intel / intel-extension-for-pytorch

A Python package for extending the official PyTorch that can easily obtain performance on Intel platform
Apache License 2.0
1.56k stars 237 forks source link

"import matplotlib" will make is_xpu_available() return False #463

Open notsyncing opened 10 months ago

notsyncing commented 10 months ago

Describe the bug

Hello, I found that a simple import matplotlib or import gradio will make is_xpu_available() returns False:

(/opt/text-generation-webui/installer_files/env) pods@c636bef83b1f:~$ source /opt/intel/oneapi/setvars.sh 

:: initializing oneAPI environment ...
   bash: BASH_VERSION = 5.1.16(1)-release
   args: Using "$@" for setvars.sh arguments: 
:: advisor -- latest
:: ccl -- latest
:: compiler -- latest
:: dal -- latest
:: debugger -- latest
:: dev-utilities -- latest
:: dnnl -- latest
:: dpcpp-ct -- latest
:: dpl -- latest
:: ipp -- latest
:: ippcp -- latest
:: mkl -- latest
:: mpi -- latest
:: tbb -- latest
:: vtune -- latest
:: oneAPI environment initialized ::

(/opt/text-generation-webui/installer_files/env) pods@c636bef83b1f:~$ python -c "import accelerate.utils; print(accelerate.utils.is_xpu_available())"
/opt/text-generation-webui/installer_files/env/lib/python3.11/site-packages/torchvision/io/image.py:13: UserWarning: Failed to load image Python extension: ''If you don't plan on using image functionality from `torchvision.io`, you can ignore this warning. Otherwise, there might be something wrong with your environment. Did you have `libjpeg` or `libpng` installed before building `torchvision` from source?
  warn(
True
(/opt/text-generation-webui/installer_files/env) pods@c636bef83b1f:~$ python -c "import matplotlib; import accelerate.utils; print(accelerate.utils.is_xpu_available())"
/opt/text-generation-webui/installer_files/env/lib/python3.11/site-packages/torchvision/io/image.py:13: UserWarning: Failed to load image Python extension: ''If you don't plan on using image functionality from `torchvision.io`, you can ignore this warning. Otherwise, there might be something wrong with your environment. Did you have `libjpeg` or `libpng` installed before building `torchvision` from source?
  warn(
False
(/opt/text-generation-webui/installer_files/env) pods@c636bef83b1f:~$ python -c "import gradio; import accelerate.utils; print(accelerate.utils.is_xpu_available())"
/opt/text-generation-webui/installer_files/env/lib/python3.11/site-packages/torchvision/io/image.py:13: UserWarning: Failed to load image Python extension: ''If you don't plan on using image functionality from `torchvision.io`, you can ignore this warning. Otherwise, there might be something wrong with your environment. Did you have `libjpeg` or `libpng` installed before building `torchvision` from source?
  warn(
False

Further, I found it's the torch.xpu.device_count() returns 0, causing is_xpu_available() returns False.

Versions

(I'm running a ubuntu 22.04 container in fedora silverblue 39)


python collect_env.py:

Collecting environment information...
PyTorch version: 2.0.1a0+cxx11.abi
PyTorch CXX11 ABI: Yes
IPEX version: 2.0.110+xpu
IPEX commit: ba7f6c127
Build type: Release

OS: Ubuntu 22.04.3 LTS (x86_64)
GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
Clang version: N/A
IGC version: 2023.2.0 (2023.2.0.20230721)
CMake version: N/A
Libc version: glibc-2.35

Python version: 3.11.5 (main, Sep 11 2023, 13:54:46) [GCC 11.2.0] (64-bit runtime)
Python platform: Linux-6.5.6-300.fc39.x86_64-x86_64-with-glibc2.35
Is XPU available: True
DPCPP runtime version: 2023.2.1
MKL version: 2023.2.0
GPU models and configuration: 
[0] _DeviceProperties(name='Intel(R) Arc(TM) A770 Graphics', platform_name='Intel(R) Level-Zero', dev_type='gpu, support_fp64=0, total_memory=15473MB, max_compute_units=512, gpu_eu_count=512)
[1] _DeviceProperties(name='Intel(R) HD Graphics 530', platform_name='Intel(R) Level-Zero', dev_type='gpu, support_fp64=1, total_memory=50833MB, max_compute_units=24, gpu_eu_count=24)
Intel OpenCL ICD version: 23.17.26241.33-647~22.04
Level Zero version: 1.3.26241.33-647~22.04

