How can I debug this failure on pocl CPU #158

picca · 2024-02-27T21:51:24Z

picca
Feb 27, 2024

Here a trace when using pyvkfft

autopkgtest [17:30:25]: test opencl: [-----------------------
Number of platforms                               1
  Platform Name                                   Portable Computing Language
  Platform Vendor                                 The pocl project
  Platform Version                                OpenCL 3.0 PoCL 5.0+debian  Linux, None+Asserts, RELOC, SPIR, LLVM 16.0.6, SLEEF, DISTRO, POCL_DEBUG
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_icd cl_pocl_content_size
  Platform Extensions with Version                cl_khr_icd                                                       0x400000 (1.0.0)
                                                  cl_pocl_content_size                                             0x400000 (1.0.0)
  Platform Numeric Version                        0xc00000 (3.0.0)
  Platform Extensions function suffix             POCL
  Platform Host timer resolution                  0ns

  Platform Name                                   Portable Computing Language
Number of devices                                 1
  Device Name                                     cpu-haswell-AMD EPYC 7B12
  Device Vendor                                   AuthenticAMD
  Device Vendor ID                                0x8086
  Device Version                                  OpenCL 3.0 PoCL HSTR: cpu-x86_64-pc-linux-gnu-haswell
  Device Numeric Version                          0xc00000 (3.0.0)
  Driver Version                                  5.0+debian
  Device OpenCL C Version                         OpenCL C 1.2 PoCL
  Device OpenCL C all versions                    OpenCL C                                                         0x400000 (1.0.0)
                                                  OpenCL C                                                         0x401000 (1.1.0)
                                                  OpenCL C                                                         0x402000 (1.2.0)
                                                  OpenCL C                                                         0xc00000 (3.0.0)
  Device OpenCL C features                        __opencl_c_3d_image_writes                                       0xc00000 (3.0.0)
                                                  __opencl_c_images                                                0xc00000 (3.0.0)
                                                  __opencl_c_atomic_order_acq_rel                                  0xc00000 (3.0.0)
                                                  __opencl_c_atomic_order_seq_cst                                  0xc00000 (3.0.0)
                                                  __opencl_c_atomic_scope_device                                   0xc00000 (3.0.0)
                                                  __opencl_c_program_scope_global_variables                        0xc00000 (3.0.0)
                                                  __opencl_c_generic_address_space                                 0xc00000 (3.0.0)
                                                  __opencl_c_subgroups                                             0xc00000 (3.0.0)
                                                  __opencl_c_atomic_scope_all_devices                              0xc00000 (3.0.0)
                                                  __opencl_c_read_write_images                                     0xc00000 (3.0.0)
                                                  __opencl_c_fp64                                                  0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp32_global_atomic_add                            0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp32_local_atomic_add                             0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp32_global_atomic_min_max                        0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp32_local_atomic_min_max                         0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp64_global_atomic_add                            0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp64_local_atomic_add                             0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp64_global_atomic_min_max                        0xc00000 (3.0.0)
                                                  __opencl_c_ext_fp64_local_atomic_min_max                         0xc00000 (3.0.0)
                                                  __opencl_c_int64                                                 0xc00000 (3.0.0)
  Latest conformance test passed                  v2022-04-19-01
  Device Type                                     CPU
  Device Profile                                  FULL_PROFILE
  Device Available                                Yes
  Compiler Available                              Yes
  Linker Available                                Yes
  Max compute units                               2
  Max clock frequency                             2249MHz
  Device Partition                                (core)
    Max number of sub-devices                     2
    Supported partition types                     equally, by counts
    Supported affinity domains                    (n/a)
  Max work item dimensions                        3
  Max work item sizes                             4096x4096x4096
  Max work group size                             4096
  Preferred work group size multiple (device)     8
  Preferred work group size multiple (kernel)     8
  Max sub-groups per work group                   128
  Sub-group sizes (Intel)                         1, 2, 4, 8, 16, 32, 64, 128, 256, 512
  Preferred / native vector sizes                 
    char                                                16 / 16      
    short                                               16 / 16      
    int                                                  8 / 8       
    long                                                 4 / 4       
    half                                                 0 / 0        (n/a)
    float                                                8 / 8       
    double                                               4 / 4        (cl_khr_fp64)
  Half-precision Floating-point support           (n/a)
  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         (cl_khr_fp64)
    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
  Address bits                                    64, Little-Endian
  Global memory size                              6195752960 (5.77GiB)
  Error Correction support                        No
  Max memory allocation                           2147483648 (2GiB)
  Unified memory for Host and Device              Yes
  Shared Virtual Memory (SVM) capabilities        (core)
    Coarse-grained buffer sharing                 Yes
    Fine-grained buffer sharing                   Yes
    Fine-grained system sharing                   No
    Atomics                                       Yes
  Unified Shared Memory (USM)                     (cl_intel_unified_shared_memory)
  Host USM capabilities (Intel)                   USM access, USM atomic access
  Device USM capabilities (Intel)                 USM access, USM atomic access
  Single-Device USM caps (Intel)                  USM access, USM atomic access
  Cross-Device USM caps (Intel)                   (n/a)
  Shared System USM caps (Intel)                  (n/a)
  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
  Atomic memory capabilities                      relaxed, acquire/release, sequentially-consistent, work-group scope, device scope, all-devices scope
  Atomic fence capabilities                       relaxed, acquire/release, sequentially-consistent, work-item scope, work-group scope, device scope
  Max size for global variable                    64000 (62.5KiB)
  Preferred total size of global vars             524288 (512KiB)
  Global Memory cache type                        Read/Write
  Global Memory cache size                        32768 (32KiB)
  Global Memory cache line size                   64 bytes
  Image support                                   Yes
    Max number of samplers per kernel             16
    Max size for 1D images from buffer            134217728 pixels
    Max 1D or 2D image array size                 2048 images
    Base address alignment for 2D image buffers   0 bytes
    Pitch alignment for 2D image buffers          0 pixels
    Max 2D image size                             8192x8192 pixels
    Max 3D image size                             2048x2048x2048 pixels
    Max number of read image args                 128
    Max number of write image args                128
    Max number of read/write image args           128
  Pipe support                                    No
  Max number of pipe args                         0
  Max active pipe reservations                    0
  Max pipe packet size                            0
  Local memory type                               Global
  Local memory size                               524288 (512KiB)
  Max number of constant args                     8
  Max constant buffer size                        524288 (512KiB)
  Generic address space support                   Yes
  Max size of kernel argument                     1024
  Queue properties (on host)                      
    Out-of-order execution                        Yes
    Profiling                                     Yes
  Device enqueue capabilities                     (n/a)
  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
  Command buffer capabilities                     kernel printf, simultaneous use, out of order
    Required queue properties for command buffer  
    Out-of-order execution                        No
    Profiling                                     No
  Prefer user sync for interop                    Yes
  Profiling timer resolution                      1ns
  Execution capabilities                          
    Run OpenCL kernels                            Yes
    Run native kernels                            Yes
    Non-uniform work-groups                       No
    Work-group collective functions               No
    Sub-group independent forward progress        Yes
    IL version                                    (n/a)
    ILs with version                              (n/a)
    SPIR versions                                 (n/a)
  printf() buffer size                            16777216 (16MiB)
  Built-in kernels                                pocl.add.i8;org.khronos.openvx.scale_image.nn.u8;org.khronos.openvx.scale_image.bl.u8;org.khronos.openvx.tensor_convert_depth.wrap.u8.f32
  Built-in kernels with version                   pocl.add.i8                                                      0x402000 (1.2.0)
                                                  org.khronos.openvx.scale_image.nn.u8                             0x402000 (1.2.0)
                                                  org.khronos.openvx.scale_image.bl.u8                             0x402000 (1.2.0)
                                                  org.khronos.openvx.tensor_convert_depth.wrap.u8.f32              0x402000 (1.2.0)
  Device Extensions                               cl_khr_byte_addressable_store 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_3d_image_writes cl_khr_command_buffer cl_pocl_pinned_buffers cl_khr_subgroups cl_intel_unified_shared_memory cl_khr_subgroup_ballot cl_khr_subgroup_shuffle cl_intel_subgroups cl_intel_required_subgroup_size cl_ext_float_atomics cl_khr_spir cl_khr_fp64 cl_khr_int64_base_atomics cl_khr_int64_extended_atomics
  Device Extensions with Version                  cl_khr_byte_addressable_store                                    0x400000 (1.0.0)
                                                  cl_khr_global_int32_base_atomics                                 0x400000 (1.0.0)
                                                  cl_khr_global_int32_extended_atomics                             0x400000 (1.0.0)
                                                  cl_khr_local_int32_base_atomics                                  0x400000 (1.0.0)
                                                  cl_khr_local_int32_extended_atomics                              0x400000 (1.0.0)
                                                  cl_khr_3d_image_writes                                           0x400000 (1.0.0)
                                                  cl_khr_command_buffer                                              0x9004 (0.9.4)
                                                  cl_pocl_pinned_buffers                                             0x1000 (0.1.0)
                                                  cl_khr_subgroups                                                 0x400000 (1.0.0)
                                                  cl_intel_unified_shared_memory                                   0x400000 (1.0.0)
                                                  cl_khr_subgroup_ballot                                           0x400000 (1.0.0)
                                                  cl_khr_subgroup_shuffle                                          0x400000 (1.0.0)
                                                  cl_intel_subgroups                                               0x400000 (1.0.0)
                                                  cl_intel_required_subgroup_size                                  0x400000 (1.0.0)
                                                  cl_ext_float_atomics                                             0x400000 (1.0.0)
                                                  cl_khr_spir                                                      0x801000 (2.1.0)
                                                  cl_khr_fp64                                                      0x400000 (1.0.0)
                                                  cl_khr_int64_base_atomics                                        0x400000 (1.0.0)
                                                  cl_khr_int64_extended_atomics                                    0x400000 (1.0.0)

