8
2

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 1 year has passed since last update.

Intel Arc A380でGPGPUをする

Posted at

はじめに

みなさん、Intel Arc A380買いました?
グローバルではA770, A750といった上位モデルがリリースされている昨今ですが、Intel Arc A380を(ようやく) GPGPUデバイスとして動かすことができたので、その備忘録です。

環境構築

2022/10/13現在、Intel ArcシリーズをLinuxでキレイにサポートしてくれている環境はないです。Linux kernel 6.1になるとフルサポートが入るとかなんとか。仕方ないので、イチから作りましょう。

…といって、某Phoronixを参考に色々作って試してみてたんですが、いまいちうまく行かなくてgithubで聞いたところ、

というドキュメントを教えてもらいました。
この手順に従えば一発で動きます。

動かしてみた結果

  • clinfo
Number of platforms                               1
  Platform Name                                   Intel(R) OpenCL HD Graphics
  Platform Vendor                                 Intel(R) Corporation
  Platform Version                                OpenCL 3.0
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_intel_command_queue_families cl_intel_subgroups cl_intel_required_subgroup_size cl_
intel_subgroups_short cl_khr_spir cl_intel_accelerator cl_intel_driver_diagnostics cl_khr_priority_hints cl_khr_throttle_hints cl_khr_create_command_queue cl_intel_subgroups_char cl_intel_subgroups_long cl_khr_il_program cl_intel_mem_force_host_memory cl_khr_subgroup_extended_types cl_khr_subgroup_non_uniform_vote cl_khr_subgroup_ballot cl_khr_subgroup_non_uniform_arithmetic cl_khr_subgroup_shuffle cl_khr_subgroup_shuffle_relative cl_khr_subgroup_clustered_reduce cl_intel_device_attribute_query cl_khr_suggested_local_work_size cl_intel_split_work_group_barrier cl_intel_spirv_media_block_io cl_intel_spirv_subgroups cl_khr_spirv_no_integer_wrap_decoration cl_intel_unified_shared_memory cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_intel_planar_yuv cl_intel_packed_yuv cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_3d_image_writes cl_intel_media_block_io cl_intel_bfloat16_conversions cl_intel_va_api_media_sharing cl_intel_sharing_format_query cl_khr_pci_bus_info cl_intel_create_buffer_with_properties cl_intel_dot_accumulate cl_intel_subgroup_local_block_io cl_intel_subgroup_matrix_multiply_accumulate cl_intel_subgroup_split_matrix_multiply_accumulate
  Platform Host timer resolution                  1ns
  Platform Extensions function suffix             INTEL

  Platform Name                                   Intel(R) OpenCL HD Graphics
Number of devices                                 1
  Device Name                                     Intel(R) Graphics [0x56a5]
  Device Vendor                                   Intel(R) Corporation
  Device Vendor ID                                0x8086
  Device Version                                  OpenCL 3.0 NEO
  Driver Version                                  22.24.23453
  Device OpenCL C Version                         OpenCL C 1.2
  Device Type                                     GPU
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               128
  Max clock frequency                             2450MHz
  Device Partition                                (core)
    Max number of sub-devices                     0
    Supported partition types                     None
    Supported affinity domains                    (n/a)
  Max work item dimensions                        3
  Max work item sizes                             1024x1024x1024
  Max work group size                             1024
  Preferred work group size multiple              64
  Max sub-groups per work group                   128
  Sub-group sizes (Intel)                         8, 16, 32
  Preferred / native vector sizes
    char                                                16 / 16
    short                                                8 / 8
    int                                                  4 / 4
    long                                                 1 / 1
    half                                                 8 / 8        (cl_khr_fp16)
    float                                                1 / 1
    double                                               1 / 1        (n/a)
  Half-precision Floating-point support           (cl_khr_fp16)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
  Single-precision Floating-point support         (core)
    Denormals                                     Yes
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 Yes
    Round to infinity                             Yes
    IEEE754-2008 fused multiply-add               Yes
    Support is emulated in software               No
    Correctly-rounded divide and sqrt operations  Yes
  Double-precision Floating-point support         (n/a)
  Address bits                                    64, Little-Endian
  Global memory size                              255012864 (243.2MiB)
  Error Correction support                        No
  Max memory allocation                           127506432 (121.6MiB)
  Unified memory for Host and Device              No
  Shared Virtual Memory (SVM) capabilities        (core)
    Coarse-grained buffer sharing                 Yes
    Fine-grained buffer sharing                   No
    Fine-grained system sharing                   No
    Atomics                                       No
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       1024 bits (128 bytes)
  Preferred alignment for atomics
    SVM                                           64 bytes
    Global                                        64 bytes
    Local                                         64 bytes
  Max size for global variable                    65536 (64KiB)
  Preferred total size of global vars             127506432 (121.6MiB)
  Global Memory cache type                        Read/Write
  Global Memory cache size                        0
  Global Memory cache line size                   64 bytes
  Image support                                   Yes
    Max number of samplers per kernel             16
    Max size for 1D images from buffer            7969152 pixels
    Max 1D or 2D image array size                 2048 images
    Base address alignment for 2D image buffers   4 bytes
    Pitch alignment for 2D image buffers          4 pixels
    Max 2D image size                             16384x16384 pixels
    Max planar YUV image size                     16384x16128 pixels
    Max 3D image size                             16384x16384x2048 pixels
    Max number of read image args                 128
    Max number of write image args                128
    Max number of read/write image args           128
  Max number of pipe args                         0
  Max active pipe reservations                    0
  Max pipe packet size                            0
  Local memory type                               Local
  Local memory size                               65536 (64KiB)
  Max number of constant args                     8
  Max constant buffer size                        127506432 (121.6MiB)
  Max size of kernel argument                     2048 (2KiB)
  Queue properties (on host)
    Out-of-order execution                        Yes
    Profiling                                     Yes
  Queue properties (on device)
    Out-of-order execution                        No
    Profiling                                     No
    Preferred size                                0
    Max size                                      0
  Max queues on device                            0
  Max events on device                            0
  Prefer user sync for interop                    Yes
  Profiling timer resolution                      52ns
  Execution capabilities
    Run OpenCL kernels                            Yes
    Run native kernels                            No
    Sub-group independent forward progress        No
    IL version                                    SPIR-V_1.2
    SPIR versions                                 1.2
  printf() buffer size                            4194304 (4MiB)
  Built-in kernels                                (n/a)
  Device Extensions                               cl_khr_byte_addressable_store cl_khr_fp16 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_icd cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_intel_command_queue_families cl_intel_subgroups cl_intel_required_subgroup_size cl_intel_subgroups_short cl_khr_spir cl_intel_accelerator cl_intel_driver_diagnostics cl_khr_priority_hints cl_khr_throttle_hints cl_khr_create_command_queue cl_intel_subgroups_char cl_intel_subgroups_long cl_khr_il_program cl_intel_mem_force_host_memory cl_khr_subgroup_extended_types cl_khr_subgroup_non_uniform_vote cl_khr_subgroup_ballot cl_khr_subgroup_non_uniform_arithmetic cl_khr_subgroup_shuffle cl_khr_subgroup_shuffle_relative cl_khr_subgroup_clustered_reduce cl_intel_device_attribute_query cl_khr_suggested_local_work_size cl_intel_split_work_group_barrier cl_intel_spirv_media_block_io cl_intel_spirv_subgroups cl_khr_spirv_no_integer_wrap_decoration cl_intel_unified_shared_memory cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_intel_planar_yuv cl_intel_packed_yuv cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_3d_image_writes cl_intel_media_block_io cl_intel_bfloat16_conversions cl_intel_va_api_media_sharing cl_intel_sharing_format_query cl_khr_pci_bus_info cl_intel_create_buffer_with_properties cl_intel_dot_accumulate cl_intel_subgroup_local_block_io cl_intel_subgroup_matrix_multiply_accumulate cl_intel_subgroup_split_matrix_multiply_accumulate

