fix conflict

CONTRIBUTORS.md

@@ -55,10 +55,12 @@ We do encourage everyone to work anything they are interested in.
 - [Marisa Kirisame](https://github.com/MarisaKirisame): @MarisaKirisame - relay
 - [Wuwei Lin](https://github.com/vinx13): @vinx13 - relay, topi
 - [Yizhi Liu](https://github.com/yzhliu) (PMC): @yzhliu - jvm, topi, relay
+- [Steven Lyubomirsky](https://github.com/slyubomirsky): @slyubomirsky - relay
 - [Hao Lu](https://github.com/hlu1): @hlu1 - nnpack, frontends
 - [Masahiro Masuda](https://github.com/masahi) (PMC): @masahi - topi, relay
 - [Thierry Moreau](https://github.com/tmoreau89) (PMC): @tmoreau89 - vta
 - [Kazutaka Morita](https://github.com/kazum): @kazum - frontends, opencl
+- [Leandro Nunes](https://github.com/leandron): @leandron - tvmc
 - [Krzysztof Parzyszek](https://github.com/kparzysz-quic): @kparzysz-quic - hexagon, llvm
 - [Andrew Reusch](https://github.com/areusch): @areusch - runtime, µTVM
 - [Jared Roesch](https://github.com/jroesch) (PMC): @jroesch - relay
@@ -111,6 +113,7 @@ We do encourage everyone to work anything they are interested in.
 - [Wuwei Lin](https://github.com/vinx13): @vinx13
 - [Andrew Liu](https://github.com/hypercubestart): @hypercubestart
 - [Henry Liu](https://github.com/optima2005): @optima2005
+- [Xin Liu](https://github.com/Meteorix): @Meteorix
 - [Steven Lyubomirsky](https://github.com/slyubomirsky): @slyubomirsky
 - [Masahiro Masuda](https://github.com/masahi): @masahi
 - [Sergey Mironov](https://github.com/grwlf): @grwlf
@@ -123,6 +126,7 @@ We do encourage everyone to work anything they are interested in.
 - [Pariksheet Pinjari](https://github.com/PariksheetPinjari909): @PariksheetPinjari909
 - [Josh Pollock](https://github.com/joshpoll): @joshpoll
 - [Jared Roesch](https://github.com/jroesch): @jroesch
+- [Giuseppe Rossini](https://github.com/giuseros): @giuseros
 - [Andrew Reusch](https://github.com/areusch): @areusch
 - [Dmitriy Smirnov](https://github.com/d-smirnov): @d-smirnov
 - [Siva](https://github.com/srkreddy1238): @srkreddy1238

Jenkinsfile

@@ -49,7 +49,7 @@ ci_gpu = "tlcpack/ci-gpu:v0.72"
 ci_cpu = "tlcpack/ci-cpu:v0.73"
 ci_wasm = "tlcpack/ci-wasm:v0.70"
 ci_i386 = "tlcpack/ci-i386:v0.72-t0"
-ci_qemu = "tlcpack/ci-qemu:v0.03"
+ci_qemu = "tlcpack/ci-qemu:v0.04"
 ci_arm = "tlcpack/ci-arm:v0.03"
 // <--- End of regex-scanned config.
 

apps/benchmark/arm_cpu_imagenet_bench.py

@@ -98,7 +98,7 @@ def evaluate_network(network, target, target_host, repeat):
         help="The model of the test device. If your device is not listed in "
         "the choices list, pick the most similar one as argument.",
     )
-    parser.add_argument("--host", type=str, default="localhost")
+    parser.add_argument("--host", type=str, default="127.0.0.1")
     parser.add_argument("--port", type=int, default=9190)
     parser.add_argument("--rpc-key", type=str, required=True)
     parser.add_argument("--repeat", type=int, default=10)

