Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
58 commits
Select commit Hold shift + click to select a range
75227c9
use group id to differentiate keys for tcp store (#41496)
Apr 7, 2022
edbb398
Switch some dy2st UT to eager mode (#41382)
0x45f Apr 7, 2022
5516f18
[Phi] Add unbind yaml and final state api (#41277)
chenwhql Apr 7, 2022
c77a263
Add yaml for matrix rank op (#41466)
From00 Apr 7, 2022
8fb8fa4
[Eager] Fix tensor type (#41468)
veyron95 Apr 7, 2022
90cb337
[Phi]Add hard_swish/kron/linspace/logit yaml file (#41298)
YuanRisheng Apr 7, 2022
50ddc0b
Add dist norm yamls (#41424)
phlrain Apr 7, 2022
0d642d3
add norm, segment_pool (#41465)
phlrain Apr 7, 2022
9714878
remove FLAGS_use_curand and change all random op CUDA implementation …
zhwesky2010 Apr 7, 2022
b3bcebb
[GPUPS] bind afs wrpper (#41227)
Thunderbrook Apr 7, 2022
9844aaf
[Phi] Add swish yaml and final state api (#41479)
chenwhql Apr 8, 2022
bc88fbb
Add conj pixel shuffle yaml (#41499)
phlrain Apr 8, 2022
0cd577c
pybind support CustomPlace (#41136)
ronny1996 Apr 8, 2022
1ed1a97
Fix cv2 import error and some issues for lamb (#41500)
sneaxiy Apr 8, 2022
770ce7c
xpu mul unittest *test=kunlun (#41140)
taixiurong Apr 8, 2022
14dba63
[ROCm] fix dcu error in device event base, test=develop (#41521)
qili93 Apr 8, 2022
f43af27
Refine statistic table (#41524)
rainyfly Apr 8, 2022
acc25d0
tensor fluid code transfer part1 (#41094)
zhiboniu Apr 8, 2022
70036d5
Fix libmct.cmake tar ownership change (#41516)
ZHUI Apr 8, 2022
0a6fe69
[Eager]Fix segment_pool/allclose/isclose/scale API bug (#41506)
Aurelius84 Apr 8, 2022
d4710df
modify unittest of lstm forward, *test=kunlun (#41534)
helen88 Apr 8, 2022
09203e4
Fix RNN OP multi-threads predict bug (#41529)
joey12300 Apr 8, 2022
33abfbe
[Eager] Remove elementwise add in conv (#41515)
chenwhql Apr 8, 2022
fcccb3f
fix test_autograd_functional_dynamic random timeout (#41457)
cxxly Apr 8, 2022
ab137a8
update (#41309)
Apr 8, 2022
c2e1294
fix running error for ipu (#41481)
gglin001 Apr 8, 2022
04a4bdf
fix group_norm (#41531)
Zjq9409 Apr 8, 2022
330582e
Fix fake quant cuda kernel (#41305)
wanghaoshuang Apr 8, 2022
b3e7973
[fleet executor] Add sink interceptor and test (#41497)
LiYuRio Apr 8, 2022
f581f5b
[new-exec] fix bug that no thread is waked up when adding task to thr…
zhiqiu Apr 9, 2022
0e048fc
fix cross entropy (#41541)
sneaxiy Apr 9, 2022
be11648
fix pylayer mem leak, test=develop (#41559)
wanghuancoder Apr 9, 2022
9cb2287
fix_ci_problem3 (#41484)
JiabinYang Apr 9, 2022
e1792a3
Add get profiler from config (#41532)
rainyfly Apr 9, 2022
b937cdc
Autotune the workspace_size_limit in conv. (#40338)
JamesLim-sy Apr 9, 2022
96ced1a
[infrt] opt support input valid places by commondline. (#41544)
winter-wang Apr 9, 2022
b3b8d34
add depthwise conv hip support (#41537)
phlrain Apr 9, 2022
9872da0
[Eager] Support allclose and linalg_cond to eager mode (#41545)
veyron95 Apr 9, 2022
ff2fba3
modify the block size of the group_norm backward (#41570)
Zjq9409 Apr 9, 2022
7a07c4a
Unittest recover (#41431)
zhaocaibei123 Apr 9, 2022
e68da18
add mkldnn int8 pass [step1] (#41579)
baoachun Apr 10, 2022
a78ca1c
predictor support trt (#41556)
