artyom-beilis
Failure of AND Fury
Built the library ... Pytorch 1.13.1
/mnt/d/srcs/pytorch_dlprim$ python3 mnist.py --device ocl:0
Traceback (most recent call last):
File "/mnt/d/srcs/pytorch_dlprim/mnist.py", line 164, in
...............
ldd /mnt/d/build_ninja/dlprim/libpt_ocl.so linux-vdso.so.1 (0x00007ffe880fb000) libc10.so => /usr/local/lib/libc10.so (0x00007f5a3c439000) libOpenCL.so.1 => /lib/x86_64-linux-gnu/libOpenCL.so.1 (0x00007f5a3c406000) libdlprim_core.so => /mnt/d/build_ninja/dlprim/dlprimitives/libdlprim_core.so (0x00007f5a3c348000) libtorch.so => /usr/local/lib/libtorch.so (0x00007f5a3c343000) libtorch_cpu.so => /usr/local/lib/libtorch_cpu.so (0x00007f5a330f4000) libsqlite3.so.0 => /lib/x86_64-linux-gnu/libsqlite3.so.0 (0x00007f5a32fa5000) libstdc++.so.6 => /lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f5a32d79000) libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f5a32c92000) libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f5a32c72000) libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f5a32a49000) /lib64/ld-linux-x86-64.so.2 (0x00007f5a3c5db000) libnuma.so.1 => /lib/x86_64-linux-gnu/libnuma.so.1 (0x00007f5a32a3c000) libopenblas.so.0 => /lib/x86_64-linux-gnu/libopenblas.so.0 (0x00007f5a305e8000) libmpi_cxx.so.40 => /lib/x86_64-linux-gnu/libmpi_cxx.so.40 (0x00007f5a305ce000) libmpi.so.40 => /lib/x86_64-linux-gnu/libmpi.so.40 (0x00007f5a30497000) libgfortran.so.5 => /lib/x86_64-linux-gnu/libgfortran.so.5 (0x00007f5a301bc000) libopen-pal.so.40 => /lib/x86_64-linux-gnu/libopen-pal.so.40 (0x00007f5a30109000) libopen-rte.so.40 => /lib/x86_64-linux-gnu/libopen-rte.so.40 (0x00007f5a3004c000) libhwloc.so.15 => /lib/x86_64-linux-gnu/libhwloc.so.15 (0x00007f5a2fff0000) libquadmath.so.0 => /lib/x86_64-linux-gnu/libquadmath.so.0 (0x00007f5a2ffa8000) libevent_core-2.1.so.7 => /lib/x86_64-linux-gnu/libevent_core-2.1.so.7 (0x00007f5a2ff73000) libevent_pthreads-2.1.so.7 => /lib/x86_64-linux-gnu/libevent_pthreads-2.1.so.7 (0x00007f5a2ff6e000) libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007f5a2ff52000) libudev.so.1 => /lib/x86_64-linux-gnu/libudev.so.1 (0x00007f5a2ff28000)
If I install the default modules from pip, then I get this error:
/mnt/d/srcs/pytorch_dlprim$ python3 mnist.py --device ocl:0A
Traceback (most recent call last):A
File "/mnt/d/srcs/pytorch_dlprim/mnist.py", line 164, in
Update: Uninstalling all previous versions and installing the whl from pytorch helped. But still does not answer why it fails with custom build?
Yes and in another comment I showed that pushing it to Intel GPU is fine (except that it is slow). I am still wondering what would be the "correct CMake" parameters to create a custom build of torch.
Я получаю похожую ошибку на этом тесте:
python mnist.py --device ocl:1 Using device: privateuseone:1 Accessing device #1:AMD Radeon R9 Fury Series (radeonsi, fiji, LLVM 17.0.6, DRM 3.57, 6.8.9-calculate) on rusticl LLVM ERROR: Cannot select: 0x7fb130370a30: f32 = and 0x7fb130370b10, Constant:i32<2147483647> 0x7fb130370b10: f32 = bitcast 0x7fb130370e90 0x7fb130370e90: i32,ch = CopyFromReg 0x5617ad66f4e0, Register:i32 %14 0x7fb13036a9d0: i32 = Register %14 0x7fb13036aab0: i32 = Constant<2147483647> In function: main Аварийный останов
Другой тест отрабатывает корректно.
