReturnT device_impl::get_info<info::device::Desc>() const;

#include <sycl/info/device_traits.def>
#undef __SYCL_PARAM_TRAITS_SPEC_SPECIALIZED
#undef __SYCL_PARAM_TRAITS_SPEC

#define __SYCL_PARAM_TRAITS_SPEC(Namespace, DescType, Desc, ReturnT, PiCode)   \
  template __SYCL_EXPORT ReturnT                                               \
  device_impl::get_info<Namespace::info::DescType::Desc>() const;

#include <sycl/info/ext_codeplay_device_traits.def>
#include <sycl/info/ext_intel_device_traits.def>
#include <sycl/info/ext_oneapi_device_traits.def>
#undef __SYCL_PARAM_TRAITS_SPEC

#ifndef __INTEL_PREVIEW_BREAKING_CHANGES
template <>
typename info::platform::version::return_type
device_impl::get_backend_info<info::platform::version>() const {
  if (getBackend() != backend::opencl) {
    throw sycl::exception(errc::backend_mismatch,
                          "the info::platform::version info descriptor can "
                          "only be queried with an OpenCL backend");
  }
  return get_platform().get_info<info::platform::version>();
}
#endif

#ifndef __INTEL_PREVIEW_BREAKING_CHANGES
template <>
typename info::device::version::return_type
device_impl::get_backend_info<info::device::version>() const {
  if (getBackend() != backend::opencl) {
    throw sycl::exception(errc::backend_mismatch,
                          "the info::device::version info descriptor can only "
                          "be queried with an OpenCL backend");
  }
  return get_info<info::device::version>();
}
#endif

#ifndef __INTEL_PREVIEW_BREAKING_CHANGES
template <>
typename info::device::backend_version::return_type
device_impl::get_backend_info<info::device::backend_version>() const {
  if (getBackend() != backend::ext_oneapi_level_zero) {
    throw sycl::exception(errc::backend_mismatch,
                          "the info::device::backend_version info descriptor "
                          "can only be queried with a Level Zero backend");
  }
  return "";
  // Currently The Level Zero backend does not define the value of this
  // information descriptor and implementations are encouraged to return the
  // empty string as per specification.
}
#endif

bool device_impl::has_extension(const std::string &ExtensionName) const {
  std::string AllExtensionNames =
      get_device_info_string(UR_DEVICE_INFO_EXTENSIONS);
  return (AllExtensionNames.find(ExtensionName) != std::string::npos);
}

bool device_impl::is_partition_supported(info::partition_property Prop) const {
  auto SupportedProperties = get_info<info::device::partition_properties>();
  return std::find(SupportedProperties.begin(), SupportedProperties.end(),
                   Prop) != SupportedProperties.end();
}

std::vector<device> device_impl::create_sub_devices(
    const ur_device_partition_properties_t *Properties,
    size_t SubDevicesCount) const {
  std::vector<ur_device_handle_t> SubDevices(SubDevicesCount);
  uint32_t ReturnedSubDevices = 0;
  const AdapterPtr &Adapter = getAdapter();
  Adapter->call<sycl::errc::invalid, UrApiKind::urDevicePartition>(
      MDevice, Properties, SubDevicesCount, SubDevices.data(),
      &ReturnedSubDevices);
  if (ReturnedSubDevices != SubDevicesCount) {
    throw sycl::exception(
        errc::invalid,
        "Could not partition to the specified number of sub-devices");
  }
  // TODO: Need to describe the subdevice model. Some sub_device management
  // may be necessary. What happens if create_sub_devices is called multiple
  // times with the same arguments?
  //
  std::vector<device> res;
  std::for_each(SubDevices.begin(), SubDevices.end(),
                [&res, this](const ur_device_handle_t &a_ur_device) {
                  device sycl_device = detail::createSyclObjFromImpl<device>(
                      MPlatform->getOrMakeDeviceImpl(a_ur_device, *MPlatform));
                  res.push_back(sycl_device);
                });
  return res;
}

