[DOCS] Try upgrade build (apache#1066)

Jenkinsfile

@@ -87,7 +87,6 @@ stage('Build') {
            cp make/config.mk .
            echo USE_CUDNN=1 >> config.mk
            echo USE_CUDA=1 >> config.mk
-           echo USE_OPENCL=1 >> config.mk
            echo USE_OPENGL=1 >> config.mk
            echo LLVM_CONFIG=llvm-config-4.0 >> config.mk
            echo USE_RPC=1 >> config.mk
@@ -105,6 +104,7 @@ stage('Build') {
         sh "mv lib/libtvm.so lib/libtvm_llvm60.so"
         pack_lib('gpu', tvm_multilib)
         sh """
+           echo USE_OPENCL=1 >> config.mk
            echo USE_ROCM=1 >> config.mk
            echo ROCM_PATH=/opt/rocm >> config.mk
            echo USE_VULKAN=1 >> config.mk
@@ -152,31 +152,6 @@ stage('Build') {
         pack_lib('i386', tvm_multilib)
       }
     }
-  },
-  'web': {
-    node('emcc') {
-      ws('workspace/tvm/build-weblib') {
-        init_git()
-        sh """
-           cp make/config.mk .
-           echo USE_CUDA=0 >> config.mk
-           echo USE_OPENCL=0 >> config.mk
-           echo LLVM_CONFIG=llvm-config >> config.mk
-           echo USE_RPC=0 >> config.mk
-           """
-        sh "${docker_run} emscripten echo testing javascript..."
-        timeout(time: max_time, unit: 'MINUTES') {
-          try {
-            sh "${docker_run} emscripten ./tests/scripts/task_web_build.sh"
-          } catch (exc) {
-            echo 'Incremental compilation failed. Fall back to build from scratch'
-            sh "${docker_run} emscripten make clean"
-            sh "${docker_run} emscripten ./tests/scripts/task_web_build.sh"
-         }
-        }
-        pack_lib('weblib', tvm_lib)
-      }
-    }
   }
 }
 
@@ -256,18 +231,6 @@ stage('Integration Test') {
       }
     }
   },
-  'web': {
-    node('emcc') {
-      ws('workspace/tvm/it-weblib') {
-        init_git()
-        unpack_lib('weblib', tvm_lib)
-        sh "${docker_run} emscripten echo testing javascript..."
-        timeout(time: max_time, unit: 'MINUTES') {
-          sh "${docker_run} emscripten ./tests/scripts/task_web_test.sh"
-        }
-      }
-    }
-  },
   'docs': {
     node('GPU' && 'linux') {
       ws('workspace/tvm/docs-python-gpu') {

python/tvm/_ffi/function.py

@@ -181,10 +181,10 @@ def register(myf):
             myf = convert_to_tvm_func(myf)
         check_call(_LIB.TVMFuncRegisterGlobal(
             c_str(func_name), myf.handle, ioverride))
+        return myf
     if f:
-        register(f)
-    else:
-        return register
+        return register(f)
+    return register
 
 
 def get_global_func(name, allow_missing=False):

python/tvm/api.py

@@ -652,6 +652,7 @@ def _make_reduce(expr, axis, where=None):
                         for i in range(size))
         return outputs[0] if size == 1 else outputs
 
+    # pylint: disable=keyword-arg-before-vararg
     def reducer(expr, axis, where=None, *args):
         if isinstance(axis, (_schedule.IterVar, list, tuple)):
             assert not args

tests/ci_build/Dockerfile.emscripten

@@ -15,4 +15,8 @@ RUN bash /install/ubuntu_install_emscripten.sh
 COPY install/ubuntu_install_python_package.sh /install/ubuntu_install_python_package.sh
 RUN bash /install/ubuntu_install_python_package.sh
 