apps/benchmark/mobile_gpu_imagenet_bench.py

@@ -98,7 +98,7 @@ def evaluate_network(network, target, target_host, dtype, repeat):
         help="The model of the test device. If your device is not listed in "
         "the choices list, pick the most similar one as argument.",
     )
-    parser.add_argument("--host", type=str, default="localhost")
+    parser.add_argument("--host", type=str, default="127.0.0.1")
     parser.add_argument("--port", type=int, default=9190)
     parser.add_argument("--rpc-key", type=str, required=True)
     parser.add_argument("--repeat", type=int, default=30)

apps/bundle_deploy/bundle.c

@@ -23,8 +23,8 @@
 #include <tvm/runtime/c_runtime_api.h>
 #include <tvm/runtime/crt/crt.h>
 #include <tvm/runtime/crt/graph_executor.h>
-#include <tvm/runtime/crt/memory.h>
 #include <tvm/runtime/crt/packed_func.h>
+#include <tvm/runtime/crt/page_allocator.h>
 
 #ifdef ENABLE_TVM_ABORT_BACKTRACE
 #include "backtrace.h"
@@ -64,8 +64,8 @@ TVM_DLL void* tvm_runtime_create(const char* json_data, const char* params_data,
   dev.device_id = device_id;
 
   // declare pointers
-  TVM_CCALL(MemoryManagerCreate(&g_memory_manager, g_crt_memory, sizeof(g_crt_memory),
-                                CRT_MEMORY_PAGE_SIZE_LOG2));
+  TVM_CCALL(PageMemoryManagerCreate(&g_memory_manager, g_crt_memory, sizeof(g_crt_memory),
+                                    CRT_MEMORY_PAGE_SIZE_LOG2));
   TVM_CCALL(TVMInitializeRuntime());
   TVMPackedFunc pf;
   TVMArgs args = TVMArgs_Create(NULL, NULL, 0);

apps/bundle_deploy/bundle_static.c

@@ -22,8 +22,8 @@
 #include <stdlib.h>
 #include <tvm/runtime/crt/crt.h>
 #include <tvm/runtime/crt/graph_executor.h>
-#include <tvm/runtime/crt/memory.h>
 #include <tvm/runtime/crt/packed_func.h>
+#include <tvm/runtime/crt/page_allocator.h>
 #include <unistd.h>
 
 #ifdef ENABLE_TVM_PLATFORM_ABORT_BACKTRACE
@@ -64,8 +64,8 @@ TVM_DLL void* tvm_runtime_create(const char* json_data, const char* params_data,
   dev.device_id = device_id;
 
   // get pointers
-  TVM_CCALL(MemoryManagerCreate(&g_memory_manager, g_crt_memory, sizeof(g_crt_memory),
-                                CRT_MEMORY_PAGE_SIZE_LOG2));
+  TVM_CCALL(PageMemoryManagerCreate(&g_memory_manager, g_crt_memory, sizeof(g_crt_memory),
+                                    CRT_MEMORY_PAGE_SIZE_LOG2));
   TVM_CCALL(TVMInitializeRuntime());
   TVMPackedFunc pf;
   TVMArgs args = TVMArgs_Create(NULL, NULL, 0);

apps/bundle_deploy/crt_config/crt_config.h

@@ -45,4 +45,7 @@
 /*! Size of the global function registry, in bytes. */
 #define TVM_CRT_GLOBAL_FUNC_REGISTRY_SIZE_BYTES 200
 
+/*! Maximum packet size, in bytes, including the length header. */
+#define TVM_CRT_MAX_PACKET_SIZE_BYTES 512
+
 #endif  // TVM_RUNTIME_CRT_CONFIG_H_

apps/microtvm/reference-vm/zephyr/README.md

@@ -21,10 +21,16 @@ This directory contains setup files for Zephyr virtual machine used for testing
 that are supported by [Zephyr Project](https://zephyrproject.org/).
 