std::vector<device> device_impl::create_sub_devices(size_t ComputeUnits) const {
  if (!is_partition_supported(info::partition_property::partition_equally)) {
    throw sycl::exception(make_error_code(errc::feature_not_supported),
                          "Device does not support "
                          "sycl::info::partition_property::partition_equally.");
  }
  // If count exceeds the total number of compute units in the device, an
  // exception with the errc::invalid error code must be thrown.
  auto MaxComputeUnits = get_info<info::device::max_compute_units>();
  if (ComputeUnits > MaxComputeUnits)
    throw sycl::exception(errc::invalid,
                          "Total counts exceed max compute units");

  size_t SubDevicesCount = MaxComputeUnits / ComputeUnits;

  ur_device_partition_property_t Prop{};
  Prop.type = UR_DEVICE_PARTITION_EQUALLY;
  Prop.value.count = static_cast<uint32_t>(ComputeUnits);

  ur_device_partition_properties_t Properties{};
  Properties.stype = UR_STRUCTURE_TYPE_DEVICE_PARTITION_PROPERTIES;
  Properties.PropCount = 1;
  Properties.pProperties = &Prop;

  return create_sub_devices(&Properties, SubDevicesCount);
}

std::vector<device>
device_impl::create_sub_devices(const std::vector<size_t> &Counts) const {
  if (!is_partition_supported(info::partition_property::partition_by_counts)) {
    throw sycl::exception(
        make_error_code(errc::feature_not_supported),
        "Device does not support "
        "sycl::info::partition_property::partition_by_counts.");
  }

  std::vector<ur_device_partition_property_t> Props{};

  // Fill the properties vector with counts and validate it
  size_t TotalCounts = 0;
  size_t NonZeroCounts = 0;
  for (auto Count : Counts) {
    TotalCounts += Count;
    NonZeroCounts += (Count != 0) ? 1 : 0;
    Props.push_back(ur_device_partition_property_t{
        UR_DEVICE_PARTITION_BY_COUNTS, {static_cast<uint32_t>(Count)}});
  }

  ur_device_partition_properties_t Properties{};
  Properties.stype = UR_STRUCTURE_TYPE_DEVICE_PARTITION_PROPERTIES;
  Properties.pProperties = Props.data();
  Properties.PropCount = Props.size();

  // If the number of non-zero values in counts exceeds the device’s maximum
  // number of sub devices (as returned by info::device::
  // partition_max_sub_devices) an exception with the errc::invalid
  // error code must be thrown.
  if (NonZeroCounts > get_info<info::device::partition_max_sub_devices>())
    throw sycl::exception(errc::invalid,
                          "Total non-zero counts exceed max sub-devices");

  // If the total of all the values in the counts vector exceeds the total
  // number of compute units in the device (as returned by
  // info::device::max_compute_units), an exception with the errc::invalid
  // error code must be thrown.
  if (TotalCounts > get_info<info::device::max_compute_units>())
    throw sycl::exception(errc::invalid,
                          "Total counts exceed max compute units");

  return create_sub_devices(&Properties, Counts.size());
}

std::vector<device> device_impl::create_sub_devices(
    info::partition_affinity_domain AffinityDomain) const {
  if (!is_partition_supported(
          info::partition_property::partition_by_affinity_domain)) {
    throw sycl::exception(
        make_error_code(errc::feature_not_supported),
        "Device does not support "
        "sycl::info::partition_property::partition_by_affinity_domain.");
  }
  if (!is_affinity_supported(AffinityDomain)) {
    throw sycl::exception(make_error_code(errc::feature_not_supported),
                          "Device does not support " +
                              affinityDomainToString(AffinityDomain) + ".");
  }

  ur_device_partition_property_t Prop;
  Prop.type = UR_DEVICE_PARTITION_BY_AFFINITY_DOMAIN;
  Prop.value.affinity_domain =
      static_cast<ur_device_affinity_domain_flags_t>(AffinityDomain);