-RUN cp /root/.emscripten /emsdk-portable/
+RUN chmod a+rwx -R /emsdk-portable
+RUN cp -r /emsdk-portable  /emsdk-portable-backup
+RUN mv /emsdk-portable  /emsdk-portable-x
+RUN mv /emsdk-portable-backup /emsdk-portable
+RUN cp /root/.emscripten /emsdk-portable/

tests/ci_build/Dockerfile.gpu

@@ -1,7 +1,6 @@
 FROM nvidia/cuda:8.0-cudnn7-devel
 
 # Base scripts
-RUN apt-get update --fix-missing
 
 COPY install/ubuntu_install_core.sh /install/ubuntu_install_core.sh
 RUN bash /install/ubuntu_install_core.sh
@@ -12,9 +11,6 @@ RUN bash /install/ubuntu_install_python.sh
 COPY install/ubuntu_install_llvm.sh /install/ubuntu_install_llvm.sh
 RUN bash /install/ubuntu_install_llvm.sh
 
-COPY install/ubuntu_install_opencl.sh /install/ubuntu_install_opencl.sh
-RUN bash /install/ubuntu_install_opencl.sh
-
 COPY install/ubuntu_install_iverilog.sh /install/ubuntu_install_iverilog.sh
 RUN bash /install/ubuntu_install_iverilog.sh
 
@@ -40,8 +36,11 @@ RUN bash /install/ubuntu_install_rocm.sh
 COPY install/ubuntu_install_opengl.sh /install/ubuntu_install_opengl.sh
 RUN bash /install/ubuntu_install_opengl.sh
 
+COPY install/ubuntu_install_opencl.sh /install/ubuntu_install_opencl.sh
+RUN bash /install/ubuntu_install_opencl.sh
+
 # Enable doxygen for c++ doc build
-RUN apt-get install -y doxygen graphviz
+RUN apt-get update && apt-get install -y doxygen graphviz
 
 # Install vulkan
 COPY install/ubuntu_install_vulkan.sh /install/ubuntu_install_vulkan.sh

tests/ci_build/install/ubuntu_install_core.sh

@@ -1,5 +1,5 @@
 # install libraries for building c++ core on ubuntu
-apt-get install -y --no-install-recommends --force-yes \
+apt-get update && apt-get install -y --no-install-recommends --force-yes \
         git make libgtest-dev cmake wget unzip libtinfo-dev libz-dev\
         libcurl4-openssl-dev libopenblas-dev g++ sudo
 

tests/ci_build/install/ubuntu_install_llvm.sh

@@ -8,6 +8,11 @@ echo deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-5.0 main\
 echo deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-5.0 main\
      >> /etc/apt/sources.list.d/llvm.list
 
+echo deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial-6.0 main\
+     >> /etc/apt/sources.list.d/llvm.list
+echo deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial-6.0 main\
+     >> /etc/apt/sources.list.d/llvm.list
+
 echo deb http://apt.llvm.org/xenial/ llvm-toolchain-xenial main\
      >> /etc/apt/sources.list.d/llvm.list
 echo deb-src http://apt.llvm.org/xenial/ llvm-toolchain-xenial main\

tests/ci_build/install/ubuntu_install_opencl.sh

@@ -1,8 +1,8 @@
 # Install OpenCL runtime in nvidia docker.
-apt-get install -y --no-install-recommends --force-yes \
-        ocl-icd-libopencl1 \
+apt-get update && apt-get install -y --no-install-recommends --force-yes \
+        ocl-icd-opencl-dev \
         clinfo && \
-        rm -rf /var/lib/apt/lists/*
+    rm -rf /var/lib/apt/lists/*
 
 mkdir -p /etc/OpenCL/vendors && \
     echo "libnvidia-opencl.so.1" > /etc/OpenCL/vendors/nvidia.icd

tests/ci_build/install/ubuntu_install_sphinx.sh

@@ -1 +1 @@
-pip install sphinx==1.6.2 sphinx-gallery sphinx_rtd_theme matplotlib Image commonmark>=0.7.3 docutils>=0.11
+pip install sphinx sphinx-gallery sphinx_rtd_theme matplotlib Image commonmark>=0.7.3 docutils>=0.11