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  Portable Computing Language
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   Success [POCL]
  clCreateContext(NULL, ...) [default]            Success [POCL]
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  Success (1)
    Platform Name                                 Portable Computing Language
    Device Name                                   cpu-haswell-AMD EPYC 7B12
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  Success (1)
    Platform Name                                 Portable Computing Language
    Device Name                                   cpu-haswell-AMD EPYC 7B12
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  No devices found in platform
  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                                 Portable Computing Language
    Device Name                                   cpu-haswell-AMD EPYC 7B12

ICD loader properties
  ICD loader Name                                 OpenCL ICD Loader
  ICD loader Vendor                               OCL Icd free software
  ICD loader Version                              2.3.2
  ICD loader Profile                              OpenCL 3.0
Testing with python3.12:

[2024-02-27 18:03:46.635599312] POCL: in fn pocl_write_file at line 228:
  *** ERROR ***  |     ERROR |  pocl_remove(/tmp/.cache/pocl/kcache/IN/BFOHOJOIPHMGKEKACCOKKKININBFDPLLGFPKK/program.bc) failed
[2024-02-27 18:03:46.636800812] POCL: in fn compile_and_link_program at line 851:
  *** ERROR ***  |     ERROR | CL_BUILD_PROGRAM_FAILURE Device cpu-haswell-AMD EPYC 7B12 failed to build the program
Device cpu-haswell-AMD EPYC 7B12 failed to build the program

__kernel __attribute__((reqd_work_group_size(30, 1, 1))) void VkFFT_main (__global float2* inputs, __global float2* outputs, __global float2* twiddleLUT) {
__local float2 sdata[90];