python test.py Accessing device #1:AMD Radeon R9 Fury Series (radeonsi, fiji, LLVM 17.0.6, DRM 3.57, 6.8.9-calculate) on rusticl REF [[[ 0 0] [164 0] [ 0 255] [ 0 255]]
[[ 0 166] [ 0 25] [ 0 9] [ 0 124]]
[[ 0 0] [ 0 255] [255 197] [ 0 0]]] DEV [[[ 0 0] [164 0] [ 0 255] [ 0 255]]
[[ 0 166] [ 0 25] [ 0 9] [ 0 124]]
[[ 0 0] [ 0 255] [255 197] [ 0 0]]] 0.0
Я получаю похожую ошибку на этом тесте:
python mnist.py --device ocl:1 Using device: privateuseone:1
- Please give the output of clinfo
- run the command
OPENCL_DEBUG_MODE=1 python mnist.py --device ocl:1(i.e. set environment variable OPENCL_DEBUG_MODE=1) - Please run
python tests\test_ops.py --device ocl:1
on rusticl
From what I know rusticl is horribly buggy. It does not work on my rx 560 at all.
Please try installing AMD drivers (I hope rocm supports it) or altenatively MESA driver (I mean OpenCL driver/platform)
- Please give the output of clinfo
Number of platforms 2 Platform Name Intel(R) CPU Runtime for OpenCL(TM) Applications Platform Vendor Intel(R) Corporation Platform Version OpenCL 2.1 LINUX Platform Profile FULL_PROFILE Platform Extensions cl_khr_icd 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_depth_images cl_khr_3d_image_writes cl_intel_exec_by_local_thread cl_khr_spir cl_khr_fp64 cl_khr_image2d_from_buffer cl_intel_vec_len_hint Platform Extensions function suffix INTEL Platform Host timer resolution 1ns
Platform Name rusticl Platform Vendor Mesa/X.org Platform Version OpenCL 3.0 Platform Profile FULL_PROFILE Platform Extensions cl_khr_byte_addressable_store cl_khr_create_command_queue cl_khr_expect_assume cl_khr_extended_versioning cl_khr_icd cl_khr_il_program cl_khr_spirv_no_integer_wrap_decoration Platform Extensions with Version cl_khr_byte_addressable_store 0x400000 (1.0.0) cl_khr_create_command_queue 0x400000 (1.0.0) cl_khr_expect_assume 0x400000 (1.0.0) cl_khr_extended_versioning 0x400000 (1.0.0) cl_khr_icd 0x400000 (1.0.0) cl_khr_il_program 0x400000 (1.0.0) cl_khr_spirv_no_integer_wrap_decoration 0x400000 (1.0.0) Platform Numeric Version 0xc00000 (3.0.0) Platform Extensions function suffix MESA Platform Host timer resolution 1ns
Platform Name Intel(R) CPU Runtime for OpenCL(TM) Applications
Number of devices 1
Device Name AMD EPYC 7542 32-Core Processor
Device Vendor Intel(R) Corporation
Device Vendor ID 0x8086
Device Version OpenCL 2.1 (Build 0)
Driver Version 18.1.0.0920
Device OpenCL C Version OpenCL C 2.0
Device Type CPU
Device Profile FULL_PROFILE
Device Available Yes
Compiler Available Yes
Linker Available Yes
Max compute units 64
Max clock frequency 0MHz
Device Partition (core)
Max number of sub-devices 64
Supported partition types by counts, equally, by names (Intel)
Supported affinity domains (n/a)
Max work item dimensions 3
Max work item sizes 8192x8192x8192
Max work group size 8192
Preferred work group size multiple (kernel) 128
Max sub-groups per work group 1
Preferred / native vector sizes
char 1 / 32
short 1 / 16
int 1 / 8
long 1 / 4
half 0 / 0 (n/a)
float 1 / 8
double 1 / 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 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 (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 270105997312 (251.6GiB)
Error Correction support No
Max memory allocation 67526499328 (62.89GiB)
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 Yes
Atomics Yes
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 0 bytes
Max size for global variable 65536 (64KiB)
Preferred total size of global vars 65536 (64KiB)
Global Memory cache type Read/Write
Global Memory cache size 524288 (512KiB)
Global Memory cache line size 64 bytes
Image support Yes
Max number of samplers per kernel 480