topi/python/topi/cuda/conv2d_transpose_nchw.py

@@ -74,6 +74,7 @@ def schedule(temp, Filter, Output):
             conv2d_56_64_64(s, Filter, temp_S, Filter_S, Out, Out_L)
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_injective(OP.tag):
             if OP not in s.outputs:

topi/python/topi/cuda/dense.py

@@ -87,6 +87,7 @@ def _schedule(Dense):
         s[Out].set_store_predicate(thread_x.var.equal(0))
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

topi/python/topi/cuda/depthwise_conv2d.py

@@ -102,6 +102,7 @@ def _schedule(PaddedInput, Filter, DepthwiseConv2d):
         s[FS].bind(tx, thread_x)
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:
@@ -178,6 +179,7 @@ def _schedule(temp, Filter, DepthwiseConv2d):
         s[FS].bind(fused, thread_x)
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

topi/python/topi/cuda/pooling.py

@@ -47,6 +47,7 @@ def _schedule(Pool):
             s[Pool].compute_at(s[Out], tx)
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:
@@ -101,6 +102,7 @@ def _schedule(PaddedInput, Pool):
             s[Pool].compute_at(s[Out], tx)
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

topi/python/topi/cuda/reduction.py

@@ -87,6 +87,7 @@ def schedule_reduce(outs):
     sch = tvm.create_schedule([x.op for x in outs])
 
     def traverse_before_reduce(operator):
+        """Internal travserse function"""
         if isinstance(operator, tvm.tensor.PlaceholderOp):
             return
         elif tag.is_injective(operator.tag):
@@ -97,6 +98,7 @@ def traverse_before_reduce(operator):
             raise RuntimeError("Unsupported operator: %s" % operator.tag)
 
     def traverse_after_reduce(operator):
+        """Internal travserse function"""
         if tag.is_broadcast(operator.tag):
             raise RuntimeError("Not yet support ewise after reduce")
         elif operator.tag == 'comm_reduce':

topi/python/topi/mali/dense.py

@@ -82,6 +82,7 @@ def fuse_and_bind(s, tensor, axis=None, num_thread=None):
 #        print(tvm.lower(s, [data, weight, bias, outs[0]], simple_mode=True))
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

topi/python/topi/mali/depthwise_conv2d.py

@@ -87,6 +87,7 @@ def tile_and_bind3d(tensor, z, y, x, z_factor=2, y_factor=None, x_factor=None):
             s[conv].compute_at(s[output], ji)
 
     def traverse(op):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(op.tag):
             if op not in s.outputs:

topi/python/topi/nn/bnn.py

@@ -43,6 +43,7 @@ def _binarize_pack(*indices):
             if j == 31:
                 return packed
             packed = packed << 1
+        raise RuntimeError("not resach")
 
     return tvm.compute(oshape, _binarize_pack, name=name, tag='binarize_pack')
 

topi/python/topi/opengl/conv2d_nchw.py

@@ -31,6 +31,7 @@ def _schedule(conv2d, data):
         s[data].opengl()
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

topi/python/topi/opengl/dense.py

@@ -31,6 +31,7 @@ def _schedule(Dense):
         s[Out].opengl()
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

topi/python/topi/opengl/pooling.py

@@ -30,6 +30,7 @@ def _schedule(Pool):
         s[Out].opengl()
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:
@@ -75,6 +76,7 @@ def _schedule(PaddedInput, Pool):
         s[Out].opengl()
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

topi/python/topi/rasp/depthwise_conv2d.py

@@ -164,6 +164,7 @@ def schedule_depthwise_conv2d(outs):
     s = tvm.create_schedule([x.op for x in outs])
 
     def traverse(op):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(op.tag):
             if op not in s.outputs:

topi/python/topi/x86/binary_dense.py

@@ -35,6 +35,7 @@ def _schedule(A, B, C):
         s[Out].vectorize(xi)
 
     def traverse(OP):
+        """Internal travserse function"""
         # inline all one-to-one-mapping operators except the last stage (output)
         if tag.is_broadcast(OP.tag):
             if OP not in s.outputs:

