sowson/darknet

activation_kernels.cl build failed on Intel HD Graphics

chenxy1988 opened this issue · 7 comments

Hi,

Sorry to trouble you, recently I am learning opencl related knowledge and using your project as an example. My laptop is a legacy dell which has a Intel onchip GPU and a ATI HD 7730M external GPU (too old :p) . the OS using ubuntu 20.04.
Since AMD no longer support HD7730M in newer Linux, so I using Intel onchip GPU to study the opencl.

I compiled the project and try to run an example in following command:

./bindarknet detector test ./cfg/coco.data ./cfg/yolov3.cfg ./yolov3.weights data/dog.jpg -i 0 -thresh 0.25

But when compiling the cl code, it reported an error message as below:

Device IDs: 1
Device ID: 0
Device name: Intel(R) HD Graphics IvyBridge M GT2
Device vendor: Intel
Device opencl availability: OpenCL 1.2 beignet 1.3
Device opencl used: 1.3
Device double precision: YES
Device max group size: 512
Device address bits: 32
ASSERTION FAILED: Double precision not supported on this device (if this is a literal, use '1.0f' not '1.0')
at file ./backend/src/backend/gen_insn_selection.cpp, function void gbe::ConvertInstructionPattern::convertBetweenFloatDouble(gbe::Selection::Opaque&, const gbe::ir::ConvertInstruction&, bool&) const, line 6227

Seems current GPU not support double float, I added an option "-cl-single-precision-constant " in src/opencl.c opencl_load_buffer() -> clBuildProgram , the diff as below:

diff --git a/src/opencl.c b/src/opencl.c
index 07c6550..9b36c35 100644
--- a/src/opencl.c
+++ b/src/opencl.c
@@ -255,6 +255,7 @@ void opencl_load_buffer(const char *buffer, const size_t size, cl_program *outpu
           //"-cl-fp32-correctly-rounded-divide-sqrt "
             "-cl-no-signed-zeros "
             "-cl-mad-enable "
+           "-cl-single-precision-constant "
           //"-cl-fast-relaxed-math "
             , NULL, NULL);
 #endif

re-compile and try again, error msg became as below:

evice address bits: 32
FATAL ERROR: CL_UNKNOWN_ERROR
opencl_load: could not compile. error: CL_UNKNOWN_ERROR
code:
In file included from :1:
/usr/lib/x86_64-linux-gnu/beignet//include/ocl.h:124:9: error: 'cl_intel_subgroups' macro redefined
:356:9: note: previous definition is here
In file included from :1:
/usr/lib/x86_64-linux-gnu/beignet//include/ocl.h:125:9: error: 'cl_intel_subgroups_short' macro redefined
:357:9: note: previous definition is here

I am no idea now, could you help to check whether my laptop environment can use the opencl with Intel GPU or not?
Many thanks ;)

Alternative information:

display information by lshw:


  *-display                 
       description: VGA compatible controller
       product: Chelsea LP [Radeon HD 7730M]
       vendor: Advanced Micro Devices, Inc. [AMD/ATI]
       physical id: 0
       bus info: pci@0000:01:00.0
       version: 00
       width: 64 bits
       clock: 33MHz
       capabilities: pm pciexpress msi vga_controller bus_master cap_list rom
       configuration: driver=radeon latency=0
       resources: irq:29 memory:a0000000-afffffff memory:c0000000-c003ffff ioport:3000(size=256) memory:c0040000-c005ffff
  *-display
       description: VGA compatible controller
       product: 3rd Gen Core processor Graphics Controller
       vendor: Intel Corporation
       physical id: 2
       bus info: pci@0000:00:02.0
       version: 09
       width: 64 bits
       clock: 33MHz
       capabilities: msi pm vga_controller bus_master cap_list
       configuration: driver=i915 latency=0
       resources: irq:28 memory:c1000000-c13fffff memory:b0000000-bfffffff ioport:4000(size=64) memory:c0000-dffff

 

display information by clinfo:


chenxy@chenxy-inspiron7520:/local$ clinfo
Number of platforms                               1
  Platform Name                                   Intel Gen OCL Driver
  Platform Vendor                                 Intel
  Platform Version                                OpenCL 2.0 beignet 1.3
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_3d_image_writes cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_spir cl_khr_icd cl_intel_accelerator cl_intel_subgroups cl_intel_subgroups_short
  Platform Extensions function suffix             Intel