Max size for 1D images from buffer 4220406208 pixels
Max 1D or 2D image array size 2048 images
Base address alignment for 2D image buffers 64 bytes
Pitch alignment for 2D image buffers 64 pixels
Max 2D image size 16384x16384 pixels
Max 3D image size 2048x2048x2048 pixels
Max number of read image args 480
Max number of write image args 480
Max number of read/write image args 480
Max number of pipe args 16
Max active pipe reservations 4095
Max pipe packet size 1024
Local memory type Global
Local memory size 32768 (32KiB)
Max number of constant args 480
Max constant buffer size 131072 (128KiB)
Max size of kernel argument 3840 (3.75KiB)
Queue properties (on host)
Out-of-order execution Yes
Profiling Yes
Local thread execution (Intel) Yes
Queue properties (on device)
Out-of-order execution Yes
Profiling Yes
Preferred size 4294967295 (4GiB)
Max size 4294967295 (4GiB)
Max queues on device 4294967295
Max events on device 4294967295
Prefer user sync for interop No
Profiling timer resolution 1ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels Yes
Sub-group independent forward progress No
IL version SPIR-V_1.0
SPIR versions 1.2
printf() buffer size 1048576 (1024KiB)
Built-in kernels (n/a)
Device Extensions cl_khr_icd 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_depth_images cl_khr_3d_image_writes cl_intel_exec_by_local_thread cl_khr_spir cl_khr_fp64 cl_khr_image2d_from_buffer cl_intel_vec_len_hint
Platform Name rusticl
Number of devices 1
Device Name AMD Radeon R9 Fury Series (radeonsi, fiji, LLVM 17.0.6, DRM 3.57, 6.8.9-calculate)
Device Vendor AMD
Device Vendor ID 0x1002
Device Version OpenCL 3.0
Device UUID 00000000-8100-0000-0000-000000000000
Driver UUID 414d442d-4d45-5341-2d44-525600000000
Valid Device LUID No
Device LUID 0000-000000000000
Device Node Mask 0
Device Numeric Version 0xc00000 (3.0.0)
Driver Version 24.0.4
Device OpenCL C Version OpenCL C 1.2
Device OpenCL C all versions OpenCL C 0xc00000 (3.0.0)
OpenCL C 0x402000 (1.2.0)
OpenCL C 0x401000 (1.1.0)
OpenCL C 0x400000 (1.0.0)
Device OpenCL C features __opencl_c_integer_dot_product_input_4x8bit_packed 0x800000 (2.0.0)
__opencl_c_integer_dot_product_input_4x8bit 0x800000 (2.0.0)
__opencl_c_int64 0x400000 (1.0.0)
__opencl_c_images 0x400000 (1.0.0)
__opencl_c_3d_image_writes 0x400000 (1.0.0)
__opencl_c_subgroups 0x400000 (1.0.0)
Latest comfornace test passed v0000-01-01-00
Device Type GPU
Device Profile EMBEDDED_PROFILE
Device Available Yes
Compiler Available Yes
Linker Available Yes
Max compute units 56
Max clock frequency 1000MHz
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 (device) 64
Preferred work group size multiple (kernel) 64
Max sub-groups per work group 16
Preferred / native vector sizes
char 1 / 1
short 1 / 1
int 1 / 1
long 1 / 1
half 0 / 0 (n/a)
float 1 / 1
double 0 / 0 (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 64, Little-Endian
Global memory size 4294967296 (4GiB)
Error Correction support No
Max memory allocation 1073741824 (1024MiB)
Unified memory for Host and Device No
Shared Virtual Memory (SVM) capabilities (core)
Coarse-grained buffer sharing No
Fine-grained buffer sharing No
Fine-grained system sharing No
Atomics No
Minimum alignment for any data type 128 bytes
Alignment of base address 4096 bits (512 bytes)
Preferred alignment for atomics
SVM 0 bytes
Global 0 bytes
Local 0 bytes
Atomic memory capabilities relaxed, work-group scope
Atomic fence capabilities relaxed, acquire/release, work-group scope
Max size for global variable 0
Preferred total size of global vars 0
Global Memory cache type None
Image support Yes
Max number of samplers per kernel 32
Max size for 1D images from buffer 268435455 pixels
Max 1D or 2D image array size 2048 images