tutorials/deployment/cross_compilation_and_rpc.py

@@ -108,8 +108,6 @@
 import numpy as np
 from tvm.contrib import rpc, util
 
-server = rpc.Server(host='0.0.0.0', port=9090, use_popen=True)
-
 ######################################################################
 # Declare and Cross Compile Kernel on Local Machine
 # -------------------------------------------------
@@ -241,47 +239,52 @@
 #    But here we set 'llvm' to enable this tutorial to run locally.
 #
 #    Also we need to build the runtime with the flag `USE_OPENCL=1`.
-
 # build kernel (different from cpu, we need bind axis for OpenCL)
-s = tvm.create_schedule(B.op)
-xo, xi = s[B].split(B.op.axis[0], factor=32)
-s[B].bind(xo, tvm.thread_axis("blockIdx.x"))
-s[B].bind(xi, tvm.thread_axis("threadIdx.x"))
-f = tvm.build(s, [A, B], "opencl", target_host="llvm", name="myadd")
+#
+# The following functions shows how we can deploy CL
+def deploy_cl():
+    s = tvm.create_schedule(B.op)
+    xo, xi = s[B].split(B.op.axis[0], factor=32)
+    s[B].bind(xo, tvm.thread_axis("blockIdx.x"))
+    s[B].bind(xi, tvm.thread_axis("threadIdx.x"))
+    f = tvm.build(s, [A, B], "opencl", target_host="llvm", name="myadd")
 
-# save files
-path_o = temp.relpath("myadd.o")
-path_cl = temp.relpath("myadd.cl")
-path_json = temp.relpath("myadd.tvm_meta.json")
-f.save(path_o)
-f.imported_modules[0].save(path_cl)
+    # save files
+    path_o = temp.relpath("myadd.o")
+    path_cl = temp.relpath("myadd.cl")
+    path_json = temp.relpath("myadd.tvm_meta.json")
+    f.save(path_o)
+    f.imported_modules[0].save(path_cl)
 
-# upload files
-remote.upload(path_o)
-remote.upload(path_cl)
-remote.upload(path_json)
+    # upload files
+    remote.upload(path_o)
+    remote.upload(path_cl)
+    remote.upload(path_json)
 
-# load files on remote device
-fhost = remote.load_module("myadd.o")
-fdev = remote.load_module("myadd.cl")
-fhost.import_module(fdev)
+    # load files on remote device
+    fhost = remote.load_module("myadd.o")
+    fdev = remote.load_module("myadd.cl")
+    fhost.import_module(fdev)
+
+    # run
+    ctx = remote.cl(0)
+    a = tvm.nd.array(np.random.uniform(size=1024).astype(A.dtype), ctx)
+    b = tvm.nd.array(np.zeros(1024, dtype=A.dtype), ctx)
+    fhost(a, b)
+    np.testing.assert_equal(b.asnumpy(), a.asnumpy() + 1)
 
-# run
-ctx = remote.cl(0)
-a = tvm.nd.array(np.random.uniform(size=1024).astype(A.dtype), ctx)
-b = tvm.nd.array(np.zeros(1024, dtype=A.dtype), ctx)
-fhost(a, b)
-np.testing.assert_equal(b.asnumpy(), a.asnumpy() + 1)
 
 #####################################################################
 # Instead of uploading files separately, there is a more convinient way.
 # You can export libraray as a tar ball.
-path_tar = temp.relpath("myadd.tar")
-f.export_library(path_tar)
-remote.upload(path_tar)
-fhost = remote.load_module("myadd.tar")
-fhost(a, b)
-np.testing.assert_equal(b.asnumpy(), a.asnumpy() + 1)
+# The following functions shows how we can deploy by tar ball
+def deploy_cl_by_tar():
+    path_tar = temp.relpath("myadd.tar")
+    f.export_library(path_tar)
+    remote.upload(path_tar)
+    fhost = remote.load_module("myadd.tar")
+    fhost(a, b)
+    np.testing.assert_equal(b.asnumpy(), a.asnumpy() + 1)
 
 # terminate the server after experiment
 server.terminate()