Merge pull request pytorch#452 from iotamudelta/ifu_20190812

IFU 20190812

.circleci/cimodel/data/pytorch_build_data.py

@@ -49,6 +49,7 @@
             ]),
             ("9.2", [X("3.6")]),
             ("10", [X("3.6")]),
+            ("10.1", [X("3.6")]),
         ]),
         ("android", [
             ("r19c", [XImportant("3.6")]),

.circleci/cimodel/data/pytorch_build_definitions.py

@@ -14,7 +14,7 @@
 
 DOCKER_IMAGE_PATH_BASE = "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/"
 
-DOCKER_IMAGE_VERSION = 323
+DOCKER_IMAGE_VERSION = 327
 
 
 @dataclass
@@ -222,7 +222,7 @@ def instantiate_configs():
             if compiler_name == "clang":
                 parms_list.append("asan")
 
-        if cuda_version in ["9.2", "10"]:
+        if cuda_version in ["9.2", "10", "10.1"]:
             # TODO The gcc version is orthogonal to CUDA version?
             parms_list.append("gcc7")
 

.circleci/scripts/binary_checkout.sh

@@ -41,8 +41,6 @@ popd
 # Clone the Builder master repo
 git clone -q https://github.com/pytorch/builder.git "$BUILDER_ROOT"
 pushd "$BUILDER_ROOT"
-git fetch origin
-git reset origin/master --hard
 echo "Using builder from "
 git --no-pager log --max-count 1
 popd

.circleci/scripts/binary_linux_test.sh

@@ -26,7 +26,7 @@ pkg="/final_pkgs/\$(ls /final_pkgs)"
 if [[ "$PACKAGE_TYPE" == conda ]]; then
   conda install -y "\$pkg" --offline
   if [[ "$DESIRED_CUDA" == 'cpu' ]]; then
-    conda install -y cpu-only -c pytorch
+    conda install -y cpuonly -c pytorch
   fi
   retry conda install -yq future numpy protobuf six
   if [[ "$DESIRED_CUDA" != 'cpu' ]]; then

.circleci/scripts/binary_linux_upload.sh

@@ -26,7 +26,7 @@ pushd /home/circleci/project/final_pkgs
 if [[ "$PACKAGE_TYPE" == conda ]]; then
   retry conda install -yq anaconda-client
   retry timeout 30 /home/circleci/project/login_to_anaconda.sh
-  anaconda upload "$(ls)" -u pytorch --label main --no-progress --force
+  anaconda upload "$(ls)" -u pytorch-nightly --label main --no-progress --force
 elif [[ "$PACKAGE_TYPE" == libtorch ]]; then
   retry pip install -q awscli
   s3_dir="s3://pytorch/libtorch/${PIP_UPLOAD_FOLDER}${DESIRED_CUDA}/"

.circleci/scripts/binary_macos_upload.sh

@@ -26,7 +26,7 @@ pushd "$workdir/final_pkgs"
 if [[ "$PACKAGE_TYPE" == conda ]]; then
   retry conda install -yq anaconda-client
   retry /Users/distiller/project/login_to_anaconda.sh
-  retry anaconda upload "$(ls)" -u pytorch --label main --no-progress --force
+  retry anaconda upload "$(ls)" -u pytorch-nightly --label main --no-progress --force
 elif [[ "$PACKAGE_TYPE" == libtorch ]]; then
   retry pip install -q awscli
   s3_dir="s3://pytorch/libtorch/${PIP_UPLOAD_FOLDER}${DESIRED_CUDA}/"

.circleci/scripts/binary_populate_env.sh

@@ -52,10 +52,10 @@ fi
 
 # We put this here so that OVERRIDE_PACKAGE_VERSION below can read from it
 export DATE="$(date -u +%Y%m%d)"
-if [[ "$(uname)" == 'Darwin' ]] || [[ "$DESIRED_CUDA" == "cu100" ]]; then
-  export PYTORCH_BUILD_VERSION="1.2.0.dev$DATE"
+if [[ "$(uname)" == 'Darwin' ]] || [[ "$DESIRED_CUDA" == "cu100" ]] || [[ "$PACKAGE_TYPE" == conda ]]; then
+  export PYTORCH_BUILD_VERSION="1.3.0.dev$DATE"
 else
-  export PYTORCH_BUILD_VERSION="1.2.0.dev$DATE+$DESIRED_CUDA"
+  export PYTORCH_BUILD_VERSION="1.3.0.dev$DATE+$DESIRED_CUDA"
 fi
 export PYTORCH_BUILD_NUMBER=1
 