Max 2D image size 16384x16384 pixels
Max 3D image size 2048x2048x2048 pixels
Max number of read image args 32
Max number of write image args 16
Max number of read/write image args 0
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 65536 (64KiB)
Max number of constant args 16
Max constant buffer size 67108864 (64MiB)
Generic address space support No
Max size of kernel argument 4096 (4KiB)
Queue properties (on host)
Out-of-order execution No
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
Prefer user sync for interop Yes
Profiling timer resolution 40ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels No
Non-uniform work-groups No
Work-group collective functions No
Sub-group independent forward progress No
IL version SPIR-V_1.0 SPIR-V_1.1 SPIR-V_1.2 SPIR-V_1.3 SPIR-V_1.4
ILs with version SPIR-V 0x400000 (1.0.0)
SPIR-V 0x401000 (1.1.0)
SPIR-V 0x402000 (1.2.0)
SPIR-V 0x403000 (1.3.0)
SPIR-V 0x404000 (1.4.0)
printf() buffer size 1048576 (1024KiB)
Built-in kernels (n/a)
Built-in kernels with version (n/a)
Device Extensions cl_khr_byte_addressable_store cl_khr_create_command_queue cl_khr_expect_assume cl_khr_extended_versioning cl_khr_icd cl_khr_il_program cl_khr_spirv_no_integer_wrap_decoration cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_integer_dot_product cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_gl_sharing cles_khr_int64 cl_khr_3d_image_writes cl_khr_pci_bus_info cl_khr_device_uuid cl_khr_subgroup_shuffle cl_khr_subgroup_shuffle_relative
Device Extensions with Version cl_khr_byte_addressable_store 0x400000 (1.0.0)
cl_khr_create_command_queue 0x400000 (1.0.0)
cl_khr_expect_assume 0x400000 (1.0.0)
cl_khr_extended_versioning 0x400000 (1.0.0)
cl_khr_icd 0x400000 (1.0.0)
cl_khr_il_program 0x400000 (1.0.0)
cl_khr_spirv_no_integer_wrap_decoration 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_integer_dot_product 0x800000 (2.0.0)
cl_khr_local_int32_base_atomics 0x400000 (1.0.0)
cl_khr_local_int32_extended_atomics 0x400000 (1.0.0)
cl_khr_gl_sharing 0x400000 (1.0.0)
cles_khr_int64 0x400000 (1.0.0)
cl_khr_3d_image_writes 0x400000 (1.0.0)
cl_khr_pci_bus_info 0x400000 (1.0.0)
cl_khr_device_uuid 0x400000 (1.0.0)
cl_khr_subgroup_shuffle 0x400000 (1.0.0)
cl_khr_subgroup_shuffle_relative 0x400000 (1.0.0)
NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) No platform
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) No platform
clCreateContext(NULL, ...) [default] No platform
clCreateContext(NULL, ...) [other] Success [INTEL]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) Success (1)
Platform Name Intel(R) CPU Runtime for OpenCL(TM) Applications
Device Name AMD EPYC 7542 32-Core Processor
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) Success (1)
Platform Name Intel(R) CPU Runtime for OpenCL(TM) Applications
Device Name AMD EPYC 7542 32-Core Processor
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 Intel(R) CPU Runtime for OpenCL(TM) Applications
Device Name AMD EPYC 7542 32-Core Processor
ICD loader properties ICD loader Name Khronos OpenCL ICD Loader ICD loader Vendor Khronos Group ICD loader Version 3.0.6 ICD loader Profile OpenCL 3.0
- run the command
OPENCL_DEBUG_MODE=1 python mnist.py --device ocl:1(i.e. set environment variable OPENCL_DEBUG_MODE=1)
Same output: Using device: privateuseone:1 Accessing device #1:AMD Radeon R9 Fury Series (radeonsi, fiji, LLVM 17.0.6, DRM 3.57, 6.8.9-calculate) on rusticl LLVM ERROR: Cannot select: 0x7f5278370a50: f32 = and 0x7f5278370b30, Constant:i32<2147483647> 0x7f5278370b30: f32 = bitcast 0x7f5278370eb0 0x7f5278370eb0: i32,ch = CopyFromReg 0x560ef0c1de00, Register:i32 %14 0x7f527836a9f0: i32 = Register %14 0x7f527836aad0: i32 = Constant<2147483647> In function: main Аварийный останов