float2 temp_0;
temp_0.x = 0.0f;
temp_0.y = 0.0f;
float2 temp_1;
temp_1.x = 0.0f;
temp_1.y = 0.0f;
float2 temp_2;
temp_2.x = 0.0f;
temp_2.y = 0.0f;
float2 w;
w.x = 0.0f;
w.y = 0.0f;
float2 loc_0;
loc_0.x = 0.0f;
loc_0.y = 0.0f;
float2 loc_1;
loc_1.x = 0.0f;
loc_1.y = 0.0f;
float2 loc_2;
loc_2.x = 0.0f;
loc_2.y = 0.0f;
unsigned int tempInt;
tempInt = 0;
unsigned int tempInt2;
tempInt2 = 0;
unsigned int shiftX;
shiftX = 0;
unsigned int shiftY;
shiftY = 0;
unsigned int shiftZ;
shiftZ = 0;
unsigned int stageInvocationID;
stageInvocationID = 0;
unsigned int blockInvocationID;
blockInvocationID = 0;
unsigned int sdataID;
sdataID = 0;
unsigned int combinedID;
combinedID = 0;
unsigned int inoutID;
inoutID = 0;
unsigned int inoutID_x;
inoutID_x = 0;
unsigned int inoutID_y;
inoutID_y = 0;
unsigned int LUTId;
LUTId = 0;
shiftX = get_group_id(0);
shiftY = get_group_id(1);
shiftY = shiftY * 30;
shiftZ = 0;
tempInt = get_global_id(2);
inoutID_y = tempInt % 3;
inoutID_y = inoutID_y * 90;
shiftZ = shiftZ + inoutID_y;
shiftZ = shiftZ + 0;
inoutID = 3 * shiftY;
inoutID = inoutID + shiftZ;
inoutID = inoutID + get_local_id(0);
combinedID = get_local_id(0) + 0;
inoutID_x = combinedID % 3;
inoutID_y = combinedID / 3;
inoutID_y = inoutID_y + shiftY;
sdataID = combinedID % 3;
sdataID = sdataID * 30;
tempInt = combinedID / 3;
sdataID = sdataID + tempInt;
sdata[sdataID] = inputs[inoutID];
combinedID = get_local_id(0) + 30;
inoutID_x = combinedID % 3;
inoutID_y = combinedID / 3;
inoutID_y = inoutID_y + shiftY;
inoutID = inoutID + 30;
sdataID = combinedID % 3;
sdataID = sdataID * 30;
tempInt = combinedID / 3;
sdataID = sdataID + tempInt;
sdata[sdataID] = inputs[inoutID];
combinedID = get_local_id(0) + 60;
inoutID_x = combinedID % 3;
inoutID_y = combinedID / 3;
inoutID_y = inoutID_y + shiftY;
inoutID = inoutID + 30;
sdataID = combinedID % 3;
sdataID = sdataID * 30;
tempInt = combinedID / 3;
sdataID = sdataID + tempInt;
sdata[sdataID] = inputs[inoutID];
barrier(CLK_LOCAL_MEM_FENCE);

stageInvocationID = get_local_id(1) + 0;
stageInvocationID = stageInvocationID % 1;
LUTId = stageInvocationID + 0;
sdataID = get_local_id(1) + 0;
sdataID = sdataID * 30;
sdataID = sdataID + get_local_id(0);
temp_0 = sdata[sdataID];
sdataID = get_local_id(1) + 1;
sdataID = sdataID * 30;
sdataID = sdataID + get_local_id(0);
temp_1 = sdata[sdataID];
sdataID = get_local_id(1) + 2;
sdataID = sdataID * 30;
sdataID = sdataID + get_local_id(0);
temp_2 = sdata[sdataID];
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_2.x = temp_2.y * w.y;
loc_2.x = -loc_2.x;
loc_2.x = fma(temp_2.x, w.x, loc_2.x);
loc_2.y = temp_2.y * w.x;
loc_2.y = fma(temp_2.x, w.y, loc_2.y);
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_1.x = temp_1.y * w.y;
loc_1.x = -loc_1.x;
loc_1.x = fma(temp_1.x, w.x, loc_1.x);
loc_1.y = temp_1.y * w.x;
loc_1.y = fma(temp_1.x, w.y, loc_1.y);
temp_1.x = loc_1.x + loc_2.x;
temp_1.y = loc_1.y + loc_2.y;
temp_2.x = loc_1.x - loc_2.x;
temp_2.y = loc_1.y - loc_2.y;
loc_0.x = temp_0.x + temp_1.x;
loc_0.y = temp_0.y + temp_1.y;
loc_1.x = fma(temp_1.x, -5.00000000000000000e-01f, temp_0.x);
loc_1.y = fma(temp_1.y, -5.00000000000000000e-01f, temp_0.y);
loc_2.x = temp_2.x * -8.66025403784438647e-01f;
loc_2.y = temp_2.y * -8.66025403784438647e-01f;
temp_0 = loc_0;
loc_0.x = -loc_2.y;
loc_0.y = loc_2.x;
temp_1.x = loc_1.x + loc_0.x;
temp_1.y = loc_1.y + loc_0.y;
loc_0.x = loc_2.y;
loc_0.y = -loc_2.x;
temp_2.x = loc_1.x + loc_0.x;
temp_2.y = loc_1.y + loc_0.y;
barrier(CLK_LOCAL_MEM_FENCE);

stageInvocationID = get_local_id(1) + 0;
blockInvocationID = stageInvocationID;
stageInvocationID = stageInvocationID % 1;
blockInvocationID = blockInvocationID - stageInvocationID;
inoutID = blockInvocationID * 3;
inoutID = inoutID + stageInvocationID;
sdataID = inoutID + 0;
sdataID = sdataID * 30;
sdataID = sdataID + gERROR

pyopencl  R2C    (2,15,15,15) axes=        None ndim=   3  -rrr  0111    float32 lut=None inplace=0  norm=   1 C   FFT: n2=1.5e-07 ninf=1.6e-07 < 3.8e-06 (0.043) 1 iFFT: n2=1.9e-07 ninf=4.0e-07 < 3.8e-06 (0.106) 0 buf=    0   OK  
[2024-02-27 18:06:33.480870950] POCL: in fn pocl_write_file at line 228:
  *** ERROR ***  |     ERROR |  pocl_remove(/tmp/.cache/pocl/kcache/PO/DIDEMFILEPACHDBLCDJLAADDNHKOHODHMKPGJ/program.bc) failed
[2024-02-27 18:06:33.482013390] POCL: in fn compile_and_link_program at line 851:
  *** ERROR ***  |     ERROR | CL_BUILD_PROGRAM_FAILURE Device cpu-haswell-AMD EPYC 7B12 failed to build the program
Device cpu-haswell-AMD EPYC 7B12 failed to build the program

__kernel __attribute__((reqd_work_group_size(8, 1, 1))) void VkFFT_main (__global float2* inputs, __global float2* outputs, __global float2* twiddleLUT) {
__local float2 sdata[120];

float2 temp_0;
temp_0.x = 0.0f;
temp_0.y = 0.0f;
float2 temp_1;
temp_1.x = 0.0f;
temp_1.y = 0.0f;
float2 temp_2;
temp_2.x = 0.0f;
temp_2.y = 0.0f;
float2 temp_3;
temp_3.x = 0.0f;
temp_3.y = 0.0f;
float2 temp_4;
temp_4.x = 0.0f;
temp_4.y = 0.0f;
float2 temp_5;
temp_5.x = 0.0f;
temp_5.y = 0.0f;
float2 temp_6;
temp_6.x = 0.0f;
temp_6.y = 0.0f;
float2 temp_7;
temp_7.x = 0.0f;
temp_7.y = 0.0f;
float2 temp_8;
temp_8.x = 0.0f;
temp_8.y = 0.0f;
float2 temp_9;
temp_9.x = 0.0f;
temp_9.y = 0.0f;
float2 temp_10;
temp_10.x = 0.0f;
temp_10.y = 0.0f;
float2 temp_11;
temp_11.x = 0.0f;
temp_11.y = 0.0f;
float2 temp_12;
temp_12.x = 0.0f;
temp_12.y = 0.0f;
float2 temp_13;
temp_13.x = 0.0f;
temp_13.y = 0.0f;
float2 temp_14;
temp_14.x = 0.0f;
temp_14.y = 0.0f;
float2 w;
w.x = 0.0f;
w.y = 0.0f;
float2 loc_0;
loc_0.x = 0.0f;
loc_0.y = 0.0f;
float2 loc_1;
loc_1.x = 0.0f;
loc_1.y = 0.0f;
float2 loc_2;
loc_2.x = 0.0f;
loc_2.y = 0.0f;
float2 loc_3;
loc_3.x = 0.0f;
loc_3.y = 0.0f;
float2 loc_4;
loc_4.x = 0.0f;
loc_4.y = 0.0f;
unsigned int tempInt;
tempInt = 0;
unsigned int tempInt2;
tempInt2 = 0;
unsigned int shiftX;
shiftX = 0;
unsigned int shiftY;
shiftY = 0;
unsigned int shiftZ;
shiftZ = 0;
unsigned int stageInvocationID;
stageInvocationID = 0;
unsigned int blockInvocationID;
blockInvocationID = 0;
unsigned int sdataID;
sdataID = 0;
unsigned int combinedID;
combinedID = 0;
unsigned int inoutID;
inoutID = 0;
unsigned int inoutID_x;
inoutID_x = 0;
unsigned int inoutID_y;
inoutID_y = 0;
unsigned int disableThreads;
disableThreads = 1;
unsigned int LUTId;
LUTId = 0;
shiftX = get_global_id(0);
shiftZ = 0;
tempInt = get_global_id(2);
inoutID_y = tempInt % 2;
inoutID_y = inoutID_y * 8;
shiftZ = shiftZ + inoutID_y;
tempInt = tempInt / 2;
inoutID_y = tempInt % 15;
inoutID_y = inoutID_y * 240;
shiftZ = shiftZ + inoutID_y;
shiftZ = shiftZ + 0;
inoutID_x = shiftX % 8;
tempInt2 = shiftX / 8;
tempInt2 = tempInt2 % 1;
tempInt = shiftX / 8;
tempInt = tempInt * 15;
tempInt2 = tempInt2 + tempInt;
if (tempInt2 < 15) {
inoutID_y = get_local_id(1) * 1;
inoutID_y = inoutID_y + tempInt2;
inoutID = inoutID_x;
tempInt = inoutID_y * 16;
inoutID = inoutID + tempInt;
inoutID = inoutID + shiftZ;
temp_0 = inputs[inoutID];
inoutID = inoutID + 16;
temp_1 = inputs[inoutID];
inoutID = inoutID + 16;
temp_2 = inputs[inoutID];
inoutID = inoutID + 16;
temp_3 = inputs[inoutID];
inoutID = inoutID + 16;
temp_4 = inputs[inoutID];
inoutID = inoutID + 16;
temp_5 = inputs[inoutID];
inoutID = inoutID + 16;
temp_6 = inputs[inoutID];
inoutID = inoutID + 16;
temp_7 = inputs[inoutID];
inoutID = inoutID + 16;
temp_8 = inputs[inoutID];
inoutID = inoutID + 16;
temp_9 = inputs[inoutID];
inoutID = inoutID + 16;
temp_10 = inputs[inoutID];
inoutID = inoutID + 16;
temp_11 = inputs[inoutID];
inoutID = inoutID + 16;
temp_12 = inputs[inoutID];
inoutID = inoutID + 16;
temp_13 = inputs[inoutID];
inoutID = inoutID + 16;
temp_14 = inputs[inoutID];
}
stageInvocationID = get_local_id(1) + 0;
stageInvocationID = stageInvocationID % 1;
LUTId = stageInvocationID + 0;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_14.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_14.x, w.x, loc_0.x);
loc_0.y = temp_14.y * w.x;
temp_14.y = fma(temp_14.x, w.y, loc_0.y);
temp_14.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_13.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_13.x, w.x, loc_0.x);
loc_0.y = temp_13.y * w.x;
temp_13.y = fma(temp_13.x, w.y, loc_0.y);
temp_13.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_12.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_12.x, w.x, loc_0.x);
loc_0.y = temp_12.y * w.x;
temp_12.y = fma(temp_12.x, w.y, loc_0.y);
temp_12.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_11.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_11.x, w.x, loc_0.x);
loc_0.y = temp_11.y * w.x;
temp_11.y = fma(temp_11.x, w.y, loc_0.y);
temp_11.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_10.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_10.x, w.x, loc_0.x);
loc_0.y = temp_10.y * w.x;
temp_10.y = fma(temp_10.x, w.y, loc_0.y);
temp_10.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_9.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_9.x, w.x, loc_0.x);
loc_0.y = temp_9.y * w.x;
temp_9.y = fma(temp_9.x, w.y, loc_0.y);
temp_9.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_8.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_8.x, w.x, loc_0.x);
loc_0.y = temp_8.y * w.x;
temp_8.y = fma(temp_8.x, w.y, loc_0.y);
temp_8.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_7.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_7.x, w.x, loc_0.x);
loc_0.y = temp_7.y * w.x;
temp_7.y = fma(temp_7.x, w.y, loc_0.y);
temp_7.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_6.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_6.x, w.x, loc_0.x);
loc_0.y = temp_6.y * w.x;
temp_6.y = fma(temp_6.x, w.y, loc_0.y);
temp_6.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_5.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_5.x, w.x, loc_0.x);
loc_0.y = temp_5.y * w.x;
temp_5.y = fma(temp_5.x, w.y, loc_0.y);
temp_5.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_4.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_4.x, w.x, loc_0.x);
loc_0.y = temp_4.y * w.x;
temp_4.y = fma(temp_4.x, w.y, loc_0.y);
temp_4.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_3.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_3.x, w.x, loc_0.x);
loc_0.y = temp_3.y * w.x;
temp_3.y = fma(temp_3.x, w.y, loc_0.y);
temp_3.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_2.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_2.x, w.x, loc_0.x);
loc_0.y = temp_2.y * w.x;
temp_2.y = fma(temp_2.x, w.y, loc_0.y);
temp_2.x = loc_0.x;
w.x = 1.00000000000000000e+00f;
w.y = 0.00000000000000000e+00f;
loc_0.x = temp_1.y * w.y;
loc_0.x = -loc_0.x;
loc_0.x = fma(temp_1.x, w.x, loc_0.x);
loc_0.y = temp_1.y * w.x;
temp_1.y = fma(temp_1.x, w.y, loc_0.y);
temp_1.x = loc_0.x;
loc_0 = temp_0;
loc_1 = temp_3;
loc_2 = temp_6;
loc_3 = temp_9;
loc_4 = temp_12;
temp_3.x = loc_1.x + loc_4.x;
temp_3.y = loc_1.y + loc_4.y;
temp_6.x = loc_2.x + loc_3.x;
temp_6.y = loc_2.y + loc_3.y;
temp_9.x = loc_2.x - loc_3.x;
temp_9.y = loc_2.y - loc_3.y;
temp_12.x = loc_1.x - loc_4.x;
temp_12.y = loc_1.y - loc_4.y;
loc_3.x = temp_3.x - temp_6.x;
loc_3.y = temp_3.y - temp_6.y;
loc_4.x = temp_9.x + temp_12.x;
loc_4.y = temp_9.y + temp_12.y;
loc_0.x = temp_0.x + temp_3.x;
loc_0.y = temp_0.y + temp_3.y;
loc_0.x = loc_0.x + temp_6.x;
loc_0.y = loc_0.y + temp_6.y;
loc_1.x = fma(temp_3.x, -5.00000000000000000e-01f, temp_0.x);
loc_1.y = fma(temp_3.y, -5.00000000000000000e-01f, temp_0.y);
loc_2.x = fma(temp_6.x, -5.00000000000000000e-01f, temp_0.x);
loc_2.y = fma(temp_6.y, -5.00000000000000000e-01f, temp_0.y);
temp_9.x = temp_9.x * 1.53884176858762670e+00f;
temp_9.y = temp_9.y * 1.53884176858762670e+00f;
temp_12.x = temp_12.x * -3.63271264002680443e-01f;
temp_12.y = temp_12.y * -3.63271264002680443e-01f;
loc_3.x = loc_3.x * -8.09016994374947424e-01f;
loc_3.y = loc_3.y * -8.09016994374947424e-01f;
loc_4.x = loc_4.x * -5.87785252292473129e-01f;
loc_4.y = loc_4.y * -5.87785252292473129e-01f;
loc_1.x = loc_1.x - loc_3.x;
loc_1.y = loc_1.y - loc_3.y;
loc_2.x = loc_2.x + loc_3.x;
loc_2.y = loc_2.y + loc_3.y;
loc_3.x = temp_9.x + loc_4.x;
loc_3.y = temp_9.y + loc_4.y;
loc_4.x = loc_4.x + temp_12.x;
loc_4.y = loc_4.y + temp_12.y;
temp_0 = loc_0;
loc_0.x = -loc_4.y;
loc_0.y = loc_4.x;
temp_3.x = loc_1.x + loc_0.x;
temp_3.y = loc_1.y + loc_0.y;
loc_0.x = -loc_3.y;
loc_0.y = loc_3.x;
temp_6.x = loc_2.x + loc_0.x;
temp_6.y = loc_2.y + loc_0.y;
loc_0.x = loc_3.y;
loc_0.y = -loc_3.x;
temp_9.x = loc_2.x + loc_0.x;
temp_9.y = loc_2.y + loc_0.y;
loc_0.x = loc_4.y;
loc_0.y = -loc_4.x;
temp_12.x = loc_1.x + loc_0.x;
temp_12.y = loc_1.y + loc_0.y;
loc_0 = temp_1;
loc_1 = temp_4;
loc_2 = temp_7;
loc_3 = temp_10;
loc_4 = temp_13;
temp_4.x = loc_1.x + loc_4.x;
temp_4.y = loc_1.y + loc_4.y;
temp_7.x = loc_2.x + loc_3.x;
temp_7.y = loc_2.y + loc_3.y;
temp_10.x = loc_2.x - loc_3.x;
temp_10.y = loc_2.y - loc_3.y;
temp_13.x = loc_1.x - loc_4.x;
temp_13.y = loc_1.y - loc_4.y;
loc_3.x = temp_4.x - temp_7.x;
loc_3.y = temp_4.y - temp_7.y;