  Platform Name                                   Intel Gen OCL Driver
Number of devices                                 1
  Device Name                                     Intel(R) HD Graphics IvyBridge M GT2
  Device Vendor                                   Intel
  Device Vendor ID                                0x8086
  Device Version                                  OpenCL 1.2 beignet 1.3
  Driver Version                                  1.3
  Device OpenCL C Version                         OpenCL C 1.2 beignet 1.3
  Device Type                                     GPU
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               16
  Max clock frequency                             1000MHz
  Device Partition                                (core)
    Max number of sub-devices                     1
    Supported partition types                     None, None, None
    Supported affinity domains                    (n/a)
  Max work item dimensions                        3
  Max work item sizes                             512x512x512
  Max work group size                             512
  Preferred work group size multiple (kernel)     16
  Preferred / native vector sizes                 
    char                                                16 / 8       
    short                                                8 / 8       
    int                                                  4 / 4       
    long                                                 2 / 2       
    half                                                 0 / 8        (n/a)
    float                                                4 / 4       
    double                                               0 / 2        (n/a)
  Half-precision Floating-point support           (n/a)
  Single-precision Floating-point support         (core)
    Denormals                                     No
    Infinity and NANs                             Yes
    Round to nearest                              Yes
    Round to zero                                 No
    Round to infinity                             No
    IEEE754-2008 fused multiply-add               No
    Support is emulated in software               No
    Correctly-rounded divide and sqrt operations  No
  Double-precision Floating-point support         (n/a)
  Address bits                                    32, Little-Endian
  Global memory size                              2147483648 (2GiB)
  Error Correction support                        No
  Max memory allocation                           1610612736 (1.5GiB)
  Unified memory for Host and Device              Yes
  Minimum alignment for any data type             128 bytes
  Alignment of base address                       1024 bits (128 bytes)
  Global Memory cache type                        Read/Write
  Global Memory cache size                        8192 (8KiB)
  Global Memory cache line size                   64 bytes
  Image support                                   Yes
    Max number of samplers per kernel             16
    Max size for 1D images from buffer            65536 pixels
    Max 1D or 2D image array size                 2048 images
    Base address alignment for 2D image buffers   4096 bytes
    Pitch alignment for 2D image buffers          1 pixels
    Max 2D image size                             8192x8192 pixels
    Max 3D image size                             8192x8192x2048 pixels
    Max number of read image args                 128
    Max number of write image args                8
  Local memory type                               Local
  Local memory size                               65536 (64KiB)
  Max number of constant args                     8
  Max constant buffer size                        134217728 (128MiB)
  Max size of kernel argument                     1024
  Queue properties                                
    Out-of-order execution                        No
    Profiling                                     Yes
  Prefer user sync for interop                    Yes
  Profiling timer resolution                      80ns
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            Yes
    SPIR versions                                 1.2
  printf() buffer size                            1048576 (1024KiB)
  Built-in kernels                                __cl_copy_region_align4;__cl_copy_region_align16;__cl_cpy_region_unalign_same_offset;__cl_copy_region_unalign_dst_offset;__cl_copy_region_unalign_src_offset;__cl_copy_buffer_rect;__cl_copy_image_1d_to_1d;__cl_copy_image_2d_to_2d;__cl_copy_image_3d_to_2d;__cl_copy_image_2d_to_3d;__cl_copy_image_3d_to_3d;__cl_copy_image_2d_to_buffer;__cl_copy_image_3d_to_buffer;__cl_copy_buffer_to_image_2d;__cl_copy_buffer_to_image_3d;__cl_fill_region_unalign;__cl_fill_region_align2;__cl_fill_region_align4;__cl_fill_region_align8_2;__cl_fill_region_align8_4;__cl_fill_region_align8_8;__cl_fill_region_align8_16;__cl_fill_region_align128;__cl_fill_image_1d;__cl_fill_image_1d_array;__cl_fill_image_2d;__cl_fill_image_2d_array;__cl_fill_image_3d;block_motion_estimate_intel;
  Device Extensions                               cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_3d_image_writes cl_khr_image2d_from_buffer cl_khr_depth_images cl_khr_spir cl_khr_icd cl_intel_accelerator cl_intel_subgroups cl_intel_subgroups_short cl_intel_motion_estimation

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  Intel Gen OCL Driver
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   Success [Intel]
  clCreateContext(NULL, ...) [default]            Success [Intel]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  Success (1)
    Platform Name                                 Intel Gen OCL Driver
    Device Name                                   Intel(R) HD Graphics IvyBridge M GT2
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  Success (1)
    Platform Name                                 Intel Gen OCL Driver
    Device Name                                   Intel(R) HD Graphics IvyBridge M GT2
  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 Gen OCL Driver
    Device Name                                   Intel(R) HD Graphics IvyBridge M GT2

ICD loader properties
  ICD loader Name                                 OpenCL ICD Loader
  ICD loader Vendor                               OCL Icd free software
  ICD loader Version                              2.2.14
  ICD loader Profile                              OpenCL 3.0

Thanks

I read the former issues which were closed in this project. I saw an issue related Intel GPU and should set NVIDIA=0 and GPU_MULTI=0 in Makefile.
I will try this tonight when I back home and keep post the testing result.

By the way, May I trun on "list (APPEND DARKNET_DEFINITIONS GPU CL_TARGET_OPENCL_VERSION=$MY_ACTUAL_OPENCL_VERSION" in CMakeLists.txt ? The warning message always output during compiling as below:
cl_version.h: CL_TARGET_OPENCL_VERSION is not defined. Defaulting to 220 (OpenCL 2.2)

Thanks ;p

Update:
I tried GPU_MULTI=0 NVIDIA=0 in make file, result the same as before.

I found some cules to root cause:

  1. I should not install beignet-ocl package due to my laptop don't have that hardware.
  2. After removing beignet, clinfo show me no platform, but I have already installed intel-ocl packages.
  3. The darknet-opencl also reponse me that no platform.
  4. I used strace to track clinfo, found that it will accord to /etc/OpenCL/vendor/intel.* to find available devices in /dev/dri/pci-*/ folder.
  5. But, the important things that my laptop has 2 graphi card, default 0 is ATI radeon 7xxxM, not match the /dev/OpenCL/vendor rule, so clinfo ( it called cl API ) return with error.

I will try to remove ATI radeon driver, use Intel as default graphi card and try it later.

close this ticket ;)

I read your paper "Darknet on OpenCL:a multi-platform tool for object detection and classification".

So cool :)

I will port the code to ARM A53 based CPU with MALI G31 GPU, if successed, i will share a performance test to you. Might be next year due to learn openCL happens in my part time and "touching fish" time during working.