CPU:
Architecture:                       x86_64
CPU op-mode(s):                     32-bit, 64-bit
Address sizes:                      39 bits physical, 48 bits virtual
Byte Order:                         Little Endian
CPU(s):                             4
On-line CPU(s) list:                0-3
Vendor ID:                          GenuineIntel
Model name:                         Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz
CPU family:                         6
Model:                              94
Thread(s) per core:                 1
Core(s) per socket:                 4
Socket(s):                          1
Stepping:                           3
CPU max MHz:                        3600.0000
CPU min MHz:                        800.0000
BogoMIPS:                           6399.96
Flags:                              fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb invpcid_single pti ssbd ibrs ibpb stibp tpr_shadow flexpriority ept vpid ept_ad fsgsbase tsc_adjust sgx bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp vnmi md_clear flush_l1d arch_capabilities
Virtualization:                     VT-x
L1d cache:                          128 KiB (4 instances)
L1i cache:                          128 KiB (4 instances)
L2 cache:                           1 MiB (4 instances)
L3 cache:                           6 MiB (1 instance)
NUMA node(s):                       1
NUMA node0 CPU(s):                  0-3
Vulnerability Gather data sampling: Vulnerable: No microcode
Vulnerability Itlb multihit:        KVM: Mitigation: VMX disabled
Vulnerability L1tf:                 Mitigation; PTE Inversion; VMX conditional cache flushes, SMT disabled
Vulnerability Mds:                  Mitigation; Clear CPU buffers; SMT disabled
Vulnerability Meltdown:             Mitigation; PTI
Vulnerability Mmio stale data:      Mitigation; Clear CPU buffers; SMT disabled
Vulnerability Retbleed:             Mitigation; IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass:    Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:           Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:           Mitigation; IBRS, IBPB conditional, STIBP disabled, RSB filling, PBRSB-eIBRS Not affected
Vulnerability Srbds:                Mitigation; Microcode
Vulnerability Tsx async abort:      Mitigation; TSX disabled

Versions of relevant libraries:
[pip3] intel-extension-for-pytorch==2.0.110+xpu
[pip3] numpy==1.24.4
[pip3] torch==2.0.1a0+cxx11.abi
[pip3] torch-grammar==0.3.3
[pip3] torchvision==0.15.2a0+cxx11.abi
[conda] intel-extension-for-pytorch 2.0.110+xpu              pypi_0    pypi
[conda] numpy                     1.24.4                   pypi_0    pypi
[conda] torch                     2.0.1a0+cxx11.abi          pypi_0    pypi
[conda] torch-grammar             0.3.3                    pypi_0    pypi
[conda] torchvision               0.15.2a0+cxx11.abi          pypi_0    pypi

xpu-smi discovery:

+-----------+--------------------------------------------------------------------------------------+
| Device ID | Device Information                                                                   |
+-----------+--------------------------------------------------------------------------------------+
| 0         | Device Name: Intel(R) Arc(TM) A770 Graphics                                          |
|           | Vendor Name: Intel(R) Corporation                                                    |
|           | UUID: 00000000-0000-0003-0000-000856a08086                                           |
|           | PCI BDF Address: 0000:03:00.0                                                        |
|           | DRM Device: /dev/dri/card1                                                           |
|           | Function Type: physical                                                              |
+-----------+--------------------------------------------------------------------------------------+
| 1         | Device Name: Intel(R) HD Graphics 530                                                |
|           | Vendor Name: Intel(R) Corporation                                                    |
|           | UUID: 00000000-0000-0200-0000-000619128086                                           |
|           | PCI BDF Address: 0000:00:02.0                                                        |
|           | DRM Device: /dev/dri/card0                                                           |
|           | Function Type: physical                                                              |
+-----------+--------------------------------------------------------------------------------------+

apt search oneapi | grep installed:

