Skip to content

vllm.platforms.xpu

logger module-attribute

logger = init_logger(__name__)

XPUPlatform

Bases: Platform

Source code in vllm/platforms/xpu.py
class XPUPlatform(Platform):
    _enum = PlatformEnum.XPU
    device_name: str = "xpu"
    device_type: str = "xpu"
    dispatch_key: str = "XPU"
    # Intel XPU's device key is "GPU" for Ray.
    # see https://github.com/ray-project/ray/blob/6a5eb5865eeb9ccf058a79b44f107e327e360673/python/ray/_private/accelerators/intel_gpu.py#L20 # noqa: E501
    ray_device_key: str = "GPU"
    device_control_env_var: str = "ONEAPI_DEVICE_SELECTOR"

    @classmethod
    def get_attn_backend_cls(cls, selected_backend: _Backend, head_size: int,
                             dtype: torch.dtype, kv_cache_dtype: Optional[str],
                             block_size: int, use_v1: bool,
                             use_mla: bool) -> str:
        if selected_backend != _Backend.IPEX:
            logger.info("Cannot use %s backend on XPU.", selected_backend)
        use_v1 = envs.VLLM_USE_V1
        if use_v1:
            logger.info("Using Flash Attention backend on V1 engine.")
            return "vllm.v1.attention.backends.flash_attn.FlashAttentionBackend"
        else:
            logger.info("Using IPEX attention backend.")
            return "vllm.attention.backends.ipex_attn.IpexAttnBackend"

    @classmethod
    def get_device_capability(
        cls,
        device_id: int = 0,
    ) -> Optional[DeviceCapability]:
        # capacity format differs from cuda's and will cause unexpected
        # failure, so use None directly
        return None

    @classmethod
    def get_device_name(cls, device_id: int = 0) -> str:
        return torch.xpu.get_device_name(device_id)

    @classmethod
    def get_device_total_memory(cls, device_id: int = 0) -> int:
        device_props = torch.xpu.get_device_properties(device_id)
        return device_props.total_memory

    @classmethod
    def is_async_output_supported(cls, enforce_eager: Optional[bool]) -> bool:
        return True

    @classmethod
    def inference_mode(cls):
        return torch.no_grad()

    @classmethod
    def check_and_update_config(cls, vllm_config: VllmConfig) -> None:
        cache_config = vllm_config.cache_config
        # in V1(or with ipex chunked prefill) block_size is 64
        if cache_config and cache_config.block_size is None:
            if envs.VLLM_USE_V1:
                cache_config.block_size = 64
            else:
                cache_config.block_size = 16

        # Instances created using VllmConfig() typically have model_config as
        # None by default. The modification involves adding a check to prevent
        # potential null exceptions check and update model config.
        if vllm_config.model_config is not None:
            model_config = vllm_config.model_config
            if model_config.dtype == torch.bfloat16:
                bf16_supported = cls.device_support_bf16()
                if not bf16_supported:
                    model_config.dtype = torch.float16
            if not model_config.enforce_eager:
                logger.warning(
                    "CUDA graph is not supported on XPU, fallback to the eager "
                    "mode.")
                model_config.enforce_eager = True

        if vllm_config.speculative_config is not None:
            raise NotImplementedError(
                "XPU does not support speculative decoding")

        if vllm_config.device_config is not None:
            assert vllm_config.device_config.device_type == "xpu"

        # check and update parallel config
        parallel_config = vllm_config.parallel_config
        if envs.VLLM_USE_V1:
            parallel_config.worker_cls =\
                "vllm.v1.worker.xpu_worker.XPUWorker"
        else:
            parallel_config.worker_cls = "vllm.worker.xpu_worker.XPUWorker"

        if parallel_config.distributed_executor_backend is None:
            if parallel_config.world_size > 1:
                parallel_config.distributed_executor_backend = "ray"
            else:
                parallel_config.distributed_executor_backend = "uni"
        elif parallel_config.distributed_executor_backend == "mp":
            # FIXME(kunshang):
            # spawn needs calling `if __name__ == '__main__':``
            # fork is not supported for xpu start new process.
            if envs.VLLM_WORKER_MULTIPROC_METHOD != "spawn":
                os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
                logger.warning(
                    "Please use spawn as start method if you want to use mp.")
        elif parallel_config.distributed_executor_backend != "ray" and \
                parallel_config.distributed_executor_backend != "uni":
            logger.warning(
                "%s is not supported on XPU, fallback to ray distributed"
                " executor backend.",
                parallel_config.distributed_executor_backend)
            parallel_config.distributed_executor_backend = "ray"

        if vllm_config.model_config and vllm_config.model_config.use_mla:
            logger.info(
                "MLA is enabled on a non-GPU platform; forcing chunked "
                "prefill and prefix caching to be disabled.")
            vllm_config.scheduler_config.enable_chunked_prefill = False
            vllm_config.scheduler_config.chunked_prefill_enabled = False
            vllm_config.scheduler_config.max_num_batched_tokens = max(
                vllm_config.scheduler_config.max_model_len,
                DEFAULT_MAX_NUM_BATCHED_TOKENS)