 ## VM Information for Developers
-Zephyr VM is published under [tlcpack/microtvm-zephyr](https://app.vagrantup.com/tlcpack/boxes/microtvm-zephyr).
+Zephyr VM is published under [tlcpack](https://app.vagrantup.com/tlcpack).
 Here is a list of different release versions and their tools.
 
-### [v0.0.4](https://app.vagrantup.com/tlcpack/boxes/microtvm-zephyr/versions/0.0.4)
+### [microtvm-zephyr-2.5](https://app.vagrantup.com/tlcpack/boxes/microtvm-zephyr-2.5/versions/0.0.1)
+
+- Zephyr [version 2.5.0]
+- Zephyr SDK [version 0.12.3]
+- nRFjProg [version 10.12.1]
+
+### [microtvm-zephyr-2.4](https://app.vagrantup.com/tlcpack/boxes/microtvm-zephyr/versions/0.0.4)
 
 - Zephyr [version 2.4.0]
 - Zephyr SDK [version 0.11.3]

apps/microtvm/reference-vm/zephyr/base-box/setup.sh

@@ -81,10 +81,8 @@ pip3 install --user -U west
 echo 'export PATH=$HOME/.local/bin:"$PATH"' >> ~/.profile
 source ~/.profile
 echo PATH=$PATH
-west init --mr v2.5.0 ~/zephyr
-cd ~/zephyr
-west update
-west zephyr-export
+REPO_ROOT=$(git rev-parse --show-toplevel)
+${REPO_ROOT}/docker/install/ubuntu_init_zephyr_project.sh ~/zephyr v2.5.0
 
 cd ~
 echo "Downloading zephyr SDK..."

cmake/modules/StandaloneCrt.cmake

@@ -44,6 +44,7 @@ if(USE_MICRO)
          "src/runtime/crt/include *.h -> include"
          "src/runtime/crt/common *.c -> src/runtime/crt/common"
          "src/runtime/crt/graph_executor *.c -> src/runtime/crt/graph_executor"
+         "src/runtime/crt/aot_executor *.c -> src/runtime/crt/aot_executor"
          "src/runtime/crt/graph_executor_module *.c -> src/runtime/crt/graph_executor_module"
          "src/runtime/crt/host crt_config.h -> template/host"
          "src/runtime/crt/host *.cc -> template/host"
@@ -97,7 +98,7 @@ if(USE_MICRO)
     set(make_quiet )
     endif(${VERBOSE})
 
-    list(APPEND crt_libraries memory graph_executor utvm_rpc_server utvm_rpc_common common)  # NOTE: listed in link order.
+    list(APPEND crt_libraries memory graph_executor aot_executor utvm_rpc_server utvm_rpc_common common)  # NOTE: listed in link order.
     foreach(crt_lib_name IN LISTS crt_libraries)
       list(APPEND crt_library_paths "host_standalone_crt/lib${crt_lib_name}.a")
     endforeach()

cmake/modules/VTA.cmake

@@ -103,7 +103,12 @@ elseif(PYTHON)
       # Rules for Pynq v2.4
       find_library(__cma_lib NAMES cma PATH /usr/lib)
     elseif(${VTA_TARGET} STREQUAL "de10nano")  # DE10-Nano rules
-      file(GLOB FPGA_RUNTIME_SRCS ${VTA_HW_PATH}/src/de10nano/*.cc ${VTA_HW_PATH}/src/*.cc)
+      file(GLOB DE10_FPGA_RUNTIME_SRCS ${VTA_HW_PATH}/src/de10nano/*.cc ${VTA_HW_PATH}/src/*.cc)
+      list(APPEND FPGA_RUNTIME_SRCS ${DE10_FPGA_RUNTIME_SRCS})
+    elseif(${VTA_TARGET} STREQUAL "intelfocl")  # Intel OpenCL for FPGA rules
+      file(GLOB FOCL_SRC ${VTA_HW_PATH}/src/oclfpga/*.cc)
+      list(APPEND FPGA_RUNTIME_SRCS ${FOCL_SRC})
+      list(APPEND FPGA_RUNTIME_SRCS ${VTA_HW_PATH}/src/vmem/virtual_memory.cc ${VTA_HW_PATH}/src/vmem/virtual_memory.h)
     endif()
     # Target lib: vta
     add_library(vta SHARED ${FPGA_RUNTIME_SRCS})
@@ -123,6 +128,10 @@ elseif(PYTHON)
       target_include_directories(vta SYSTEM PUBLIC 3rdparty)
       target_include_directories(vta SYSTEM PUBLIC
         "/usr/local/intelFPGA_lite/18.1/embedded/ds-5/sw/gcc/arm-linux-gnueabihf/include")
+    elseif(${VTA_TARGET} STREQUAL "intelfocl")  # Intel OpenCL for FPGA rules
+      target_include_directories(vta PUBLIC 3rdparty)
+      set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++17")
+      target_link_libraries(vta -lOpenCL)
     endif()
   endif()
 