@@ -72,12 +72,13 @@ export BUILD_PYTHONLESS="${BUILD_PYTHONLESS:-}"
 export DESIRED_DEVTOOLSET="$DESIRED_DEVTOOLSET"
 
 export DATE="$DATE"
-export NIGHTLIES_DATE_PREAMBLE=1.2.0.dev
+export NIGHTLIES_DATE_PREAMBLE=1.3.0.dev
 export PYTORCH_BUILD_VERSION="$PYTORCH_BUILD_VERSION"
 export PYTORCH_BUILD_NUMBER="$PYTORCH_BUILD_NUMBER"
 export OVERRIDE_PACKAGE_VERSION="$PYTORCH_BUILD_VERSION"
 
-export TORCH_PACKAGE_NAME='torch-nightly'
+# TODO: We don't need this anymore IIUC
+export TORCH_PACKAGE_NAME='torch'
 export TORCH_CONDA_BUILD_FOLDER='pytorch-nightly'
 
 export USE_FBGEMM=1

.circleci/verbatim-sources/job-specs-custom.yml

@@ -1,7 +1,7 @@
   pytorch_short_perf_test_gpu:
     environment:
       BUILD_ENVIRONMENT: pytorch-short-perf-test-gpu
-      DOCKER_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9-cudnn7-py3:323"
+      DOCKER_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9-cudnn7-py3:327"
       PYTHON_VERSION: "3.6"
       USE_CUDA_DOCKER_RUNTIME: "1"
     resource_class: gpu.medium
@@ -42,7 +42,7 @@
     environment:
       BUILD_ENVIRONMENT: pytorch-python-doc-push
       # TODO: stop hardcoding this
-      DOCKER_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9-cudnn7-py3:323"
+      DOCKER_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9-cudnn7-py3:327"
     resource_class: large
     machine:
       image: ubuntu-1604:201903-01
@@ -92,7 +92,7 @@
   pytorch_cpp_doc_push:
     environment:
       BUILD_ENVIRONMENT: pytorch-cpp-doc-push
-      DOCKER_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9-cudnn7-py3:323"
+      DOCKER_IMAGE: "308535385114.dkr.ecr.us-east-1.amazonaws.com/pytorch/pytorch-linux-xenial-cuda9-cudnn7-py3:327"
     resource_class: large
     machine:
       image: ubuntu-1604:201903-01

CMakeLists.txt

@@ -10,6 +10,12 @@ cmake_policy(SET CMP0025 NEW)
 # ---[ Project and semantic versioning.
 project(Caffe2 CXX C)
 
+if (${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
+  set(LINUX TRUE)
+else()
+  set(LINUX FALSE)
+endif()
+
 set(CMAKE_INSTALL_MESSAGE NEVER)
 
 set(CMAKE_CXX_STANDARD 11)
@@ -121,7 +127,9 @@ cmake_dependent_option(
     "USE_NCCL" OFF)
 option(USE_NNAPI "Use NNAPI" OFF)
 option(USE_NNPACK "Use NNPACK" ON)
-option(USE_NUMA "Use NUMA (only available on Linux)" ON)
+cmake_dependent_option(
+    USE_NUMA "Use NUMA. Only available on Linux." ON
+    "LINUX" OFF)
 cmake_dependent_option(
     USE_NVRTC "Use NVRTC. Only available if USE_CUDA is on." OFF
     "USE_CUDA" OFF)

aten/src/ATen/ATen.h

@@ -17,13 +17,13 @@
 #include <ATen/Tensor.h>
 #include <ATen/TensorGeometry.h>
 #include <ATen/TensorOperators.h>
+#include <ATen/TensorOptions.h>
 #include <ATen/Version.h>
 #include <ATen/core/ATenGeneral.h>
 #include <ATen/core/Generator.h>
 #include <c10/core/Layout.h>
 #include <ATen/core/Scalar.h>
 #include <c10/core/Storage.h>
-#include <c10/core/TensorOptions.h>
 #include <ATen/core/Reduction.h>
 #include <c10/util/Exception.h>
 #include <ATen/core/ATenDispatch.h>