intel-basekit/all,now 2023.2.0-49384 amd64 [installed]
intel-basekit-getting-started/all,now 2023.2.0-49384 all [installed,automatic]
intel-level-zero-gpu/jammy,now 1.3.26241.33-647~22.04 amd64 [installed]
intel-oneapi-advisor/all,now 2023.2.0-49486 amd64 [installed,automatic]
intel-oneapi-ccl-2021.10.0/all,now 2021.10.0-49084 amd64 [installed,automatic]
intel-oneapi-ccl-devel/all,now 2021.10.0-49084 amd64 [installed,automatic]
intel-oneapi-ccl-devel-2021.10.0/all,now 2021.10.0-49084 amd64 [installed,automatic]
intel-oneapi-common-licensing/all,now 2023.2.0-49462 all [installed,automatic]
intel-oneapi-common-licensing-2023.2.0/all,now 2023.2.0-49462 all [installed,automatic]
intel-oneapi-common-vars/all,now 2023.2.0-49462 all [installed,automatic]
intel-oneapi-compiler-dpcpp-cpp-2023.2.1/all,now 2023.2.1-16 amd64 [installed,automatic]
intel-oneapi-compiler-dpcpp-cpp-common-2023.2.1/all,now 2023.2.1-16 all [installed,automatic]
intel-oneapi-compiler-dpcpp-cpp-runtime-2023.2.0/all,now 2023.2.0-49495 amd64 [installed,automatic]
intel-oneapi-compiler-dpcpp-cpp-runtime-2023.2.1/all,now 2023.2.1-16 amd64 [installed,automatic]
intel-oneapi-compiler-shared-2023.2.1/all,now 2023.2.1-16 amd64 [installed,automatic]
intel-oneapi-compiler-shared-common-2023.2.1/all,now 2023.2.1-16 all [installed,automatic]
intel-oneapi-compiler-shared-runtime-2023.2.0/all,now 2023.2.0-49495 amd64 [installed,automatic]
intel-oneapi-compiler-shared-runtime-2023.2.1/all,now 2023.2.1-16 amd64 [installed,automatic]
intel-oneapi-condaindex/all,now 2023.2.0-49417 amd64 [installed,automatic]
intel-oneapi-dal-2023.2.0/all,now 2023.2.0-49572 amd64 [installed,automatic]
intel-oneapi-dal-common-2023.2.0/all,now 2023.2.0-49572 all [installed,automatic]
intel-oneapi-dal-common-devel-2023.2.0/all,now 2023.2.0-49572 all [installed,automatic]
intel-oneapi-dal-devel/all,now 2023.2.0-49572 amd64 [installed,automatic]
intel-oneapi-dal-devel-2023.2.0/all,now 2023.2.0-49572 amd64 [installed,automatic]
intel-oneapi-dev-utilities/all,now 2021.10.0-49423 amd64 [installed,automatic]
intel-oneapi-dev-utilities-2021.10.0/all,now 2021.10.0-49423 amd64 [installed,automatic]
intel-oneapi-dev-utilities-eclipse-cfg/all,now 2021.10.0-49423 all [installed,automatic]
intel-oneapi-diagnostics-utility/all,now 2022.4.0-49091 amd64 [installed,automatic]
intel-oneapi-dnnl/all,now 2023.2.0-49516 amd64 [installed,automatic]
intel-oneapi-dnnl-devel/all,now 2023.2.0-49516 amd64 [installed,automatic]
intel-oneapi-dpcpp-cpp-2023.2.1/all,now 2023.2.1-16 amd64 [installed,automatic]
intel-oneapi-dpcpp-ct/all,now 2023.2.0-49330 amd64 [installed,automatic]
intel-oneapi-dpcpp-ct-2023.2.0/all,now 2023.2.0-49330 amd64 [installed,automatic]
intel-oneapi-dpcpp-ct-eclipse-cfg/all,now 2023.2.0-49330 all [installed,automatic]
intel-oneapi-dpcpp-debugger/all,now 2023.2.0-49330 amd64 [installed,automatic]
intel-oneapi-dpcpp-debugger-2023.2.0/all,now 2023.2.0-49330 amd64 [installed,automatic]
intel-oneapi-icc-eclipse-plugin-cpp-2023.2.1/all,now 2023.2.1-16 all [installed,automatic]
intel-oneapi-ipp-2021.9.0/all,now 2021.9.0-49452 amd64 [installed,automatic]