docker/Dockerfile.ci_qemu

@@ -62,6 +62,7 @@ RUN bash /install/ubuntu_install_qemu.sh
 
 # Zephyr SDK deps
 COPY install/ubuntu_install_zephyr.sh /install/ubuntu_install_zephyr.sh
+COPY install/ubuntu_init_zephyr_project.sh /install/ubuntu_init_zephyr_project.sh
 RUN bash /install/ubuntu_install_zephyr.sh
 ENV ZEPHYR_BASE=/opt/zephyrproject/zephyr
 

tests/scripts/task_ci_python_setup.sh → docker/install/ubuntu_init_zephyr_project.sh

@@ -16,18 +16,10 @@
 # specific language governing permissions and limitations
 # under the License.
 
-set -e
-set -u
-set -o pipefail
+DOWNLOAD_DIR=$1
+ZEPHYR_BRANCH=$2
 
-# Script to setup additional python env.
-#
-# Use the following command to install the
-# package to /workspace/.local, these additional
-# packages will have precedence over the system packages.
-#
-# command: python3 -m pip install --user <package>==<version>
-#
-echo "Addtiional setup in" ${CI_IMAGE_NAME}
-
-python3 -m pip install --user tlcpack-sphinx-addon==0.1.4 synr==0.3.0
+west init --mr ${ZEPHYR_BRANCH} ${DOWNLOAD_DIR}
+cd ${DOWNLOAD_DIR}
+west update
+west zephyr-export

docker/install/ubuntu_install_arm_compute_lib.sh

@@ -58,8 +58,8 @@ git clone "$repo_url" "$repo_dir"
 
 cd "$repo_dir"
 
-# pin version to v20.11
-git checkout 49b8f90
+# pin version to v21.02
+git checkout "v21.02"
 
 if [ "$architecture_type" != "aarch64" ]; then
   build_type="cross_compile"

docker/install/ubuntu_install_sbt.sh

@@ -25,7 +25,8 @@ set -o pipefail
 apt-get update && apt-get install -y apt-transport-https
 
 # Install the necessary dependencies for sbt
-echo "deb https://dl.bintray.com/sbt/debian /" | tee -a /etc/apt/sources.list.d/sbt.list
+echo "deb https://repo.scala-sbt.org/scalasbt/debian all main" | sudo tee /etc/apt/sources.list.d/sbt.list
+echo "deb https://repo.scala-sbt.org/scalasbt/debian /" | sudo tee /etc/apt/sources.list.d/sbt_old.list
 apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv 2EE0EA64E40A89B84B2DF73499E82A75642AC823
 