- Please run
python tests\test_ops.py --device ocl:1
OPENCL_DEBUG_MODE=1 python tests/test_op.py --device ocl:1
Mean 1d Accessing device #1:AMD Radeon R9 Fury Series (radeonsi, fiji, LLVM 17.0.6, DRM 3.57, 6.8.9-calculate) on rusticl torch.Size([1, 3, 4]) torch.Size([1, 3, 4]) y 0.000000 x0 0.000000 Mean 2d torch.Size([2, 1, 1]) torch.Size([2, 1, 1]) y 0.000000 x0 0.000000 Mean 1d squeeze torch.Size([3, 4]) torch.Size([3, 4]) y 0.000000 x0 0.000000 Mean 2d squeeze torch.Size([3]) torch.Size([3]) y 0.000000 x0 0.000000 Mean all squeeze torch.Size([]) torch.Size([]) y 0.000000 x0 0.000000 Sum 1d torch.Size([1, 3, 4]) torch.Size([1, 3, 4]) y 0.000000 x0 0.000000 Sum 2d torch.Size([2, 1, 1]) torch.Size([2, 1, 1]) y 0.000001 x0 0.000000 Sum 1d squeeze torch.Size([3, 4]) torch.Size([3, 4]) y 0.000000 x0 0.000000 Sum 2d squeeze torch.Size([3]) torch.Size([3]) y 0.000000 x0 0.000000 LogSoftmax LLVM ERROR: Cannot select: 0x7f1f8c1fed90: f32 = and 0x7f1f8c1fee70, Constant:i32<2147483647> 0x7f1f8c1fee70: f32 = bitcast 0x7f1f8c1ff1f0 0x7f1f8c1ff1f0: i32,ch = CopyFromReg 0x560aa816c510, Register:i32 %14 0x7f1f8c1f8d30: i32 = Register %14 0x7f1f8c1f8e10: i32 = Constant<2147483647> In function: main Аварийный останов
From what I know rusticl is horribly buggy. It does not work on my rx 560 at all.
Касаемо AMD это не так. Я используя pytorch_dlprim и rusticl уже обучил много своих тестовых нейросеток.
Моя карта (AMD Fury) не поддерживается ROCm. Я попробую что-то придумать, чтобы с другой версий OpenCL потестировать. В прошлый раз (issue 72) я тестировал с несколькими версиями OpenCl, и были ошибки. Но, как я понял, ты уже внёс какие-то фиксы в pytorch_dlprim, поэтому стоит перетестировать?
ты уже внёс какие-то фиксы в pytorch_dlprim, поэтому стоит перетестировать?
For sure - and move to pytorch 2.4 since 1.13 support will likely to end soon. Note also import is simplified now all you need is to import pytorch_ocl - look at readme.
But not critical - the OCL code is the same for 1.13 and 2.4
Also please change OPENCL_DEBUG_MODE=1 to =2 I hoped it was an exception but I mistaken. I want to see exactly at what operator it fails. But anyway it looks like driver failure. Back than Mesa worked quite fine for me.
Regarding support of ROCM - does their opencl driver recognise it or not? It is good to ask their support since they should provide some solution for working OpenCL driver (either old one pal - that worked actually very well) or somehow enable rocm. It isn't request to support their entire infrastructure hip etc, it is barely OpenCL driver as provided in official specs.
OPENCL_DEBUG_MODE=2 python tests/test_op.py --device ocl:1
LogSoftmax
in: at::Tensor ptdlprim::empty_strided(at::IntArrayRef, at::IntArrayRef, c10::optionalc10::ScalarType, c10::optionalc10::Layout, c10::optionalc10::Device, c10::optional
Full output:
https://pastebin.com/hazfFBvz
OPENCL_DEBUG_MODE=2 python mnist.py --device ocl:1
in: at::Tensor ptdlprim::copy_from(const at::Tensor&, const at::Tensor&, bool)
in: at::Tensor ptdlprim::make_contiguous_as_target_type(const at::Tensor&, const at::Tensor&)
in: at::Tensor ptdlprim::convolution_overrideable(const at::Tensor&, const at::Tensor&, const c10::optionalat::Tensor&, c10::IntArrayRef, c10::IntArrayRef, c10::IntArrayRef, bool, c10::IntArrayRef, int64_t)
in: at::Tensor& ptdlprim::add_out(const at::Tensor&, const at::Tensor&, const c10::Scalar&, at::Tensor&)
in: at::Tensor ptdlprim::allocate_empty(at::IntArrayRef, c10::optionalc10::ScalarType, c10::optionalc10::Layout, c10::optionalc10::Device, c10::optional
Full output:
https://pastebin.com/0TGnp3wj