  ur_device_partition_properties_t Properties{};
  Properties.stype = UR_STRUCTURE_TYPE_DEVICE_PARTITION_PROPERTIES;
  Properties.PropCount = 1;
  Properties.pProperties = &Prop;

  uint32_t SubDevicesCount = 0;
  const AdapterPtr &Adapter = getAdapter();
  Adapter->call<sycl::errc::invalid, UrApiKind::urDevicePartition>(
      MDevice, &Properties, 0, nullptr, &SubDevicesCount);

  return create_sub_devices(&Properties, SubDevicesCount);
}

std::vector<device> device_impl::create_sub_devices() const {
  if (!is_partition_supported(
          info::partition_property::ext_intel_partition_by_cslice)) {
    throw sycl::exception(
        make_error_code(errc::feature_not_supported),
        "Device does not support "
        "sycl::info::partition_property::ext_intel_partition_by_cslice.");
  }

  ur_device_partition_property_t Prop;
  Prop.type = UR_DEVICE_PARTITION_BY_CSLICE;

  ur_device_partition_properties_t Properties{};
  Properties.stype = UR_STRUCTURE_TYPE_DEVICE_PARTITION_PROPERTIES;
  Properties.pProperties = &Prop;
  Properties.PropCount = 1;

  uint32_t SubDevicesCount = 0;
  const AdapterPtr &Adapter = getAdapter();
  Adapter->call<UrApiKind::urDevicePartition>(MDevice, &Properties, 0, nullptr,
                                              &SubDevicesCount);

  return create_sub_devices(&Properties, SubDevicesCount);
}

ur_native_handle_t device_impl::getNative() const {
  auto Adapter = getAdapter();
  ur_native_handle_t Handle;
  Adapter->call<UrApiKind::urDeviceGetNativeHandle>(getHandleRef(), &Handle);
  if (getBackend() == backend::opencl) {
    __SYCL_OCL_CALL(clRetainDevice, ur::cast<cl_device_id>(Handle));
  }
  return Handle;
}

bool device_impl::has(aspect Aspect) const {
  size_t return_size = 0;

  switch (Aspect) {
  case aspect::host:
    // Deprecated
    return false;
  case aspect::cpu:
    return is_cpu();
  case aspect::gpu:
    return is_gpu();
  case aspect::accelerator:
    return is_accelerator();
  case aspect::custom:
    return false;
  // TODO: Implement this for FPGA emulator.
  case aspect::emulated:
    return false;
  case aspect::host_debuggable:
    return false;
  case aspect::fp16:
    return has_extension("cl_khr_fp16");
  case aspect::fp64:
    return has_extension("cl_khr_fp64");
  case aspect::int64_base_atomics:
    return has_extension("cl_khr_int64_base_atomics");
  case aspect::int64_extended_atomics:
    return has_extension("cl_khr_int64_extended_atomics");
  case aspect::atomic64:
    return get_info<info::device::atomic64>();
  case aspect::image:
    return get_info<info::device::image_support>();
  case aspect::online_compiler:
    return get_info<info::device::is_compiler_available>();
  case aspect::online_linker:
    return get_info<info::device::is_linker_available>();
  case aspect::queue_profiling:
    return get_info<info::device::queue_profiling>();
  case aspect::usm_device_allocations:
    return get_info<info::device::usm_device_allocations>();
  case aspect::usm_host_allocations:
    return get_info<info::device::usm_host_allocations>();
  case aspect::ext_intel_mem_channel:
    return get_info<info::device::ext_intel_mem_channel>();
  case aspect::ext_oneapi_cuda_cluster_group:
    return get_info<info::device::ext_oneapi_cuda_cluster_group>();
  case aspect::usm_atomic_host_allocations:
    return (
        get_device_info_impl<ur_device_usm_access_capability_flags_t,
                             info::device::usm_host_allocations>::get(*this) &
        UR_DEVICE_USM_ACCESS_CAPABILITY_FLAG_ATOMIC_CONCURRENT_ACCESS);
  case aspect::usm_shared_allocations:
    return get_info<info::device::usm_shared_allocations>();
  case aspect::usm_atomic_shared_allocations:
    return (
        get_device_info_impl<ur_device_usm_access_capability_flags_t,
                             info::device::usm_shared_allocations>::get(*this) &
        UR_DEVICE_USM_ACCESS_CAPABILITY_FLAG_ATOMIC_CONCURRENT_ACCESS);
  case aspect::usm_restricted_shared_allocations:
    return get_info<info::device::usm_restricted_shared_allocations>();
  case aspect::usm_system_allocations:
    return get_info<info::device::usm_system_allocations>();
  case aspect::ext_intel_device_id:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_DEVICE_ID, 0, nullptr, &return_size) ==
           UR_RESULT_SUCCESS;
  case aspect::ext_intel_pci_address:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_PCI_ADDRESS, 0, nullptr, &return_size) ==
           UR_RESULT_SUCCESS;
  case aspect::ext_intel_gpu_eu_count:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_GPU_EU_COUNT, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_gpu_eu_simd_width:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_GPU_EU_SIMD_WIDTH, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_gpu_slices:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_GPU_EU_SLICES, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_gpu_subslices_per_slice:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_GPU_SUBSLICES_PER_SLICE, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_gpu_eu_count_per_subslice:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_GPU_EU_COUNT_PER_SUBSLICE, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_gpu_hw_threads_per_eu:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_GPU_HW_THREADS_PER_EU, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_free_memory:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_GLOBAL_MEM_FREE, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_memory_clock_rate:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_MEMORY_CLOCK_RATE, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_memory_bus_width:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_MEMORY_BUS_WIDTH, 0, nullptr,
               &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_device_info_uuid: {
    auto Result = getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
        MDevice, UR_DEVICE_INFO_UUID, 0, nullptr, &return_size);
    if (Result != UR_RESULT_SUCCESS) {
      return false;
    }