 # Note: The settings in vta/hardware/chisel/project/build.properties

docker/install/ubuntu_install_zephyr.sh

@@ -55,9 +55,11 @@ pip3 install west
 #EOF
 #chmod a+x /usr/local/bin/west
 
-west init --mr v2.5.0 /opt/zephyrproject
-cd /opt/zephyrproject
-west update
+# Init ZephyrProject
+ZEPHYR_PROJECT_PATH=/opt/zephyrproject
+ZEPHYR_INIT_SCRIPT=$(find -name "ubuntu_init_zephyr_project.sh")
+bash ${ZEPHYR_INIT_SCRIPT} ${ZEPHYR_PROJECT_PATH} v2.5-branch
+cd ${ZEPHYR_PROJECT_PATH}
 
 # This step is required because of the way docker/bash.sh works. It sets the user home directory to
 # /workspace (or the TVM root, anyhow), and this means that zephyr expects a ~/.cache directory to be
@@ -67,8 +69,6 @@ west update
 mkdir zephyr/.cache
 chmod o+rwx zephyr/.cache
 
-west zephyr-export
-
 #/opt/west/bin/pip3 install -r /opt/zephyrproject/zephyr/scripts/requirements.txt
 pip3 install -r /opt/zephyrproject/zephyr/scripts/requirements.txt
 

docs/conf.py

@@ -47,18 +47,31 @@
 # General information about the project.
 project = "tvm"
 author = "Apache Software Foundation"
-copyright = "2020, %s" % author
+copyright = "2020 - 2021, %s" % author
 github_doc_root = "https://github.com/apache/tvm/tree/main/docs/"
 
 os.environ["TVM_BUILD_DOC"] = "1"
+
+
+def git_describe_version(original_version):
+    """Get git describe version."""
+    ver_py = os.path.join(curr_path, "..", "version.py")
+    libver = {"__file__": ver_py}
+    exec(compile(open(ver_py, "rb").read(), ver_py, "exec"), libver, libver)
+    _, gd_version = libver["git_describe_version"]()
+    if gd_version != original_version:
+        print("Use git describe based version %s" % gd_version)
+    return gd_version
+
+
 # Version information.
 import tvm
 from tvm import topi
 from tvm import te
 from tvm import testing
 
-version = tvm.__version__
-release = tvm.__version__
+version = git_describe_version(tvm.__version__)
+release = version
 
 # Add any Sphinx extension module names here, as strings. They can be
 # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom ones
@@ -245,7 +258,6 @@
         "opt_gemm.py",
         "opt_conv_cuda.py",
         "opt_conv_tensorcore.py",
-        "opt_matmul_auto_tensorcore.py",
     ],
     "autotvm": [
         "tune_simple_template.py",

docs/deploy/arm_compute_lib.rst

@@ -178,7 +178,7 @@ An example configuration for `test_config.json`:
 
     {
       "connection_type": "local",
-      "host": "localhost",
+      "host": "127.0.0.1",
       "port": 9090,
       "target": "llvm -mtriple=aarch64-linux-gnu -mattr=+neon",
       "device_key": "",

docs/deploy/tensorrt.rst

@@ -181,6 +181,8 @@ Operator support
 +------------------------+------------------------------------+
 | nn.batch_norm          |                                    |
 +------------------------+------------------------------------+
+| nn.layer_norm          |                                    |
++------------------------+------------------------------------+
 | nn.softmax             |                                    |
 +------------------------+------------------------------------+
 | nn.conv2d              |                                    |
@@ -253,6 +255,8 @@ Operator support
 +------------------------+------------------------------------+
 | nn.adaptive_avg_pool2d |                                    |
 +------------------------+------------------------------------+
+| nn.batch_matmul        |                                    |
++------------------------+------------------------------------+
 | clip                   | Requires TensorRT 5.1.5 or greater |
 +------------------------+------------------------------------+
 | nn.leaky_relu          | Requires TensorRT 5.1.5 or greater |
@@ -277,6 +281,8 @@ Operator support
 +------------------------+------------------------------------+
 | nn.conv3d_transpose    | Requires TensorRT 6.0.1 or greater |
 +------------------------+------------------------------------+
+| erf                    | Requires TensorRT 7.0.0 or greater |
++------------------------+------------------------------------+
 
 
 Adding a new operator

docs/dev/index.rst

@@ -27,7 +27,7 @@ This page is organized as follows:
 - The `Logical Architecture Components`_ section describes the logical components.
   The sections after are specific guides focused on each logical component, organized
   by the component's name.
-- Feel free to also checkout the :ref:`dev-how-to` for useful development tips.
+- Feel free to also check out the :ref:`dev-how-to` for useful development tips.
 
