Skip to content

Commit

Permalink
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into…
Browse files Browse the repository at this point in the history 
… pten/auot_gen_include_headers
  • Loading branch information
@chenwhql
chenwhql committed Jan 21, 2022
2 parents c1c6a97 + a14dc68 commit d6d5185
Show file tree
Hide file tree
Showing 241 changed files with 4,315 additions and 2,581 deletions.
61 changes: 39 additions & 22 deletions cmake/pten.cmake
View file
Original file line number Diff line number Diff line change
Expand Up @@ -145,38 +145,55 @@ function(kernel_library TARGET)
list(LENGTH gpu_srcs gpu_srcs_len)
list(LENGTH xpu_srcs xpu_srcs_len)

if (${common_srcs_len} GREATER 0)
# If the kernel has a device independent public implementation,
# we will use this implementation and will not adopt the implementation
# under specific devices
# Build Target according different src organization
if((${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0 OR
${xpu_srcs_len} GREATER 0) AND ${common_srcs_len} GREATER 0)
# If the common_srcs depends on specific device srcs, build target using this rule.
if (WITH_GPU)
if (${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0)
nv_library(${TARGET}_part SRCS ${cpu_srcs} ${gpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
nv_library(${TARGET} SRCS ${common_srcs} DEPS ${TARGET}_part)
endif()
elseif (WITH_ROCM)
if (${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0)
hip_library(${TARGET}_part SRCS ${cpu_srcs} ${gpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
hip_library(${TARGET} SRCS ${common_srcs} DEPS ${TARGET}_part)
endif()
else()
if (${cpu_srcs_len} GREATER 0 OR ${xpu_srcs_len} GREATER 0)
cc_library(${TARGET}_part SRCS ${cpu_srcs} ${xpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
cc_library(${TARGET} SRCS ${common_srcs} DEPS ${TARGET}_part)
endif()
endif()
elseif (${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0 OR ${xpu_srcs_len} GREATER 0)
if (WITH_GPU)
nv_library(${TARGET} SRCS ${common_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
if (${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0)
nv_library(${TARGET} SRCS ${cpu_srcs} ${gpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
endif()
elseif (WITH_ROCM)
hip_library(${TARGET} SRCS ${common_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
if (${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0)
hip_library(${TARGET} SRCS ${cpu_srcs} ${gpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
endif()
else()
cc_library(${TARGET} SRCS ${common_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
if (${cpu_srcs_len} GREATER 0 OR ${xpu_srcs_len} GREATER 0)
cc_library(${TARGET} SRCS ${cpu_srcs} ${xpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
endif()
endif()
else()
# If the kernel has a header file declaration, but no corresponding
# implementation can be found, this is not allowed
if (${cpu_srcs_len} EQUAL 0 AND ${gpu_srcs_len} EQUAL 0 AND
${xpu_srcs_len} EQUAL 0)
message(FATAL_ERROR "Cannot find any implementation for ${TARGET}")
if (${common_srcs_len} EQUAL 0)
message(FATAL_ERROR "Cannot find any implementation for ${TARGET}")
else()
# If the kernel has a device independent public implementation,
# we will use this implementation and will not adopt the implementation
# under specific devices
if (WITH_GPU)
if (${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0)
nv_library(${TARGET} SRCS ${cpu_srcs} ${gpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
endif()
nv_library(${TARGET} SRCS ${common_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
elseif (WITH_ROCM)
if (${cpu_srcs_len} GREATER 0 OR ${gpu_srcs_len} GREATER 0)
hip_library(${TARGET} SRCS ${cpu_srcs} ${gpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
endif()
hip_library(${TARGET} SRCS ${common_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
else()
if (${cpu_srcs_len} GREATER 0 OR ${xpu_srcs_len} GREATER 0)
cc_library(${TARGET} SRCS ${cpu_srcs} ${xpu_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
endif()
cc_library(${TARGET} SRCS ${common_srcs} DEPS ${kernel_library_DEPS} ${kernel_deps})
endif()
endif()
endif()
endif()