loc_4.x = temp_10.x + temp_13.x;
loc_4.y = temp_10.y + temp_13.y;
loc_0.x = temp_1.x + temp_4.x;
loc_0.y = temp_1.y + temp_4.y;
loc_0.x = loc_0.x + temp_7.x;
loc_0.y = loc_0.y + temp_7.y;
loc_1.x = fma(temp_4.x, -5.00000000000000000e-01f, temp_1.x);
loc_1.y = fma(temp_4.y, -5.00000000000000000e-01f, temp_1.y);
loc_2.x = fma(temp_7.x, -5.00000000000000000e-01f, temp_1.x);
loc_2.y = fma(temp_7.y, -5.00000000000000000e-01f, temp_1.y);
temp_10.x = temp_10.x * 1.53884176858762670e+00f;
temp_10.y = temp_10.y * 1.53884176858762670e+00f;
temp_13.x = temp_13.x * -3.63271264002680443e-01f;
temp_13.y = temp_13.y * -3.63271264002680443e-01f;
loc_3.x = loc_3.x * -8.09016994374947424e-01f;
loc_3.y = loc_3.y * -8.09016994374947424e-01f;
loc_4.x = loc_4.x * -5.87785252292473129e-01f;
loc_4.y = loc_4.y * -5.87785252292473129e-01f;
loc_1.x = loc_1.x - loc_3.x;
loc_1.y = loc_1.y - loc_3.y;
loc_2.x = loc_2.x + loc_3.x;
loc_2.y = loc_2.y + loc_3.y;
loc_3.x = temp_10.x + loc_4.x;
loc_3.y = temp_10.y + loc_4.y;
loc_4.x = loc_4.x + temp_13.x;
loc_4.y = loc_4.y + temp_13.y;
temp_1 = loc_0;
loc_0.x = -loc_4.y;
loc_0.y = loc_4.x;
temp_4.x = loc_1.x + loc_0.x;
temp_4.y = loc_1.y + loc_0.y;
loc_0.x = -loc_3.y;
loc_0.y = loc_3.x;
temp_7.x = loc_2.x + loc_0.x;
temp_7.y = loc_2.y + loc_0.y;
loc_0.x = loc_3.y;
loc_0.y = -loc_3.x;
temp_10.x = loc_2.x + loc_0.x;
temp_10.y = loc_2.y + loc_0.y;
loc_0.x = loc_4.y;
loc_0.y = -loc_4.x;
temp_13.x = loc_1.x + loc_0.x;
temp_13.y = loc_1.y + loc_0.y;
loc_0 = temp_2;
loc_1 = temp_5;
loc_2 = temp_8;
loc_3 = temp_11;
loc_4 = temp_14;
temp_5.x = loc_1.x + loc_4.x;
temp_5.y = loc_1.y + loc_4.y;
temp_8.x = loc_2.x + loc_3.x;
temp_8.y = loc_2.y + loc_3.y;
temp_11.x = loc_2.x - loc_3.x;
temp_11.y = loc_2.y - loc_3.y;
temp_14.x = loc_1.x - loc_4.x;
temp_14.y = loc_1.y - loc_4.y;
loc_3.x = temp_5.x - temp_8.x;
loc_3.y = temp_5.y - temp_8.y;
loc_4.x = temp_11.x + temp_14.x;
loc_4.y = temp_11.y + temp_14.y;
loc_0.x = temp_2.x + temp_5.x;
loc_0.y = temp_2.y + temp_5.y;
loc_0.x = loc_0.x + temp_8.x;
loc_0.y = loc_0.y + temp_8.y;
loc_1.x = fma(temp_5.x, -5.00000000000000000e-01f, temp_2.x);
loc_1.y = fma(temp_5.y, -5.00000000000000000e-01f, temp_2.y);
loc_2.x = fma(temp_8.x, -5.00000000000000000e-01f, temp_2.x);
loc_2.y = fma(temp_8.y, -5.00000000000000000e-01f, temp_2.y);
temp_11.x = temp_11.x * 1.53884176858762670e+00f;
temp_11.y = temp_11.y * 1.53884176858762670e+00f;
temp_14.x = temp_14.x * -3.63271264002680443e-01f;
temp_14.y = temp_14.y * -3.63271264002680443e-01f;
loc_3.x = loc_3.x * -8.09016994374947424e-01f;
loc_3.y = loc_3.y * -8.09016994374947424e-01f;
loc_4.x = loc_4.x * -5.87785252292473129e-01f;
loc_4.y = loc_4.y * -5.87785252292473129e-01f;
loc_1.x = loc_1.x - loc_3.x;
loc_1.y = loc_1.y - loc_3.y;
loc_2.x = loc_2.x + loc_3.x;
loc_2.y = loc_2.y + loc_3.y;
loc_3.x = temp_11.x + loc_4.x;
loc_3.y = temp_11.y + loc_4.y;
loc_4.x = loc_4.x + temp_14.x;
loc_4.y = loc_4.y + temp_14.y;
temp_2 = loc_0;
loc_0.x = -loc_4.y;
loc_0.y = loc_4.x;
temp_5.x = loc_1.x + loc_0.x;
temp_5.y = loc_1.y + loc_0.y;
loc_0.x = -loc_3.y;
loc_0.y = loc_3.x;
temp_8.x = loc_2.x + loc_0.x;
temp_8.y = loc_2.y + loc_0.y;
loc_0.x = loc_3.y;
loc_0.y = -loc_3.x;
temp_11.x = loc_2.x + loc_0.x;
temp_11.y = loc_2.y + loc_0.y;
loc_0.x = loc_4.y;
loc_0.y = -loc_4.x;
temp_14.x = loc_1.x + loc_0.x;
temp_14.y = loc_1.y + loc_0.y;
loc_1 = temp_1;
loc_2 = temp_2;
temp_1.x = loc_1.x + loc_2.x;
temp_1.y = loc_1.y + loc_2.y;
temp_2.x = loc_1.x - loc_2.x;
temp_2.y = loc_1.y - loc_2.y;
loc_0.x = temp_0.x + temp_1.x;
loc_0.y = temp_0.y + temp_1.y;
loc_1.x = fma(temp_1.x, -5.00000000000000000e-01f, temp_0.x);
loc_1.y = fma(temp_1.y, -5.00000000000000000e-01f, temp_0.y);
loc_2.x = temp_2.x * -8.66025403784438647e-01f;
loc_2.y = temp_2.y * -8.66025403784438647e-01f;
temp_0 = loc_0;
loc_0.x = -loc_2.y;
loc_0.y = loc_2.x;
temp_1.x = loc_1.x + loc_0.x;
temp_1.y = loc_1.y + loc_0.y;
loc_0.x = loc_2.y;
loc_0.y = -loc_2.x;
temp_2.x = loc_1.x + loc_0.x;
temp_2.y = loc_1.y + loc_0.y;
w.x = 9.13545457642600895e-01f;
w.y = -4.06736643075800208e-01f;
loc_1.x = temp_4.y * w.y;
loc_1.x = -loc_1.x;
loc_1.x = fma(temp_4.x, w.x, loc_1.x);
loc_1.y = temp_4.y * w.x;
loc_1.y = fma(temp_4.x, w.y, loc_1.y);
w.x = 6.69130606358858214e-01f;
w.y = -7.43144825477394235e-01f;
loc_2.x = temp_5.y * w.y;
loc_2.x = -loc_2.x;
loc_2.x = fma(temp_5.x, w.x, loc_2.x);
loc_2.y = temp_5.y * w.x;
loc_2.y = fma(temp_5.x, w.y, loc_2.y);
temp_4.x = loc_1.x + loc_2.x;
temp_4.y = loc_1.y + loc_2.y;
temp_5.x = loc_1.x - loc_2.x;
temp_5.y = loc_1.y - loc_2.y;
loc_0.x = temp_3.x + temp_4.x;
loc_0.y = temp_3.y + temp_4.y;
loc_1.x = fma(temp_4.x, -5.00000000000000000e-01f, temp_3.x);
loc_1.y = fma(temp_4.y, -5.00000000000000000e-01f, temp_3.y);
loc_2.x = temp_5.x * -8.66025403784438647e-01f;
loc_2.y = temp_5.y * -8.66025403784438647e-01f;
temp_3 = loc_0;
loc_0.x = -loc_2.y;
loc_0.y = loc_2.x;
temp_4.x = loc_1.x + loc_0.x;
temp_4.y = loc_1.y + loc_0.y;
loc_0.x = loc_2.y;
loc_0.y = -loc_2.x;
temp_5.x = loc_1.x + loc_0.x;
temp_5.y = loc_1.y + loc_0.y;
w.x = 6.69130606358858214e-01f;
w.y = -7.43144825477394235e-01f;
loc_1.x = temp_7.y * w.y;
loc_1.x = -loc_1.x;
loc_1.x = fma(temp_7.x, w.x, loc_1.x);
loc_1.y = temp_7.y * w.x;
loc_1.y = fma(temp_7.x, w.y, loc_1.y);
w.x = -1.04528463267653471e-01f;
w.y = -9.94521895368273337e-01f;
loc_2.x = temp_8.y * w.y;
loc_2.x = -loc_2.x;
loc_2.x = fma(temp_8.x, w.x, loc_2.x);
loc_2.y = temp_8.y * w.x;
loc_2.y = fma(temp_8.x, w.y, loc_2.y);
temp_7.x = loc_1.x + loc_2.x;
temp_7.y = loc_1.y + loc_2.y;
temp_8.x = loc_1.x - loc_2.x;
temp_8.y = loc_1.y - loc_2.y;
loc_0.x = temp_6.x + temp_7.x;
loc_0.y = temp_6.y + temp_7.y;
loc_1.x = fma(temp_7.x, -5.00000000000000000e-01f, temp_6.x);
loc_1.y = fma(temp_7.y, -5.00000000000000000e-01f, temp_6.y);
loc_2.x = temp_8.x * -8.66025403784438647e-01f;
loc_2.y = temp_8.y * -8.66025403784438647e-01f;
temp_6 = loc_0;
loc_0.x = -loc_2.y;
loc_0.y = loc_2.x;
temp_7.x = loc_1.x + loc_0.x;
temp_7.y = loc_1.y + loc_0.y;
loc_0.x = loc_2.y;
loc_0.y = -loc_2.x;
temp_8.x = loc_1.x + loc_0.x;
temp_8.y = loc_1.y + loc_0.y;
w.x = 3.09016994374947424e-01f;
w.y = -9.51056516295153572e-01f;
loc_1.x = temp_10.y * w.y;
loc_1.x = -loc_1.x;
loc_1.x = fma(temp_10.x, w.x, loc_1.x);
loc_1.y = temp_10.y * w.x;
loc_1.y = fma(temp_10.x, w.y, loc_1.y);
w.x = -8.09016994374947424e-01f;
w.y = -5.87785252292473129e-01f;
loc_2.x = temp_11.y * w.y;
loc_2.x = -loc_2.x;
loc_2.x = fma(temp_11.x, w.x, loc_2.x);
loc_2.y = temp_11.y * w.x;
loc_2.y = fma(temp_11.x, w.y, loc_2.y);
temp_10.x = loc_1.x + loc_2.x;
temp_10.y = loc_1.y + loc_2.y;
temp_11.x = loc_1.x - loc_2.x;
temp_11.y = loc_1.y - loc_2.y;
loc_0.x = temp_9.x + temp_10.x;
loc_0.y = temp_9.y + temp_10.y;
loc_1.x = fma(temp_10.x, -5.00000000000000000e-01f, temp_9.x);
loc_1.y = fma(temp_10.y, -5.00000000000000000e-01f, temp_9.y);
loc_2.x = temp_11.x * -8.66025403784438647e-01f;
loc_2.y = temp_11.y * -8.66025403784438647e-01f;
temp_9 = loc_0;
loc_0.x = -loc_2.y;
loc_0.y = loc_2.x;
temp_10.x = loc_1.x + loc_0.x;
temp_10.y = loc_1.y + loc_0.y;
loc_0.x = loc_2.y;
loc_0.y = -loc_2.x;
temp_11.x = loc_1.x + loc_0.x;
temp_11.y = loc_1.y + loc_0.y;
w.x = -1.04528463267653471e-01f;
w.y = -9.94521895368273337e-01f;
loc_1.x = temp_13.y * w.y;
loc_1.x = -loc_1.x;
loc_1.x = fma(temp_13.x, w.x, loc_1.x);
loc_1.y = temp_13.y * w.x;
loc_1.y = fma(temp_13.x, w.y, loc_1.y);
w.x = -9.78147600733805638e-01f;
w.y = 2.07911690817759337e-01f;
loc_2.x = temp_14.y * w.y;
loc_2.x = -loc_2.x;
loc_2.x = fma(temp_14.x, w.x, loc_2.x);
loc_2.y = tERROR
test_simple_fft (pyvkfft.test.test_fft.TestFFT.test_simple_fft)
Test the simple fft API ... ok
test_zz_multi_gpu_fft_threaded_pyopencl (pyvkfft.test.test_fft.TestFFT.test_zz_multi_gpu_fft_threaded_pyopencl)
Test the pyvkfft.fft (cached) interface with multiple GPU and parallel threads with pyopencl ... skipped '0<2 devices available'
test_zz_multi_gpu_fft_threads_cupy (pyvkfft.test.test_fft.TestFFT.test_zz_multi_gpu_fft_threads_cupy)
Test the pyvkfft.fft (cached) interface with multiple GPU and parallel threads with cupy ... skipped 'cupy is not available'
test_zz_multi_gpu_fft_threads_pycuda (pyvkfft.test.test_fft.TestFFT.test_zz_multi_gpu_fft_threads_pycuda)
Test the pyvkfft.fft (cached) interface with multiple GPU and parallel threads with pycuda streams ... skipped 'pycuda is not available'

