Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[AutoTVM] Failed to run autotvm example on GEMM #4717

Closed
Airtnp opened this issue Jan 15, 2020 · 1 comment
Closed

[AutoTVM] Failed to run autotvm example on GEMM #4717

Airtnp opened this issue Jan 15, 2020 · 1 comment

Comments

@Airtnp
Copy link

Airtnp commented Jan 15, 2020

I was trying to use code from https://zhuanlan.zhihu.com/p/75203171 (it's a Chinese post, but it's essential the same as Writing tunable template and Using auto-tuner) with modified size 4096^3. However, in my AWS instance, it keeps reporting tons of similar error like this

DEBUG:autotvm:No: 199 GFLOPS: 0.00/0.00 result: MeasureResult(costs=(TVMError('Traceback (most recent call last):\n [bt] (3) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(TVMFuncCall+0x61) [0x7f51e3d495e1]\n [bt] (2) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(ZNSt17_Function_handlerIFvN3tvm7runtime7TVMArgsEPNS1_11TVMRetValueEEZNS1_15TypedPackedFuncIFRNS0_5StageES7_NS0_7IterVarEEE17AssignTypedLambdaIZNS1_8Registry15set_body_methodIS7_S8_IS9_EEERSD_MT_FT0_DpT1_EEUlS7_S9_E_EEvSG_EUlRKS2_S4_E_E9_M_invokeERKSt9_Any_dataOS2_OS4+0x98) [0x7f51e3570c48]\n [bt] (1) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::Stage::vectorize(tvm::IterVar)+0xb0) [0x7f51e38e7c70]\n [bt] (0) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(dmlc::LogMessageFatal::~LogMessageFatal()+0x32) [0x7f51e35375e2]\n File "/home/ubuntu/tvm/src/schedule/schedule_lang.cc", line 377\nTVMError: Check failed: var->iter_type == kDataPar || var->iter_type == kOpaque || var->iter_type == kUnrolled || var->iter_type == kVectorized || var->iter_type == kTensorized || var->iter_type == kParallelized: Cannot vectorize on CommReduce',),), error_no=2, all_cost=0.006806135177612305, timestamp=1579123976.0153828) [('tile_x', [-1, 8, 32]), ('tile_y', [-1, 16, 32]), ('tile_k', [-1, 2048]), ('reorder', (2, 0, 1)), ('ann', ['unroll', 'vec', 'unroll'])],,None,10215900

I tried with code in Writing tunable template and Using auto-tuner and it also reports errors like

No: 9 GFLOPS: 0.00/0.00 result: MeasureResult(costs=(RuntimeError('Traceback (most recent call last):\n [bt] (3) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(TVMFuncCall+0x61) [0x7f6bdfd335e1]\n [bt] (2) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(std::_Function_handler<void (tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*), tvm::runtime::RPCModuleNode::WrapRemote(void*)::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#1}>::_M_invoke(std::_Any_data const&, tvm::runtime::TVMArgs&&, tvm::runtime::TVMRetValue*&&)+0x41) [0x7f6bdfd914f1]\n [bt] (1) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(tvm::runtime::RPCSession::CallFunc(void*, tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*, void* ()(int, tvm::runtime::TVMArgValue const&), tvm::runtime::PackedFunc const)+0x167) [0x7f6bdfd9afb7]\n [bt] (0) /home/ubuntu/.local/lib/python3.6/site-packages/tvm-0.6.0-py3.6-linux-x86_64.egg/tvm/libtvm.so(dmlc::LogMessageFatal::~Lo',),), error_no=4, all_cost=10.207964420318604, timestamp=1579122195.5362496) [('tile_y', [-1, 128]), ('tile_x', [-1, 8])],,None,46

What's wrong with the code? Thank you for your help!

@tqchen
Copy link
Member

tqchen commented Jan 15, 2020

Thanks for asking the question, please open a new trouble shooting thread on https://discuss.tvm.ai/

@tqchen tqchen closed this as completed Jan 15, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants