The Runtime Error about nnvm_quick_start.py

Development
#1

This first my time to learn TVM.
so hen I compiled TVM. Two questions ocurred.
Without enable LLVM , I run the nnvm_quick_start.py
I have the Check failed: bf != nullptr Target llvm is not enabled

Then I tried to compile with LLVM.I downloaded the pre-built version LLVM.But there is no
llvm-config,How can i get this config?

I can successful run opt_conv_cuda.py

#2

If you can run opt_conv_cuda.py, then you build the tvm with llvm successfully.
You should also be able to run nnvm_quick_start.py.

#3

Which OS are you using ?

#4

Continuing the discussion from The Runtime Error about nnvm_quick_start.py:

Thank you for reply.
I finally enable the llvm lib.
But there is some new issues.
Traceback (most recent call last): File "nnvm_quick_start.py", line 83, in <module> net, target, shape={"data": data_shape}, params=params) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\nnvm-0.8.0-py3.6.egg\nnvm\compiler\build_module.py", line 305, in build graph = graph.apply("GraphCompile") File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\nnvm-0.8.0-py3.6.egg\nnvm\graph.py", line 234, in apply check_call(_LIB.NNGraphApplyPasses(self.handle, npass, cpass, ctypes.byref(ghandle))) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\nnvm-0.8.0-py3.6.egg\nnvm\_base.py", line 75, in check_call raise NNVMError(py_str(_LIB.NNGetLastError())) nnvm._base.NNVMError: TVMCall CFunc Error: Traceback (most recent call last): File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\_ffi\_ctypes\function.py", line 55, in cfun rv = local_pyfunc(*pyargs) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\nnvm-0.8.0-py3.6.egg\nnvm\compiler\build_module.py", line 124, in _build return tvm.build(funcs, target=target, target_host=target_host) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\build_module.py", line 504, in build mdev = codegen.build_module(fdevice, str(target_device)) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\codegen.py", line 20, in build_module return _Build(lowered_func, target) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\_ffi\_ctypes\function.py", line 185, in __call__ ctypes.byref(ret_val), ctypes.byref(ret_tcode))) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\_ffi\base.py", line 66, in check_call raise TVMError(py_str(_LIB.TVMGetLastError())) tvm._ffi.base.TVMError: TVMCall CFunc Error: Traceback (most recent call last): File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\_ffi\_ctypes\function.py", line 55, in cfun rv = local_pyfunc(*pyargs) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\autotvm\measure\measure_methods.py", line 560, in tvm_callback_cuda_compile ptx = nvcc.compile_cuda(code, target="ptx", arch=AutotvmGlobalScope.current.cuda_target_arch) File "C:\Users\wall_painter\AppData\Roaming\Python\Python36\site-packages\tvm-0.5.dev0-py3.6-win-amd64.egg\tvm\contrib\nvcc.py", line 82, in compile_cuda raise RuntimeError(msg) RuntimeError: Compilation error: nvcc fatal : Host compiler targets unsupported OS.

Before this error.
There is something like:cl.exe not in the path

But i still can run opt_conv_cuda.py

#5

Windows,I just compiled llvm by myself.LLVM prebuild version on Windows doesn’t have config

#6

you don’t need llvm-config as long as cmake can find llvm. On windows I recommend building llvm from source. It’s easy and always works.

#7

Yes…I finished yesterday.