======================================================================
ERROR: test_c2c (pyvkfft.test.test_fft.TestFFT.test_c2c)
Run C2C tests
----------------------------------------------------------------------
multiprocessing.pool.RemoteTraceback: 
"""
Traceback (most recent call last):
  File "/usr/lib/python3.12/multiprocessing/pool.py", line 125, in worker
    result = (True, func(*args, **kwds))
                    ^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/pyvkfft/accuracy.py", line 596, in test_accuracy_kwargs
    t = test_accuracy(**kwargs)
        ^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/pyvkfft/accuracy.py", line 322, in test_accuracy
    app = clVkFFTApp(d0.shape, d0.dtype, queue, ndim=ndim, norm=norm,
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/pyvkfft/opencl.py", line 189, in __init__
    check_vkfft_result(res, shape, dtype, ndim, inplace, norm, r2c, dct, dst, axes, "opencl:%s:%s" %
  File "/usr/lib/python3/dist-packages/pyvkfft/base.py", line 446, in check_vkfft_result
    raise RuntimeError("VkFFT error %d: %s %s" % (res, r.name, s))
RuntimeError: VkFFT error 4031: VKFFT_ERROR_FAILED_TO_COMPILE_PROGRAM C2C (3,30,3) complex64 3D inplace norm=1 [opencl:Portable Computing Language:cpu-haswell-AMD EPYC 7B12]
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/pyvkfft/test/test_fft.py", line 508, in test_c2c
    self.run_fft_parallel(vkwargs)
  File "/usr/lib/python3/dist-packages/pyvkfft/test/test_fft.py", line 429, in run_fft_parallel
    for res in pool.imap(test_accuracy_kwargs, vkwargs):
  File "/usr/lib/python3.12/multiprocessing/pool.py", line 873, in next
    raise value
RuntimeError: VkFFT error 4031: VKFFT_ERROR_FAILED_TO_COMPILE_PROGRAM C2C (3,30,3) complex64 3D inplace norm=1 [opencl:Portable Computing Language:cpu-haswell-AMD EPYC 7B12]

======================================================================
ERROR: test_r2c (pyvkfft.test.test_fft.TestFFT.test_r2c)
Run R2C tests
----------------------------------------------------------------------
multiprocessing.pool.RemoteTraceback: 
"""
Traceback (most recent call last):
  File "/usr/lib/python3.12/multiprocessing/pool.py", line 125, in worker
    result = (True, func(*args, **kwds))
                    ^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/pyvkfft/accuracy.py", line 596, in test_accuracy_kwargs
    t = test_accuracy(**kwargs)
        ^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/pyvkfft/accuracy.py", line 322, in test_accuracy
    app = clVkFFTApp(d0.shape, d0.dtype, queue, ndim=ndim, norm=norm,
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3/dist-packages/pyvkfft/opencl.py", line 189, in __init__
    check_vkfft_result(res, shape, dtype, ndim, inplace, norm, r2c, dct, dst, axes, "opencl:%s:%s" %
  File "/usr/lib/python3/dist-packages/pyvkfft/base.py", line 446, in check_vkfft_result
    raise RuntimeError("VkFFT error %d: %s %s" % (res, r.name, s))
RuntimeError: VkFFT error 4031: VKFFT_ERROR_FAILED_TO_COMPILE_PROGRAM R2C (15,15,2,14+2) float32 [-4,-3,-1] inplace norm=1 [opencl:Portable Computing Language:cpu-haswell-AMD EPYC 7B12]
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/pyvkfft/test/test_fft.py", line 552, in test_r2c
    self.run_fft_parallel(vkwargs)
  File "/usr/lib/python3/dist-packages/pyvkfft/test/test_fft.py", line 429, in run_fft_parallel
    for res in pool.imap(test_accuracy_kwargs, vkwargs):
  File "/usr/lib/python3.12/multiprocessing/pool.py", line 873, in next
    raise value
RuntimeError: VkFFT error 4031: VKFFT_ERROR_FAILED_TO_COMPILE_PROGRAM R2C (15,15,2,14+2) float32 [-4,-3,-1] inplace norm=1 [opencl:Portable Computing Language:cpu-haswell-AMD EPYC 7B12]