    assert(return_size <= 16);
    unsigned char UUID[16];

    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_UUID, 16 * sizeof(unsigned char), UUID,
               nullptr) == UR_RESULT_SUCCESS;
  }
  case aspect::ext_intel_max_mem_bandwidth:
    // currently not supported
    return false;
  case aspect::ext_intel_current_clock_throttle_reasons:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_CURRENT_CLOCK_THROTTLE_REASONS, 0,
               nullptr, &return_size) == UR_RESULT_SUCCESS;
  case aspect::ext_intel_fan_speed:
    return getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
               MDevice, UR_DEVICE_INFO_FAN_SPEED, 0, nullptr, &return_size) ==
           UR_RESULT_SUCCESS;
  case aspect::ext_intel_power_limits:
    return (getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
                MDevice, UR_DEVICE_INFO_MIN_POWER_LIMIT, 0, nullptr,
                &return_size) == UR_RESULT_SUCCESS) &&
           (getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
                MDevice, UR_DEVICE_INFO_MAX_POWER_LIMIT, 0, nullptr,
                &return_size) == UR_RESULT_SUCCESS);
  case aspect::ext_oneapi_srgb:
    return get_info<info::device::ext_oneapi_srgb>();
  case aspect::ext_oneapi_native_assert:
    return useNativeAssert();
  case aspect::ext_oneapi_cuda_async_barrier: {
    int async_barrier_supported;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_ASYNC_BARRIER, sizeof(int),
            &async_barrier_supported, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && async_barrier_supported;
  }
  case aspect::ext_intel_legacy_image: {
    ur_bool_t legacy_image_support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_IMAGE_SUPPORT, sizeof(ur_bool_t),
            &legacy_image_support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && legacy_image_support;
  }
  case aspect::ext_oneapi_bindless_images: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_IMAGES_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_images_shared_usm: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_IMAGES_SHARED_USM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_images_1d_usm: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_IMAGES_1D_USM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_images_2d_usm: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_IMAGES_2D_USM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_external_memory_import: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_EXTERNAL_MEMORY_IMPORT_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_external_semaphore_import: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_EXTERNAL_SEMAPHORE_IMPORT_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_mipmap: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_MIPMAP_SUPPORT_EXP, sizeof(ur_bool_t),
            &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_mipmap_anisotropy: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_MIPMAP_ANISOTROPY_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_mipmap_level_reference: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_MIPMAP_LEVEL_REFERENCE_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_sampled_image_fetch_1d_usm: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice,
            UR_DEVICE_INFO_BINDLESS_SAMPLED_IMAGE_FETCH_1D_USM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_sampled_image_fetch_1d: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_SAMPLED_IMAGE_FETCH_1D_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_sampled_image_fetch_2d_usm: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice,
