Can I use version 11 or above?
If you are asking about whether you can build onnxruntime with CUDA 11.1. This has been tested and you should be able to use CUDA 11.1.
With CUDA 11.1, tests fail :-(
...
git clone ...
git checkout 718ca7f # (version 1.6.0)
CUDACXX=/usr/local/cuda-11.1/bin/nvcc ./build.sh --config Release --build_wheel --parallel --use_cuda --cuda_home /usr/local/cuda-11.1 --cudnn_home /usr/lib/x86_64-linux-gnu/
Generate:
...
[----------] 8 tests from LinearRegressorTest/LinearRegressorTest (3 ms total)
1:
1: [----------] 5 tests from CApiTestWithProviders/CApiTestWithProvider
1: [ RUN ] CApiTestWithProviders/CApiTestWithProvider.simple/0
1: Running simple inference with default provider
1: [ OK ] CApiTestWithProviders/CApiTestWithProvider.simple/0 (1 ms)
1: [ RUN ] CApiTestWithProviders/CApiTestWithProvider.simple/1
1: Running simple inference with cuda provider
1: [ OK ] CApiTestWithProviders/CApiTestWithProvider.simple/1 (3 ms)
1: [ RUN ] CApiTestWithProviders/CApiTestWithProvider.simple/2
1: [ OK ] CApiTestWithProviders/CApiTestWithProvider.simple/2 (0 ms)
1: [ RUN ] CApiTestWithProviders/CApiTestWithProvider.simple/3
1: [ OK ] CApiTestWithProviders/CApiTestWithProvider.simple/3 (0 ms)
1: [ RUN ] CApiTestWithProviders/CApiTestWithProvider.simple/4
1: Running simple inference with default provider
1: [ OK ] CApiTestWithProviders/CApiTestWithProvider.simple/4 (0 ms)
1: [----------] 5 tests from CApiTestWithProviders/CApiTestWithProvider (5 ms total)
1:
1: [----------] Global test environment tear-down
1: [==========] 2431 tests from 198 test suites ran. (27154 ms total)
1: [ PASSED ] 2418 tests.
1: [ FAILED ] 13 tests, listed below:
1: [ FAILED ] QAttentionTest.QAttentionBatch1
1: [ FAILED ] QAttentionTest.QAttentionBatch1_Float16
1: [ FAILED ] QAttentionTest.QAttentionBatch2
1: [ FAILED ] QAttentionTest.QAttentionMaskPartialSequence
1: [ FAILED ] QAttentionTest.QAttentionMaskExceedSequence
1: [ FAILED ] QAttentionTest.QAttentionNoMaskIndex
1: [ FAILED ] QAttentionTest.QAttentionUnidirectional_CUDA
1: [ FAILED ] MatmulIntegerOpTest.MatMulInteger_int8_t
1: [ FAILED ] MatmulIntegerOpTest.MatMulInteger_int8_t_A_ND
1: [ FAILED ] MatmulIntegerOpTest.MatMulInteger_int8_t_B_ND
1: [ FAILED ] MatmulIntegerOpTest.MatMulInteger_int8_t_A_ND_B_ND
1: [ FAILED ] MatmulIntegerOpTest.MatMulInteger_int8_t_A_Has_Zero_Point
1: [ FAILED ] MatmulIntegerOpTest.MatMulInteger_int8_t_No_Zero_Point
1:
1: 13 FAILED TESTS
1: YOU HAVE 6 DISABLED TESTS
1:
1/3 Test #1: onnxruntime_test_all .............***Failed 27.70 sec
test 2
Start 2: onnx_test_pytorch_converted
2: Test command: /mnt/workspace/onnxruntime/build/Linux/Release/onnx_test_runner "/mnt/workspace/onnxruntime/cmake/external/onnx/onnx/backend/test/data/pytorch-converted"
2: Test timeout computed to be: 3600
2: 2020-12-19 23:32:42.739477607 [E:onnxruntime:Default, testcase_driver.cc:39 RunParallel] Running tests in parallel: at most 12 models at any time
2: 2020-12-19 23:32:42.781376554 [E:onnxruntime:Default, testcase_driver.cc:61 RunModelsAsync] Running tests finished. Generating report
2: result:
2: Models: 57
2: Total test cases: 57
2: Succeeded: 57
2: Not implemented: 0
2: Failed: 0
2: Stats by Operator type:
2: Not implemented(0):
2: Failed:
2: Failed Test Cases:
2/3 Test #2: onnx_test_pytorch_converted ...... Passed 0.07 sec
test 3
Start 3: onnx_test_pytorch_operator
3: Test command: /mnt/workspace/onnxruntime/build/Linux/Release/onnx_test_runner "/mnt/workspace/onnxruntime/cmake/external/onnx/onnx/backend/test/data/pytorch-operator"
3: Test timeout computed to be: 3600
3: 2020-12-19 23:32:42.811606633 [E:onnxruntime:Default, testcase_driver.cc:39 RunParallel] Running tests in parallel: at most 12 models at any time
3: 2020-12-19 23:32:42.833771281 [E:onnxruntime:Default, testcase_driver.cc:61 RunModelsAsync] Running tests finished. Generating report
3: result:
3: Models: 24
3: Total test cases: 24
3: Succeeded: 24
3: Not implemented: 0
3: Failed: 0
3: Stats by Operator type:
3: Not implemented(0):
3: Failed:
3: Failed Test Cases:
3/3 Test #3: onnx_test_pytorch_operator ....... Passed 0.05 sec
67% tests passed, 1 tests failed out of 3
Total Test time (real) = 27.84 sec
The following tests FAILED:
1 - onnxruntime_test_all (Failed)
Errors while running CTest
Traceback (most recent call last):
File "/home/geantvert/workspace/onnxruntime/tools/ci_build/build.py", line 2088, in <module>
sys.exit(main())
File "/home/geantvert/workspace/onnxruntime/tools/ci_build/build.py", line 2026, in main
run_onnxruntime_tests(args, source_dir, ctest_path, build_dir, configs)
File "/home/geantvert/workspace/onnxruntime/tools/ci_build/build.py", line 1450, in run_onnxruntime_tests
run_subprocess(ctest_cmd, cwd=cwd, dll_path=dll_path)
File "/home/geantvert/workspace/onnxruntime/tools/ci_build/build.py", line 523, in run_subprocess
return run(*args, cwd=cwd, capture_stdout=capture_stdout, shell=shell, env=my_env)
File "/mnt/workspace/onnxruntime/tools/python/util/run.py", line 39, in run
completed_process = subprocess.run(
File "/usr/lib/python3.8/subprocess.py", line 512, in run
raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['/usr/bin/ctest', '--build-config', 'Release', '--verbose', '--timeout', '3600']' returned non-zero exit status 8.
Most helpful comment
If you are asking about whether you can build onnxruntime with CUDA 11.1. This has been tested and you should be able to use CUDA 11.1.