----------------------------------------------------------------------
Ran 10 tests in 2179.732s

FAILED (errors=2, skipped=5)

picca · 2024-02-27T21:53:24Z

picca
Feb 27, 2024
Author

I am adding integration tests via autopkgtest in the Debian package of pyvkfft

I discussed with Vincent-Favre and he told that your are the only one understanding these traces :)

So my question is how can I debug this ???

thanks for your help

Frederic

1 reply

DTolm Feb 28, 2024
Maintainer

Hello,

Can you run one of the failing tests separately and see if it fails this way? Parts of the kernels printed in the log are corrupted and don't show what might have gone wrong.

The following message:

[2024-02-27 18:06:33.480870950] POCL: in fn pocl_write_file at line 228:
  *** ERROR ***  |     ERROR |  pocl_remove(/tmp/.cache/pocl/kcache/PO/DIDEMFILEPACHDBLCDJLAADDNHKOHODHMKPGJ/program.bc) failed

also raises some suspicion that the issue might be on pocl side. Testing the failing system only might help to see if the error is consistent.

Best regards,
Dmitrii

picca · 2024-02-29T10:26:38Z

picca
Feb 29, 2024
Author

Hello, here a full log from the Debian CI.

https://ci.debian.net/packages/p/pyvkfft/testing/amd64/43404320/

can you tell me if this log is usefull. I would like to produce usable log from our CI.

thanks

Frederic

1 reply

DTolm Feb 29, 2024
Maintainer

Hello,

This log is very useful (the only issue is that some parts of the kernels are cut off). However, the fact that the errors happen for different sizes now shows that there may be some issue with caching kernel binaries at pocl. We will probably need to ask someone from pocl what can make pocl_remove fail.

Best regards,
Dmitrii

picca · 2024-02-29T22:21:13Z

picca
Feb 29, 2024
Author

Do you know your entries in the pocl community ?

1 reply

DTolm Feb 29, 2024
Maintainer

I haven't really looked into CPU OpenCL execution - I once checked that it was possible but haven't optimized or done much for it. So I haven't talked to anyone in pocl yet as this is the first time I have seen the issue.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

How can I debug this failure on pocl CPU #158

Uh oh!

{{title}}

Uh oh!

Replies: 3 comments 3 replies

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Uh oh!

How can I debug this failure on pocl CPU #158

Uh oh!

picca Feb 27, 2024

Replies: 3 comments · 3 replies

Uh oh!

picca Feb 27, 2024 Author

Uh oh!

DTolm Feb 28, 2024 Maintainer

Uh oh!

picca Feb 29, 2024 Author

Uh oh!

DTolm Feb 29, 2024 Maintainer

Uh oh!

picca Feb 29, 2024 Author

Uh oh!

DTolm Feb 29, 2024 Maintainer

picca
Feb 27, 2024

Replies: 3 comments 3 replies

picca
Feb 27, 2024
Author

DTolm Feb 28, 2024
Maintainer

picca
Feb 29, 2024
Author

DTolm Feb 29, 2024
Maintainer

picca
Feb 29, 2024
Author

DTolm Feb 29, 2024
Maintainer