            UR_DEVICE_INFO_BINDLESS_SAMPLED_IMAGE_FETCH_2D_USM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_sampled_image_fetch_2d: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_SAMPLED_IMAGE_FETCH_2D_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_sampled_image_fetch_3d: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_SAMPLED_IMAGE_FETCH_3D_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_images_gather: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_IMAGES_GATHER_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_cubemap: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_CUBEMAP_SUPPORT_EXP, sizeof(ur_bool_t),
            &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_cubemap_seamless_filtering: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_CUBEMAP_SEAMLESS_FILTERING_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_image_array: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_IMAGE_ARRAY_SUPPORT_EXP, sizeof(ur_bool_t),
            &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_unique_addressing_per_dim: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice,
            UR_DEVICE_INFO_BINDLESS_UNIQUE_ADDRESSING_PER_DIM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_images_sample_1d_usm: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_SAMPLE_1D_USM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_bindless_images_sample_2d_usm: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_BINDLESS_SAMPLE_2D_USM_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_intel_esimd: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_ESIMD_SUPPORT, sizeof(ur_bool_t), &support,
            nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_ballot_group:
  case aspect::ext_oneapi_fixed_size_group:
  case aspect::ext_oneapi_opportunistic_group: {
    return (this->getBackend() == backend::ext_oneapi_level_zero) ||
           (this->getBackend() == backend::opencl) ||
           (this->getBackend() == backend::ext_oneapi_cuda);
  }
  case aspect::ext_oneapi_tangle_group: {
    // TODO: tangle_group is not currently supported for CUDA devices. Add when
    //       implemented.
    return (this->getBackend() == backend::ext_oneapi_level_zero) ||
           (this->getBackend() == backend::opencl);
  }
  case aspect::ext_intel_matrix: {
    using arch = sycl::ext::oneapi::experimental::architecture;
    const arch supported_archs[] = {
        arch::intel_cpu_spr,     arch::intel_cpu_gnr,
        arch::intel_cpu_dmr,     arch::intel_gpu_pvc,
        arch::intel_gpu_dg2_g10, arch::intel_gpu_dg2_g11,
        arch::intel_gpu_dg2_g12, arch::intel_gpu_bmg_g21,
        arch::intel_gpu_lnl_m,   arch::intel_gpu_arl_h,
        arch::intel_gpu_ptl_h,   arch::intel_gpu_ptl_u,
    };
    try {
      return std::any_of(
          std::begin(supported_archs), std::end(supported_archs),
          [=](const arch a) { return this->extOneapiArchitectureIs(a); });
    } catch (const sycl::exception &) {
      // If we're here it means the device does not support architecture
      // querying
      return false;
    }
  }
  case aspect::ext_oneapi_is_composite: {
    auto components = get_info<
        sycl::ext::oneapi::experimental::info::device::component_devices>();
    // Any device with ext_oneapi_is_composite aspect will have at least two
    // constituent component devices.
    return components.size() >= 2;
  }
  case aspect::ext_oneapi_is_component: {
    typename sycl_to_ur<device>::type Result;
    bool CallSuccessful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            getHandleRef(),
            UrInfoCode<ext::oneapi::experimental::info::device::
                           composite_device>::value,
            sizeof(Result), &Result, nullptr) == UR_RESULT_SUCCESS;

    return CallSuccessful && Result != nullptr;
  }
  case aspect::ext_oneapi_graph: {
    ur_device_command_buffer_update_capability_flags_t UpdateCapabilities;
    bool CallSuccessful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_COMMAND_BUFFER_UPDATE_CAPABILITIES_EXP,
            sizeof(UpdateCapabilities), &UpdateCapabilities,
            nullptr) == UR_RESULT_SUCCESS;
    if (!CallSuccessful) {
      return false;
    }