if (${common_srcs_len} GREATER 0 OR ${cpu_srcs_len} GREATER 0 OR
Expand Down
2 changes: 1 addition & 1 deletion paddle/fluid/distributed/service/brpc_utils.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ limitations under the License. */
#include "paddle/fluid/framework/data_type.h"
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/scope.h"
#include "paddle/fluid/framework/selected_rows.h"
#include "paddle/fluid/framework/selected_rows_utils.h"
#include "paddle/fluid/framework/tensor_util.h"
#include "paddle/fluid/framework/var_type.h"
#include "paddle/fluid/platform/port.h"
Expand Down
2 changes: 1 addition & 1 deletion paddle/fluid/distributed/table/depends/large_scale_kv.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,7 @@
#include "paddle/fluid/framework/generator.h"
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/rw_lock.h"
#include "paddle/fluid/framework/selected_rows.h"
#include "paddle/fluid/framework/selected_rows_utils.h"
#include "paddle/fluid/framework/tensor.h"
#include "paddle/fluid/framework/threadpool.h"
#include "paddle/fluid/framework/variable.h"
Expand Down
2 changes: 1 addition & 1 deletion paddle/fluid/eager/legacy/tensor_helper.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,7 @@
#include "paddle/fluid/framework/lod_tensor_array.h"
#include "paddle/fluid/framework/reader.h"
#include "paddle/fluid/framework/scope.h"
#include "paddle/fluid/framework/selected_rows.h"
#include "paddle/fluid/framework/selected_rows_utils.h"
#include "paddle/fluid/framework/var_type_traits.h"
#include "paddle/fluid/platform/place.h"

Expand Down
24 changes: 8 additions & 16 deletions paddle/fluid/framework/CMakeLists.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -55,14 +55,6 @@ proto_library(trainer_desc_proto SRCS trainer_desc.proto DEPS framework_proto

cc_library(string_array SRCS string_array.cc DEPS utf8proc)

cc_library(ddim SRCS ddim.cc DEPS eigen3 boost enforce)
cc_test(ddim_test SRCS ddim_test.cc DEPS ddim)
if(WITH_GPU)
nv_test(dim_test SRCS dim_test.cu DEPS ddim)
elseif(WITH_ROCM)
hip_test(dim_test SRCS dim_test.cu DEPS ddim)
endif()
cc_test(unroll_array_ops_test SRCS unroll_array_ops_test.cc)
cc_library(data_type SRCS data_type.cc DEPS framework_proto ddim device_context)
cc_test(data_type_test SRCS data_type_test.cc DEPS data_type place tensor)
if(WITH_GPU)
Expand Down Expand Up @@ -102,7 +94,7 @@ else()
endif()
cc_library(lod_tensor SRCS lod_tensor.cc DEPS ddim mixed_vector place tensor framework_proto version)

cc_test(lod_tensor_test SRCS lod_tensor_test.cc DEPS lod_tensor memory)
cc_test(lod_tensor_test SRCS lod_tensor_test.cc DEPS lod_utils lod_tensor memory)

if(WITH_GPU)
nv_test(lod_tensor_gpu_test SRCS lod_tensor_test.cu DEPS lod_tensor)
Expand All @@ -118,7 +110,7 @@ cc_test(reader_test SRCS reader_test.cc DEPS reader)
cc_library(threadpool SRCS threadpool.cc DEPS enforce)
cc_test(threadpool_test SRCS threadpool_test.cc DEPS threadpool)

cc_library(var_type_traits SRCS var_type_traits.cc DEPS lod_tensor selected_rows framework_proto scope)
cc_library(var_type_traits SRCS var_type_traits.cc DEPS lod_tensor selected_rows_utils framework_proto scope)
if (WITH_GPU)
target_link_libraries(var_type_traits dynload_cuda)
endif()
Expand Down Expand Up @@ -172,7 +164,7 @@ cc_library(data_layout_transform SRCS data_layout_transform.cc DEPS tensor math_
cc_test(data_layout_transform_test SRCS data_layout_transform_test.cc DEPS data_layout_transform)