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  Intel(R) OpenCL HD Graphics
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   Success [INTEL]
  clCreateContext(NULL, ...) [default]            Success [INTEL]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  Success (1)
    Platform Name                                 Intel(R) OpenCL HD Graphics
    Device Name                                   Intel(R) Graphics [0x56a5]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  Success (1)
    Platform Name                                 Intel(R) OpenCL HD Graphics
    Device Name                                   Intel(R) Graphics [0x56a5]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL)  Success (1)
    Platform Name                                 Intel(R) OpenCL HD Graphics
    Device Name                                   Intel(R) Graphics [0x56a5]

ICD loader properties
  ICD loader Name                                 OpenCL ICD Loader
  ICD loader Vendor                               OCL Icd free software
  ICD loader Version                              2.2.11
  ICD loader Profile                              OpenCL 2.1
        NOTE:   your OpenCL library only supports OpenCL 2.1,
                but some installed platforms support OpenCL 3.0.
                Programs using 3.0 features may crash
                or behave unexpectedly

A380,というか、Xe−HPGってdoubleサポートしてないっぽいんですよね…残念。

  • clpeak

より、clpeakを動かしてみました。

$ ./clpeak

Platform: Intel(R) OpenCL HD Graphics
  Device: Intel(R) Graphics [0x56a5]
    Driver version  : 22.24.23453 (Linux x64)
    Compute units   : 128
    Clock frequency : 2450 MHz

    Global memory bandwidth (GBPS)
      float   : 109.08
      float2  : 111.76
      float4  : 115.32
      float8  : 115.60
      float16 : 117.49

    Single-precision compute (GFLOPS)
      float   : 3282.88
      float2  : 2806.91
      float4  : 2623.91
      float8  : 2597.03
      float16 : 2448.10

    Half-precision compute (GFLOPS)
      half   : 4921.48
      half2  : 4911.95
      half4  : 4909.31
      half8  : 4908.17
      half16 : 4853.99

    No double precision support! Skipped

    Integer compute (GIOPS)
      int   : 1183.67
      int2  : 1356.22
      int4  : 1205.06
      int8  : 1385.80
      int16 : 1167.13

    Integer compute Fast 24bit (GIOPS)
      int   : 1178.15
      int2  : 1351.02
      int4  : 1202.26
      int8  : 1377.92
      int16 : 1165.42

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 4.67
      enqueueReadBuffer               : 4.33
      enqueueWriteBuffer non-blocking : 5.02
      enqueueReadBuffer non-blocking  : 4.62
      enqueueMapBuffer(for read)      : 5.17
        memcpy from mapped ptr        : 10.10
      enqueueUnmap(after write)       : 5.11
        memcpy to mapped ptr          : 10.39

    Kernel launch latency : 11.83 us

まとめ

A380をGPGPUデバイスとして使うのは色々大変でした。
が、全部OSSでできるようにしているIntelはすごいと思います。まぁAMDもGPUOpenという試みでやってますし、確かAMDのが先駆者だった気もしますが。

doubleをサポートしていない点、それからなんかPCIeの帯域が低い点(これはなぜかx1であがっているからっぽい) が気になります。
まぁこの環境、Resizable BARに対応してないので、それのせいな気がしないでもないですけど…うーんなんでx1になるんだろ。謎。

PCIeについてはまた何か変わったらそのうち追記します。
これでA770がいつローンチしても大丈夫だな!!(?

8
2
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
8
2

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?