    @classmethod
    def is_pin_memory_available(cls):
        logger.warning("Pin memory is not supported on XPU.")
        return False

    @classmethod
    def get_current_memory_usage(cls,
                                 device: Optional[torch.types.Device] = None
                                 ) -> float:
        torch.xpu.reset_peak_memory_stats(device)
        return torch.xpu.max_memory_allocated(device)

    @classmethod
    def device_support_bf16(cls) -> bool:
        device_name = cls.get_device_name().lower()
        if cls.is_client_gpu_a770():
            logger.warning("Intel Arc A770 have bfloat16 accuracy known issue,"
                           " fallback to float16")
            return False
        else:
            logger.info(
                "Device name %s supports bfloat16. Please file an issue "
                "if you encounter any accuracy problems with bfloat16.",
                device_name)
            return True

    @classmethod
    def is_data_center_gpu(cls) -> bool:
        device_name = cls.get_device_name().lower()
        return device_name.count("data center gpu") > 0

    @classmethod
    def is_client_gpu_a770(cls) -> bool:
        device_name = cls.get_device_name().lower()
        return device_name.count("a770") > 0

    @classmethod
    def get_device_communicator_cls(cls) -> str:
        return "vllm.distributed.device_communicators.xpu_communicator.XpuCommunicator"  # noqa

    @classmethod
    def supports_v1(cls, model_config: ModelConfig) -> bool:
        return True

    @classmethod
    def device_count(cls) -> int:
        return torch.xpu.device_count()

_enum class-attribute instance-attribute

_enum = XPU

device_control_env_var class-attribute instance-attribute

device_control_env_var: str = 'ONEAPI_DEVICE_SELECTOR'

device_name class-attribute instance-attribute

device_name: str = 'xpu'

device_type class-attribute instance-attribute

device_type: str = 'xpu'

dispatch_key class-attribute instance-attribute

dispatch_key: str = 'XPU'

ray_device_key class-attribute instance-attribute

ray_device_key: str = 'GPU'

check_and_update_config classmethod

check_and_update_config(vllm_config: VllmConfig) -> None
Source code in vllm/platforms/xpu.py
@classmethod
def check_and_update_config(cls, vllm_config: VllmConfig) -> None:
    cache_config = vllm_config.cache_config
    # in V1(or with ipex chunked prefill) block_size is 64
    if cache_config and cache_config.block_size is None:
        if envs.VLLM_USE_V1:
            cache_config.block_size = 64
        else:
            cache_config.block_size = 16

    # Instances created using VllmConfig() typically have model_config as
    # None by default. The modification involves adding a check to prevent
    # potential null exceptions check and update model config.
    if vllm_config.model_config is not None:
        model_config = vllm_config.model_config
        if model_config.dtype == torch.bfloat16:
            bf16_supported = cls.device_support_bf16()
            if not bf16_supported:
                model_config.dtype = torch.float16
        if not model_config.enforce_eager:
            logger.warning(
                "CUDA graph is not supported on XPU, fallback to the eager "
                "mode.")
            model_config.enforce_eager = True

    if vllm_config.speculative_config is not None:
        raise NotImplementedError(
            "XPU does not support speculative decoding")

    if vllm_config.device_config is not None:
        assert vllm_config.device_config.device_type == "xpu"

    # check and update parallel config
    parallel_config = vllm_config.parallel_config
    if envs.VLLM_USE_V1:
        parallel_config.worker_cls =\
            "vllm.v1.worker.xpu_worker.XPUWorker"
    else:
        parallel_config.worker_cls = "vllm.worker.xpu_worker.XPUWorker"

    if parallel_config.distributed_executor_backend is None:
        if parallel_config.world_size > 1:
            parallel_config.distributed_executor_backend = "ray"
        else:
            parallel_config.distributed_executor_backend = "uni"
    elif parallel_config.distributed_executor_backend == "mp":
        # FIXME(kunshang):
        # spawn needs calling `if __name__ == '__main__':``
        # fork is not supported for xpu start new process.
        if envs.VLLM_WORKER_MULTIPROC_METHOD != "spawn":
            os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
            logger.warning(
                "Please use spawn as start method if you want to use mp.")
    elif parallel_config.distributed_executor_backend != "ray" and \
            parallel_config.distributed_executor_backend != "uni":
        logger.warning(
            "%s is not supported on XPU, fallback to ray distributed"
            " executor backend.",
            parallel_config.distributed_executor_backend)
        parallel_config.distributed_executor_backend = "ray"

    if vllm_config.model_config and vllm_config.model_config.use_mla:
        logger.info(
            "MLA is enabled on a non-GPU platform; forcing chunked "
            "prefill and prefix caching to be disabled.")
        vllm_config.scheduler_config.enable_chunked_prefill = False
        vllm_config.scheduler_config.chunked_prefill_enabled = False
        vllm_config.scheduler_config.max_num_batched_tokens = max(
            vllm_config.scheduler_config.max_model_len,
            DEFAULT_MAX_NUM_BATCHED_TOKENS)