cc_library(data_transform SRCS data_transform.cc DEPS math_function tensor
framework_proto selected_rows data_device_transform data_type_transform data_layout_transform)
framework_proto selected_rows_utils data_device_transform data_type_transform data_layout_transform)

cc_library(attribute SRCS attribute.cc DEPS framework_proto boost enforce)
cc_test(program_desc_test SRCS program_desc_test.cc DEPS proto_desc
Expand Down Expand Up @@ -390,8 +382,8 @@ cc_library(prune SRCS prune.cc DEPS framework_proto boost)
cc_test(prune_test SRCS prune_test.cc DEPS op_info prune recurrent_op device_context)
cc_test(var_type_inference_test SRCS var_type_inference_test.cc DEPS op_registry
proto_desc)
cc_library(selected_rows SRCS selected_rows.cc DEPS tensor)
cc_test(selected_rows_test SRCS selected_rows_test.cc DEPS selected_rows)
cc_library(selected_rows_utils SRCS selected_rows_utils.cc DEPS tensor)
cc_test(selected_rows_utils_test SRCS selected_rows_utils_test.cc DEPS selected_rows_utils)

cc_test(op_kernel_type_test SRCS op_kernel_type_test.cc DEPS place device_context framework_proto op_kernel_type)
cc_test(cow_ptr_tests SRCS details/cow_ptr_test.cc)
Expand All @@ -414,7 +406,7 @@ cc_library(save_load_util SRCS save_load_util.cc DEPS tensor scope layer)
cc_test(save_load_util_test SRCS save_load_util_test.cc DEPS save_load_util tensor scope layer)
cc_library(generator SRCS generator.cc DEPS enforce place)

cc_library(pten_utils SRCS pten_utils.cc DEPS lod_tensor selected_rows place pten var_type_traits pten_api_utils op_info)
cc_library(pten_utils SRCS pten_utils.cc DEPS lod_tensor selected_rows_utils place pten var_type_traits pten_api_utils op_info)

# Get the current working branch
execute_process(
Expand Down Expand Up @@ -446,8 +438,8 @@ set(FLUID_FRAMEWORK_MODULES proto_desc memory lod_tensor executor data_feed_prot

cc_library(paddle_framework DEPS ${FLUID_FRAMEWORK_MODULES})

if(WITH_TESTING AND TEST selected_rows_test)
set_tests_properties(selected_rows_test PROPERTIES TIMEOUT 120)
if(WITH_TESTING AND TEST selected_rows_utils_test)
set_tests_properties(selected_rows_utils_test PROPERTIES TIMEOUT 120)
endif()

cc_test(scope_guard_test SRCS scope_guard_test.cc)
Expand Down
2 changes: 1 addition & 1 deletion paddle/fluid/framework/data_transform.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ limitations under the License. */
#include <vector>

#include "paddle/fluid/framework/op_kernel_type.h"
#include "paddle/fluid/framework/selected_rows.h"
#include "paddle/fluid/framework/selected_rows_utils.h"
#include "paddle/fluid/framework/tensor.h"
#include "paddle/fluid/framework/variable.h"
#include "paddle/fluid/operators/math/math_function.h"
Expand Down
2 changes: 1 addition & 1 deletion paddle/fluid/framework/data_type_transform.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@ limitations under the License. */

#include "paddle/fluid/framework/data_type_transform.h"

#include "paddle/fluid/framework/selected_rows.h"
#include "paddle/fluid/framework/selected_rows_utils.h"
#include "paddle/fluid/platform/transform.h"

namespace paddle {
Expand Down
Loading

1 comment on commit d6d5185

@paddle-bot-old
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

✅Congratulation! Your pull request passed all required CI. You could ask reviewer(s) to approve and merge. 🎉

Please sign in to comment.