    /* The kernel handle update capability is not yet required for the
     * ext_oneapi_graph aspect */
    ur_device_command_buffer_update_capability_flags_t RequiredCapabilities =
        UR_DEVICE_COMMAND_BUFFER_UPDATE_CAPABILITY_FLAG_KERNEL_ARGUMENTS |
        UR_DEVICE_COMMAND_BUFFER_UPDATE_CAPABILITY_FLAG_LOCAL_WORK_SIZE |
        UR_DEVICE_COMMAND_BUFFER_UPDATE_CAPABILITY_FLAG_GLOBAL_WORK_SIZE |
        UR_DEVICE_COMMAND_BUFFER_UPDATE_CAPABILITY_FLAG_GLOBAL_WORK_OFFSET |
        UR_DEVICE_COMMAND_BUFFER_UPDATE_CAPABILITY_FLAG_KERNEL_HANDLE;

    return has(aspect::ext_oneapi_limited_graph) &&
           (UpdateCapabilities & RequiredCapabilities) == RequiredCapabilities;
  }
  case aspect::ext_oneapi_limited_graph: {
    bool SupportsCommandBuffers = false;
    bool CallSuccessful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_COMMAND_BUFFER_SUPPORT_EXP,
            sizeof(SupportsCommandBuffers), &SupportsCommandBuffers,
            nullptr) == UR_RESULT_SUCCESS;
    if (!CallSuccessful) {
      return false;
    }

    return SupportsCommandBuffers;
  }
  case aspect::ext_oneapi_private_alloca: {
    // Extension only supported on SPIR-V targets.
    backend be = getBackend();
    return be == sycl::backend::ext_oneapi_level_zero ||
           be == sycl::backend::opencl;
  }
  case aspect::ext_oneapi_queue_profiling_tag: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_TIMESTAMP_RECORDING_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_oneapi_virtual_mem: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_VIRTUAL_MEMORY_SUPPORT, sizeof(ur_bool_t),
            &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  case aspect::ext_intel_fpga_task_sequence: {
    return is_accelerator();
  }
  case aspect::ext_oneapi_atomic16: {
    // Likely L0 doesn't check it properly. Need to double-check.
    return has_extension("cl_ext_float_atomics");
  }
  case aspect::ext_oneapi_virtual_functions: {
    // TODO: move to UR like e.g. aspect::ext_oneapi_virtual_mem
    backend BE = getBackend();
    bool isCompatibleBE = BE == sycl::backend::ext_oneapi_level_zero ||
                          BE == sycl::backend::opencl;
    return (is_cpu() || is_gpu()) && isCompatibleBE;
  }
  case aspect::ext_intel_spill_memory_size: {
    backend BE = getBackend();
    bool isCompatibleBE = BE == sycl::backend::ext_oneapi_level_zero;
    return is_gpu() && isCompatibleBE;
  }
  case aspect::ext_oneapi_async_memory_alloc: {
    ur_bool_t support = false;
    bool call_successful =
        getAdapter()->call_nocheck<UrApiKind::urDeviceGetInfo>(
            MDevice, UR_DEVICE_INFO_ASYNC_USM_ALLOCATIONS_SUPPORT_EXP,
            sizeof(ur_bool_t), &support, nullptr) == UR_RESULT_SUCCESS;
    return call_successful && support;
  }
  }

  return false; // This device aspect has not been implemented yet.
}

bool device_impl::useNativeAssert() const { return MUseNativeAssert; }

std::string device_impl::getDeviceName() const {
  std::call_once(MDeviceNameFlag,
                 [this]() { MDeviceName = get_info<info::device::name>(); });