 This guide provides a few complementary views of the architecture.
 First, we review a single end-to-end compilation flow and discuss the key data structures and the transformations.
@@ -42,7 +42,7 @@ In this guide, we will study an example compilation flow in the compiler. The fi
 
 - Import: The frontend component ingests a model into an IRModule, which contains a collection of functions that internally represent the model.
 - Transformation: The compiler transforms an IRModule to another functionally equivalent or approximately
-  equivalent(e.g. in the case of quantization) IRModule. Many of the transformatons are target (backend) independent.
+  equivalent(e.g. in the case of quantization) IRModule. Many of the transformations are target (backend) independent.
   We also allow target to affect the configuration of the transformation pipeline.
 - Target Translation: The compiler translates(codegen) the IRModule to an executable format specified by the target.
   The target translation result is encapsulated as a `runtime.Module` that can be exported, loaded, and executed on the target runtime environment.
@@ -103,7 +103,7 @@ Many low-level optimizations can be handled in the target phase by the LLVM, CUD
 Search-space and Learning-based Transformations
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-The transformation passes we described so far are deterministic and rule-based. One design goal of the TVM stack is to support high-performance code optimizations for different hardware platforms. To do so, we will need to investigate as many optimizations choices as possible, including but not limited to, multi-dimensional tensor access, loop tiling behavior, special accelerator memory hierarchy, and threading.
+The transformation passes we described so far are deterministic and rule-based. One design goal of the TVM stack is to support high-performance code optimizations for different hardware platforms. To do so, we will need to investigate as many optimization choices as possible, including but not limited to, multi-dimensional tensor access, loop tiling behavior, special accelerator memory hierarchy, and threading.
 
 It is hard to define a heuristic to make all of the choices. Instead, we will take a search and learning-based approach.
 We first define a collection of actions we can take to transform a program. Example actions include loop transformations, inlining,

docs/dev/pass_infra.rst

@@ -394,7 +394,7 @@ Python Frontend
 
 Only some simple APIs are needed for the frontend side. For example, we can
 provide users the following APIs to create and execute a pass (full
-implementation is provided in `python/tvm/relay/transform.py`_ and
+implementation is provided in `python/tvm/relay/transform/transform.py`_ and
 `python/tvm/ir/transform.py`_). The backend
 receives the information and decides which function it should use to create
 a Pass object.
@@ -460,7 +460,7 @@ users so that they can customize their own pass or pass pipeline.
 
 For all the passes that are implemented in the C++ backend, we provide
 corresponding Python APIs in `python/tvm/ir/transform.py`_ and
-`python/tvm/relay/transform.py`_, respectively. For instance,
+`python/tvm/relay/transform/transform.py`_, respectively. For instance,
 const folding has a Python API like the following:
 
 .. code:: python
@@ -538,7 +538,7 @@ optimization pipeline and debug Relay and tir passes, please refer to the
 
 .. _src/relay/pass/fold_constant.cc: https://github.com/apache/tvm/blob/main/src/relay/pass/fold_constant.cc
 
-.. _python/tvm/relay/transform.py: https://github.com/apache/tvm/blob/main/python/tvm/relay/transform.py
+.. _python/tvm/relay/transform/transform.py: https://github.com/apache/tvm/blob/main/python/tvm/relay/transform/transform.py
 
 .. _include/tvm/relay/transform.h: https://github.com/apache/tvm/blob/main/include/tvm/relay/transform.h