intel-oneapi-ipp-common-2021.9.0/all,now 2021.9.0-49452 all [installed,automatic]
intel-oneapi-ipp-common-devel-2021.9.0/all,now 2021.9.0-49452 all [installed,automatic]
intel-oneapi-ipp-devel/all,now 2021.9.0-49452 amd64 [installed,automatic]
intel-oneapi-ipp-devel-2021.9.0/all,now 2021.9.0-49452 amd64 [installed,automatic]
intel-oneapi-ippcp-2021.8.0/all,now 2021.8.0-49490 amd64 [installed,automatic]
intel-oneapi-ippcp-common-2021.8.0/all,now 2021.8.0-49490 all [installed,automatic]
intel-oneapi-ippcp-common-devel-2021.8.0/all,now 2021.8.0-49490 all [installed,automatic]
intel-oneapi-ippcp-devel/all,now 2021.8.0-49490 amd64 [installed,automatic]
intel-oneapi-ippcp-devel-2021.8.0/all,now 2021.8.0-49490 amd64 [installed,automatic]
intel-oneapi-libdpstd-devel/all,now 2022.2.0-49284 amd64 [installed,automatic]
intel-oneapi-libdpstd-devel-2022.2.0/all,now 2022.2.0-49284 amd64 [installed,automatic]
intel-oneapi-mkl/all,now 2023.2.0-49495 amd64 [installed]
intel-oneapi-mkl-2023.2.0/all,now 2023.2.0-49495 amd64 [installed,automatic]
intel-oneapi-mkl-common-2023.2.0/all,now 2023.2.0-49495 all [installed,automatic]
intel-oneapi-mkl-common-devel-2023.2.0/all,now 2023.2.0-49495 all [installed,automatic]
intel-oneapi-mkl-devel/all,now 2023.2.0-49495 amd64 [installed,automatic]
intel-oneapi-mkl-devel-2023.2.0/all,now 2023.2.0-49495 amd64 [installed,automatic]
intel-oneapi-mpi-2021.10.0/all,now 2021.10.0-49371 amd64 [installed,automatic]
intel-oneapi-mpi-devel-2021.10.0/all,now 2021.10.0-49371 amd64 [installed,automatic]
intel-oneapi-openmp-2023.2.0/all,now 2023.2.0-49495 amd64 [installed,automatic]
intel-oneapi-openmp-2023.2.1/all,now 2023.2.1-16 amd64 [installed,automatic]
intel-oneapi-openmp-common-2023.2.0/all,now 2023.2.0-49495 all [installed,automatic]
intel-oneapi-openmp-common-2023.2.1/all,now 2023.2.1-16 all [installed,automatic]
intel-oneapi-tbb-2021.10.0/all,now 2021.10.0-49541 amd64 [installed,automatic]
intel-oneapi-tbb-common-2021.10.0/all,now 2021.10.0-49541 all [installed,automatic]
intel-oneapi-tbb-common-devel-2021.10.0/all,now 2021.10.0-49541 all [installed,automatic]
intel-oneapi-tbb-devel/all,now 2021.10.0-49541 amd64 [installed,automatic]
intel-oneapi-tbb-devel-2021.10.0/all,now 2021.10.0-49541 amd64 [installed,automatic]
intel-oneapi-vtune/all,now 2023.2.0-49484 amd64 [installed,automatic]
intel-oneapi-vtune-eclipse-plugin-vtune/all,now 2023.2.0-49484 all [installed,automatic]
level-zero/jammy,now 1.11.0-647~22.04 amd64 [installed]
xpu-smi/jammy,now 1.2.13-24~22.04 amd64 [installed]
notsyncing commented 10 months ago

btw, IPEX works well in vladmandic/automatic on the same machine and container base image.

notsyncing commented 10 months ago

I found that if I exchange the import order like this:

import accelerate; import matplotlib; print(accelerate.utils.is_xpu_available())

Then it works. Is this intentional, or any bugs?

jingxu10 commented 9 months ago

pls import ipex first. importing a package loads dynamic library files. There might be dynamic library version that doesn't work with IPEX, so importing IPEX first could avoid this problem.