aten/src/ATen/Context.cpp

@@ -2,7 +2,7 @@
 
 #include <ATen/Context.h>
 
-#include <c10/core/TensorOptions.h>
+#include <ATen/core/TensorOptions.h>
 
 #include <thread>
 #include <mutex>

aten/src/ATen/Declarations.cwrap

@@ -1768,40 +1768,6 @@
         - THTensor* self
         - THTensor* self
 ]]
-[[
-  name: _th_atan2
-  types:
-    - floating_point
-  backends:
-    - CPU
-    - CUDA
-  variants:
-    - function
-  cname: atan2
-  return: argument 0
-  arguments:
-    - arg: THTensor* result
-      output: True
-    - arg: THTensor* self
-      broadcast: other fallback
-    - THTensor* other
-]]
-[[
-  name: _th_atan2_
-  types:
-    - floating_point
-  backends:
-    - CPU
-    - CUDA
-  cname: atan2
-  variants: function
-  return: argument 0
-  arguments:
-    - THTensor* self
-    - arg: THTensor* self
-      broadcast: other fallback inplace
-    - THTensor* other
-]]
 [[
   name: _th_pow
   cname: pow
@@ -2355,75 +2321,6 @@
     - THTensor* batch1
     - THTensor* batch2
 ]]
-[[
-  name: _th_addcmul
-  cname: addcmul
-  variants:
-    - function
-  backends:
-    - CUDA
-  return: argument 0
-  arguments:
-    - arg: THTensor* result
-      output: True
-    - arg: THTensor* self
-      broadcast: tensor1,tensor2 fallback
-    - arg: real value
-      default: AS_REAL(1)
-      kwarg_only: True
-    - THTensor* tensor1
-    - THTensor* tensor2
-]]
-[[
-  name: _th_addcmul_
-  options:
-    - cname: addcmul
-      variants: function
-      backends:
-        - CUDA
-      return: argument 0
-      arguments:
-        - THTensor* self
-        - arg: THTensor* self
-          broadcast: tensor1,tensor2 inplace fallback
-        - arg: real value
-          default: AS_REAL(1)
-          kwarg_only: True
-        - THTensor* tensor1
-        - THTensor* tensor2
-]]
-[[
-  name: _th_addcdiv
-  cname: addcdiv
-  variants:
-    - function
-  return: argument 0
-  arguments:
-    - arg: THTensor* result
-      output: True
-    - arg: THTensor* self
-      broadcast: tensor1,tensor2 fallback
-    - arg: real value
-      default: AS_REAL(1)
-      kwarg_only: True
-    - THTensor* tensor1
-    - THTensor* tensor2
-]]
-[[
-  name: _th_addcdiv_
-  cname: addcdiv
-  variants: function
-  return: argument 0
-  arguments:
-    - THTensor* self
-    - arg: THTensor* self
-      broadcast: tensor1,tensor2 inplace fallback
-    - arg: real value
-      default: AS_REAL(1)
-      kwarg_only: True
-    - THTensor* tensor1
-    - THTensor* tensor2
-]]
 [[
   name: _th_gels
   cname: gels

aten/src/ATen/Dispatch.h

@@ -33,7 +33,7 @@ struct ScalarTypeToCType<at::ScalarType::Half> {
   // due to ambiguous reference which can't to be resolved. For some reason it cant pick between at::detail and at::cuda::detail.
   // For repro example, please see: https://gist.github.com/izdeby/952ae7cf256ddb740a73776d39a7e7ba
   // TODO: remove once the bug is fixed.
-  static at::Half t;
+  static type t;
 };
 
 template<>
@@ -44,7 +44,7 @@ struct ScalarTypeToCType<at::ScalarType::BFloat16> {
   // due to ambiguous reference which can't to be resolved. For some reason it cant pick between at::detail and at::cuda::detail.
   // For repro example, please see: https://gist.github.com/izdeby/952ae7cf256ddb740a73776d39a7e7ba
   // TODO: remove once the bug is fixed.
-  static at::BFloat16 t;
+  static type t;
 };
 
 template<>
@@ -55,7 +55,7 @@ struct ScalarTypeToCType<at::ScalarType::Bool> {
   // due to ambiguous reference which can't to be resolved. For some reason it cant pick between at::detail and at::cuda::detail.
   // For repro example, please see: https://gist.github.com/izdeby/952ae7cf256ddb740a73776d39a7e7ba
   // TODO: remove once the bug is fixed.
-  static bool t;
+  static type t;
 };
 
 template<>

aten/src/ATen/InitialTensorOptions.h

@@ -1,6 +1,6 @@
 #pragma once
 
-#include <c10/core/TensorOptions.h>
+#include <ATen/core/TensorOptions.h>
 
 namespace at {
 

aten/src/ATen/TensorOptions.h

@@ -1,2 +1,2 @@
 #pragma once
-#include <c10/core/TensorOptions.h>
+#include <ATen/core/TensorOptions.h>

aten/src/ATen/core/DeprecatedTypeProperties.h

@@ -3,7 +3,7 @@
 #include <c10/core/Backend.h>
 #include <c10/core/ScalarType.h>
 #include <c10/core/Layout.h>
-#include <c10/core/TensorOptions.h>
+#include <ATen/core/TensorOptions.h>
 #include <c10/core/Storage.h>
 #include <ATen/core/DeprecatedTypePropertiesRegistry.h>
 #include <ATen/core/Generator.h>