  return MDeviceName;
}

ext::oneapi::experimental::architecture device_impl::getDeviceArch() const {
  std::call_once(MDeviceArchFlag, [this]() {
    MDeviceArch =
        get_info<ext::oneapi::experimental::info::device::architecture>();
  });

  return MDeviceArch;
}

// On the first call this function queries for device timestamp
// along with host synchronized timestamp and stores it in member variable
// MDeviceHostBaseTime. Subsequent calls to this function would just retrieve
// the host timestamp, compute difference against the host timestamp in
// MDeviceHostBaseTime and calculate the device timestamp based on the
// difference.
//
// The MDeviceHostBaseTime is refreshed with new device and host timestamp
// after a certain interval (determined by TimeTillRefresh) to account for
// clock drift between host and device.
//
uint64_t device_impl::getCurrentDeviceTime() {
  using namespace std::chrono;
  uint64_t HostTime =
      duration_cast<nanoseconds>(steady_clock::now().time_since_epoch())
          .count();

  // To account for potential clock drift between host clock and device clock.
  // The value set is arbitrary: 200 seconds
  constexpr uint64_t TimeTillRefresh = 200e9;
  assert(HostTime >= MDeviceHostBaseTime.second);
  uint64_t Diff = HostTime - MDeviceHostBaseTime.second;

  // If getCurrentDeviceTime is called for the first time or we have to refresh.
  if (!MDeviceHostBaseTime.second || Diff > TimeTillRefresh) {
    const auto &Adapter = getAdapter();
    auto Result = Adapter->call_nocheck<UrApiKind::urDeviceGetGlobalTimestamps>(
        MDevice, &MDeviceHostBaseTime.first, &MDeviceHostBaseTime.second);
    // We have to remember base host timestamp right after UR call and it is
    // going to be used for calculation of the device timestamp at the next
    // getCurrentDeviceTime() call. We need to do it here because getAdapter()
    // and urDeviceGetGlobalTimestamps calls may take significant amount of
    // time, for example on the first call to getAdapter adapters may need to be
    // initialized. If we use timestamp from the beginning of the function then
    // the difference between host timestamps of the current
    // getCurrentDeviceTime and the next getCurrentDeviceTime will be incorrect
    // because it will include execution time of the code before we get device
    // timestamp from urDeviceGetGlobalTimestamps.
    HostTime =
        duration_cast<nanoseconds>(steady_clock::now().time_since_epoch())
            .count();
    if (Result == UR_RESULT_ERROR_INVALID_OPERATION) {
      // NOTE(UR port): Removed the call to GetLastError because  we shouldn't
      // be calling it after ERROR_INVALID_OPERATION: there is no
      // adapter-specific error.
      throw detail::set_ur_error(
          sycl::exception(
              make_error_code(errc::feature_not_supported),
              "Device and/or backend does not support querying timestamp."),
          UR_RESULT_ERROR_INVALID_OPERATION);
    } else {
      Adapter->checkUrResult<errc::feature_not_supported>(Result);
    }
    // Until next sync we will compute device time based on the host time
    // returned in HostTime, so make this our base host time.
    MDeviceHostBaseTime.second = HostTime;
    Diff = 0;
  }
  return MDeviceHostBaseTime.first + Diff;
}

bool device_impl::isGetDeviceAndHostTimerSupported() {
  const auto &Adapter = getAdapter();
  uint64_t DeviceTime = 0, HostTime = 0;
  auto Result = Adapter->call_nocheck<UrApiKind::urDeviceGetGlobalTimestamps>(
      MDevice, &DeviceTime, &HostTime);
  return Result != UR_RESULT_ERROR_INVALID_OPERATION;
}

bool device_impl::extOneapiCanCompile(
    ext::oneapi::experimental::source_language Language) {
  try {
    return sycl::ext::oneapi::experimental::detail::
        is_source_kernel_bundle_supported(getBackend(), Language);
  } catch (sycl::exception &) {
    return false;
  }
}