device_count classmethod

device_count() -> int
Source code in vllm/platforms/xpu.py
@classmethod
def device_count(cls) -> int:
    return torch.xpu.device_count()

device_support_bf16 classmethod

device_support_bf16() -> bool
Source code in vllm/platforms/xpu.py
@classmethod
def device_support_bf16(cls) -> bool:
    device_name = cls.get_device_name().lower()
    if cls.is_client_gpu_a770():
        logger.warning("Intel Arc A770 have bfloat16 accuracy known issue,"
                       " fallback to float16")
        return False
    else:
        logger.info(
            "Device name %s supports bfloat16. Please file an issue "
            "if you encounter any accuracy problems with bfloat16.",
            device_name)
        return True

get_attn_backend_cls classmethod

get_attn_backend_cls(
    selected_backend: _Backend,
    head_size: int,
    dtype: dtype,
    kv_cache_dtype: Optional[str],
    block_size: int,
    use_v1: bool,
    use_mla: bool,
) -> str
Source code in vllm/platforms/xpu.py
@classmethod
def get_attn_backend_cls(cls, selected_backend: _Backend, head_size: int,
                         dtype: torch.dtype, kv_cache_dtype: Optional[str],
                         block_size: int, use_v1: bool,
                         use_mla: bool) -> str:
    if selected_backend != _Backend.IPEX:
        logger.info("Cannot use %s backend on XPU.", selected_backend)
    use_v1 = envs.VLLM_USE_V1
    if use_v1:
        logger.info("Using Flash Attention backend on V1 engine.")
        return "vllm.v1.attention.backends.flash_attn.FlashAttentionBackend"
    else:
        logger.info("Using IPEX attention backend.")
        return "vllm.attention.backends.ipex_attn.IpexAttnBackend"

get_current_memory_usage classmethod

get_current_memory_usage(
    device: Optional[Device] = None,
) -> float
Source code in vllm/platforms/xpu.py
@classmethod
def get_current_memory_usage(cls,
                             device: Optional[torch.types.Device] = None
                             ) -> float:
    torch.xpu.reset_peak_memory_stats(device)
    return torch.xpu.max_memory_allocated(device)

get_device_capability classmethod

get_device_capability(
    device_id: int = 0,
) -> Optional[DeviceCapability]
Source code in vllm/platforms/xpu.py
@classmethod
def get_device_capability(
    cls,
    device_id: int = 0,
) -> Optional[DeviceCapability]:
    # capacity format differs from cuda's and will cause unexpected
    # failure, so use None directly
    return None

get_device_communicator_cls classmethod

get_device_communicator_cls() -> str
Source code in vllm/platforms/xpu.py
@classmethod
def get_device_communicator_cls(cls) -> str:
    return "vllm.distributed.device_communicators.xpu_communicator.XpuCommunicator"  # noqa

get_device_name classmethod

get_device_name(device_id: int = 0) -> str
Source code in vllm/platforms/xpu.py
@classmethod
def get_device_name(cls, device_id: int = 0) -> str:
    return torch.xpu.get_device_name(device_id)

get_device_total_memory classmethod

get_device_total_memory(device_id: int = 0) -> int
Source code in vllm/platforms/xpu.py
@classmethod
def get_device_total_memory(cls, device_id: int = 0) -> int:
    device_props = torch.xpu.get_device_properties(device_id)
    return device_props.total_memory

inference_mode classmethod

inference_mode()
Source code in vllm/platforms/xpu.py
@classmethod
def inference_mode(cls):
    return torch.no_grad()

is_async_output_supported classmethod

is_async_output_supported(
    enforce_eager: Optional[bool],
) -> bool
Source code in vllm/platforms/xpu.py
@classmethod
def is_async_output_supported(cls, enforce_eager: Optional[bool]) -> bool:
    return True

is_client_gpu_a770 classmethod

is_client_gpu_a770() -> bool
Source code in vllm/platforms/xpu.py
@classmethod
def is_client_gpu_a770(cls) -> bool:
    device_name = cls.get_device_name().lower()
    return device_name.count("a770") > 0

is_data_center_gpu classmethod

is_data_center_gpu() -> bool
Source code in vllm/platforms/xpu.py
@classmethod
def is_data_center_gpu(cls) -> bool:
    device_name = cls.get_device_name().lower()
    return device_name.count("data center gpu") > 0

is_pin_memory_available classmethod

is_pin_memory_available()
Source code in vllm/platforms/xpu.py
@classmethod
def is_pin_memory_available(cls):
    logger.warning("Pin memory is not supported on XPU.")
    return False

supports_v1 classmethod

supports_v1(model_config: ModelConfig) -> bool
Source code in vllm/platforms/xpu.py
@classmethod
def supports_v1(cls, model_config: ModelConfig) -> bool:
    return True