jiweibo Apr 10, 2022
c00d869
add mkldnn compute_propagate_scales int8 pass (#41592)
baoachun Apr 10, 2022
81c4072
[Yaml] Modify api and add unittests for full api final state. (#41437)
2742195759 Apr 10, 2022
91d6f47
fix warpctc grad kernel dep eror (#41598)
chenwhql Apr 10, 2022
c1394c6
[KP]fix bug when TruncatedNormal cannot fall back in cpu (#41565)
Liu-xiandong Apr 10, 2022
795d712
fix some ops (#41577)
sneaxiy Apr 11, 2022
9107dc6
Switch test_transformer to eager mode and fix roll error (#41548)
0x45f Apr 11, 2022
cd2a4cd
[Yaml] add yaml for Uniform random and add unit test. (#41517)
2742195759 Apr 11, 2022
437bebd
[Yaml] Add assign yaml (#41428)
MingMingShangTian Apr 11, 2022
89bfa96
Modify op-benchamrk script (#41470)
ZzSean Apr 11, 2022
36d7684
[Phi]Add multi_dot/maxout/multiplex op yaml (#41550)
YuanRisheng Apr 11, 2022
9287d5a
Add no need buffer config (#41605)
phlrain Apr 11, 2022
368f1dd
fix arg_max for int type, *test=kunlun (#41522)
ykkk2333 Apr 11, 2022
8fc9c41
fix for gaussian random (#41572)
jakpiase Apr 11, 2022
535810b
update lite compile cmake (#41512)
shentanyue Apr 11, 2022
fc621df
support more ops (#41421)
gglin001 Apr 11, 2022
c64d9a4
add backend for heter training (#41526)
Apr 11, 2022
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion cmake/external/libmct.cmake
Original file line number Diff line number Diff line change
Expand Up @@ -45,7 +45,7 @@ ExternalProject_Add(
PREFIX ${LIBMCT_PREFIX_DIR}
DOWNLOAD_DIR ${LIBMCT_DOWNLOAD_DIR}
DOWNLOAD_COMMAND wget --no-check-certificate ${LIBMCT_URL} -c -q -O ${LIBMCT_NAME}.tar.gz
&& tar zxvf ${LIBMCT_NAME}.tar.gz
&& tar --no-same-owner -zxvf ${LIBMCT_NAME}.tar.gz
DOWNLOAD_NO_PROGRESS 1
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_INSTALL_PREFIX=${LIBMCT_INSTALL_ROOT}
Expand Down
2 changes: 1 addition & 1 deletion cmake/external/lite.cmake
Original file line number Diff line number Diff line change
Expand Up @@ -86,7 +86,7 @@ if (NOT LITE_SOURCE_DIR OR NOT LITE_BINARY_DIR)
GIT_REPOSITORY "${GIT_URL}/PaddlePaddle/Paddle-Lite.git"
GIT_TAG ${LITE_GIT_TAG}
PREFIX ${LITE_PREFIX_DIR}
PATCH_COMMAND mkdir -p ${LITE_PREFIX_DIR}/src/extern_lite-build/lite/gen_code && touch ${LITE_PREFIX_DIR}/src/extern_lite-build/lite/gen_code/__generated_code__.cc && sed -i "/aarch64-linux-gnu-gcc/d" ${LITE_PREFIX_DIR}/src/extern_lite/cmake/cross_compiling/armlinux.cmake && sed -i "/aarch64-linux-gnu-g++/d" ${LITE_PREFIX_DIR}/src/extern_lite/cmake/cross_compiling/armlinux.cmake
PATCH_COMMAND mkdir -p ${LITE_PREFIX_DIR}/src/extern_lite-build/lite/gen_code && touch ${LITE_PREFIX_DIR}/src/extern_lite-build/lite/gen_code/__generated_code__.cc && sed -i "/aarch64-linux-gnu-gcc/d" ${LITE_PREFIX_DIR}/src/extern_lite/cmake/os/armlinux.cmake && sed -i "/aarch64-linux-gnu-g++/d" ${LITE_PREFIX_DIR}/src/extern_lite/cmake/os/armlinux.cmake
UPDATE_COMMAND ""
BUILD_COMMAND ${LITE_BUILD_COMMAND}
INSTALL_COMMAND ""
Expand Down
15 changes: 15 additions & 0 deletions cmake/external/poplar.cmake
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,19 @@
# See the License for the specific language governing permissions and
# limitations under the License.