// Returns the strongest guarantee that can be provided by the host device for
// threads created at threadScope from a coordination scope given by
// coordinationScope
sycl::ext::oneapi::experimental::forward_progress_guarantee
device_impl::getHostProgressGuarantee(
    ext::oneapi::experimental::execution_scope,
    ext::oneapi::experimental::execution_scope) {
  return sycl::ext::oneapi::experimental::forward_progress_guarantee::
      weakly_parallel;
}

// Returns the strongest progress guarantee that can be provided by this device
// for threads created at threadScope from the coordination scope given by
// coordinationScope.
sycl::ext::oneapi::experimental::forward_progress_guarantee
device_impl::getProgressGuarantee(
    ext::oneapi::experimental::execution_scope threadScope,
    ext::oneapi::experimental::execution_scope coordinationScope) const {
  using forward_progress_guarantee =
      ext::oneapi::experimental::forward_progress_guarantee;
  using execution_scope = ext::oneapi::experimental::execution_scope;
  const int executionScopeSize = 4;
  (void)coordinationScope;
  int threadScopeNum = static_cast<int>(threadScope);
  // we get the immediate progress guarantee that is provided by each scope
  // between root_group and threadScope and then return the weakest of these.
  // Counterintuitively, this corresponds to taking the max of the enum values
  // because of how the forward_progress_guarantee enum values are declared.
  int guaranteeNum = static_cast<int>(
      getImmediateProgressGuarantee(execution_scope::root_group));
  for (int currentScope = executionScopeSize - 2; currentScope > threadScopeNum;
       --currentScope) {
    guaranteeNum = std::max(guaranteeNum,
                            static_cast<int>(getImmediateProgressGuarantee(
                                static_cast<execution_scope>(currentScope))));
  }
  return static_cast<forward_progress_guarantee>(guaranteeNum);
}

bool device_impl::supportsForwardProgress(
    ext::oneapi::experimental::forward_progress_guarantee guarantee,
    ext::oneapi::experimental::execution_scope threadScope,
    ext::oneapi::experimental::execution_scope coordinationScope) const {
  using ReturnT =
      std::vector<ext::oneapi::experimental::forward_progress_guarantee>;
  auto guarantees = getProgressGuaranteesUpTo<ReturnT>(
      getProgressGuarantee(threadScope, coordinationScope));
  return std::find(guarantees.begin(), guarantees.end(), guarantee) !=
         guarantees.end();
}

// Returns the progress guarantee provided for a coordination scope
// given by coordination_scope for threads created at a scope
// immediately below coordination_scope. For example, for root_group
// coordination scope it returns the progress guarantee provided
// at root_group for threads created at work_group.
ext::oneapi::experimental::forward_progress_guarantee
device_impl::getImmediateProgressGuarantee(
    ext::oneapi::experimental::execution_scope coordination_scope) const {
  using forward_progress_guarantee =
      ext::oneapi::experimental::forward_progress_guarantee;
  using execution_scope = ext::oneapi::experimental::execution_scope;
  if (is_cpu() && getBackend() == backend::opencl) {
    switch (coordination_scope) {
    case execution_scope::root_group:
      return forward_progress_guarantee::parallel;
    case execution_scope::work_group:
    case execution_scope::sub_group:
      return forward_progress_guarantee::weakly_parallel;
    default:
      throw sycl::exception(sycl::errc::invalid,
                            "Work item is not a valid coordination scope!");
    }
  } else if (is_gpu() && getBackend() == backend::ext_oneapi_level_zero) {
    switch (coordination_scope) {
    case execution_scope::root_group:
    case execution_scope::work_group:
      return forward_progress_guarantee::concurrent;
    case execution_scope::sub_group:
      return forward_progress_guarantee::weakly_parallel;
    default:
      throw sycl::exception(sycl::errc::invalid,
                            "Work item is not a valid coordination scope!");
    }
  }
  return forward_progress_guarantee::weakly_parallel;
}

} // namespace detail
} // namespace _V1
} // namespace sycl