macro(find_popart_version popart_version_file)
file(READ ${popart_version_file} popart_version_file_content)
string(REGEX MATCH "(POPART_VERSION_STRING)[ \t\r\n](\")([0-9]+\.[0-9]+\.[0-9]+)(\\+)([A-Za-z0-9_]*)(\")" POPART_VERSION ${popart_version_file_content})
string(REPLACE "POPART_VERSION_STRING" "" POPART_VERSION "${POPART_VERSION}")
string(REPLACE "\"" "" POPART_VERSION "${POPART_VERSION}")
string(REPLACE " " "" POPART_VERSION "${POPART_VERSION}")
if(NOT POPART_VERSION)
set(POPART_VERSION "Unknown version")
else()
message(STATUS "Current PopART version is ${POPART_VERSION}")
endif()
endmacro()

if(WITH_IPU)
set(POPLAR_DIR CACHE PATH "Path to a Poplar install")
set(POPART_DIR CACHE PATH "Path to a Popart install")
Expand Down Expand Up @@ -64,6 +77,8 @@ if(WITH_IPU)
message(FATAL_ERROR "You must provide a path to a Popart build using -DPOPART_DIR=/path/to/popart/build")
endif()

find_popart_version("${POPART_DIR}/include/popart/version.hpp")

add_definitions(-DONNX_NAMESPACE=onnx)
add_custom_target(extern_poplar DEPENDS poplar popart-only)
endif()
2 changes: 1 addition & 1 deletion cmake/external/xpu.cmake
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ ENDIF()

if(NOT DEFINED XPU_BASE_URL)
SET(XPU_BASE_URL_WITHOUT_DATE "https://baidu-kunlun-product.cdn.bcebos.com/KL-SDK/klsdk-dev")
SET(XPU_BASE_URL "${XPU_BASE_URL_WITHOUT_DATE}/20220402")
SET(XPU_BASE_URL "${XPU_BASE_URL_WITHOUT_DATE}/20220408")
else()
SET(XPU_BASE_URL "${XPU_BASE_URL}")
endif()
Expand Down
7 changes: 6 additions & 1 deletion cmake/inference_lib.cmake
Original file line number Diff line number Diff line change
Expand Up @@ -398,7 +398,8 @@ function(version version_file)
"WITH_GPU: ${WITH_GPU}\n"
"WITH_ROCM: ${WITH_ROCM}\n"
"WITH_ASCEND_CL: ${WITH_ASCEND_CL}\n"
"WITH_ASCEND_CXX11: ${WITH_ASCEND_CXX11}\n")
"WITH_ASCEND_CXX11: ${WITH_ASCEND_CXX11}\n"
"WITH_IPU: ${WITH_IPU}\n")
if(WITH_GPU)
file(APPEND ${version_file}
"CUDA version: ${CUDA_VERSION}\n"
Expand All @@ -414,6 +415,10 @@ function(version version_file)
"Ascend Toolkit version: ${ASCEND_TOOLKIT_VERSION}\n"
"Ascend Driver version: ${ASCEND_DRIVER_VERSION}\n")
endif()
if(WITH_IPU)
file(APPEND ${version_file}
"PopART version: ${POPART_VERSION}\n")
endif()
file(APPEND ${version_file} "CXX compiler version: ${CMAKE_CXX_COMPILER_VERSION}\n")
if(TENSORRT_FOUND)
file(APPEND ${version_file}
Expand Down
6 changes: 4 additions & 2 deletions paddle/fluid/distributed/collective/ProcessGroupNCCL.cc
Original file line number Diff line number Diff line change
Expand Up @@ -110,15 +110,17 @@ void ProcessGroupNCCL::BroadcastUniqueNCCLID(
std::vector<ncclUniqueId>& nccl_ids) { // NOLINT
if (rank_ == 0) {
for (size_t i = 0; i < nccl_ids.size(); i++) {
auto key = "ProcessGroupNCCL/nccl_ids/" + std::to_string(i);
auto key = "ProcessGroupNCCL/nccl_ids/" + std::to_string(gid_) + "/" +
std::to_string(i);
auto nccl_id = std::vector<uint8_t>(
reinterpret_cast<uint8_t*>(&nccl_ids[i]),
reinterpret_cast<uint8_t*>(&nccl_ids[i]) + NCCL_UNIQUE_ID_BYTES);
store_->set(key, nccl_id);
}
} else {
for (size_t i = 0; i < nccl_ids.size(); i++) {
auto key = "ProcessGroupNCCL/nccl_ids/" + std::to_string(i);
auto key = "ProcessGroupNCCL/nccl_ids/" + std::to_string(gid_) + "/" +
std::to_string(i);
auto ret = store_->get(key);
std::memcpy(&nccl_ids[i], ret.data(), ret.size());
}
Expand Down
Loading