ttudu
dac6f5425d
delete GPU/CPUDynamicKernel
4 years ago
zhoufeng
f49b195c39
extract common as an independent shared library
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
4 years ago
i-robot
6c8f942737
!30544 fix codecheck
Merge pull request !30544 from xulei/fix_codecheck
4 years ago
i-robot
c4843c4085
!29958 upgrade ascend 20220211
Merge pull request !29958 from shenwei41/upgrade_ascend_20220211
4 years ago
shenwei41
ff75314719
upgrade ascend 20220211
4 years ago
z00512249
0de0e0cd4b
fix cpu matrix_set_diag && matrix_band_part kernel codex && pclint-plus
4 years ago
xulei
4cf320cdbd
fix codecheck
4 years ago
kswang
f26870d437
clean code
4 years ago
ttudu
451ebd1bd1
dynamic_shape
4 years ago
marui
1d529f2ff3
Add OWNERS files for ccsrc subdirectories
4 years ago
marui
d47e79b04c
Refactor ccsrc directories and CMakeLists files
4 years ago
liubuyu
43c79eb853
mindspore path adjust
5 years ago
mindspore-ci-bot
99a6033aa2
!3032 Refactor the akg op registers
Merge pull request !3032 from DeshiChen/0711_akg_op_register_master
5 years ago
VectorSL
d22a597689
gpu fix addn bug and supported list bug
5 years ago
dayschan
78f35814d5
Refactor the akg op registers.
Since akg supports both Ascend and Gpu, but their supported type and format are different,
so we use two directory "ascend" and "gpu" to store their registers respectively, and use
an attribute "processor" to distinguish them.
Main changes:
1) Add two op register class "AkgAscendRegOp" and "AkgGpuRegOp", inherited from the original AkgRegOp.
2) Rewrite akg ascend op registers with new interface, move them into directory "ascend".
3) Rename the imply_type from "AutoDiff" to "AKG".
4) Modify function FindOp, check the processor when imply_type is "AKG".
5) Modify function CheckRepetition, remove the judgement for impl_path, check processor instead.
TODO: Remove op registers in akg root path.
5 years ago
mindspore-ci-bot
f65586cefa
!3035 Refactoring for base and abstract types.
Merge pull request !3035 from ZhangQinghua/master
5 years ago
Zhang Qinghua
87715ed127
Refactoring for base and abstract types: Moving Base into base folder; Splitting the abstract type and infer&specialize routines;
5 years ago
mindspore-ci-bot
be9b3c53dc
!3027 Adaptation for ps mode.
Merge pull request !3027 from ZPaC/adaptation-for-ps-mode
5 years ago
WilliamLian
9682d08d96
refactor primitive hook function
5 years ago
ZPaC
3618b0843d
Adaptation for ps mode.
5 years ago
mindspore-ci-bot
e03bd975a9
!1898 data dump
Merge pull request !1898 from caifubi/data-dump
5 years ago
WilliamLian
ea9b5468bb
fix bug of hccl kernel info
5 years ago
mindspore-ci-bot
6dcfbbefce
!3017 Add ps optimizer kernels
Merge pull request !3017 from ZPaC/add-ps-optimizer-kernel
5 years ago
mindspore-ci-bot
95048955ef
!2925 Add random normal op at GPU
Merge pull request !2925 from peixu_ren/custom_gpu
5 years ago
caifubi
c577952c9a
Async Data Dump
5 years ago
ZPaC
6bb83ad3e1
Add ps optimizer kernels.
5 years ago
ZPaC
bb6373d690
Add ps embedding lookup kernels
5 years ago
peixu_ren
cccb230f7b
Add random normal cuda implementation on GPU
5 years ago
mindspore-ci-bot
f201bd655c
!2979 Add push pull kernel
Merge pull request !2979 from ZPaC/add-push-pull-kernel
5 years ago
ZPaC
e0da486e2f
Add push pull kernels
5 years ago
mindspore-ci-bot
ee199007ed
!2942 change reduce node's reduce axis attr whem it's using in special foramt
Merge pull request !2942 from lianliguang/convert-reduce-axis-when-selected-hw-special-format-for-reduce-kernel
5 years ago
mindspore-ci-bot
cf5a27e97d
!2962 gpu support SmoothL1Loss kernel
Merge pull request !2962 from chenweifeng/smoothl1loss
5 years ago
mindspore-ci-bot
d1e7b9770f
!2964 Increase the max tensor size
Merge pull request !2964 from jiangzhenguang/Increase-the-max-tensor-size
5 years ago
jzg
fb90ff164b
increase the max size of tensor.
5 years ago
mindspore-ci-bot
7fd22f8ede
!2971 Reg op info from local config file
Merge pull request !2971 from jjfeing/master
5 years ago
wilfChen
0fdc304a8e
gpu support smoothl1loss
5 years ago
jzg
94d0d45ab4
increase the max size of tensor.
5 years ago
WilliamLian
4200925383
convert the reduce axis attr when the reduce node selected the special format
5 years ago
jjfeing
53277f8c02
reg op info from local config file
5 years ago
yujianfeng
3fdc3629af
Check attr exists before getting it in embeddinglookup cpu kernel
5 years ago
mindspore-ci-bot
0c7a5f76a3
!2779 gpu support CTCLoss kernel
Merge pull request !2779 from chenweifeng/ctcloss
5 years ago
wilfChen
d54154a1f9
Gpu support ctcloss kernel
5 years ago
kswang
4094c4688d
use multi thread for reduce sparse gradient
5 years ago
mindspore-ci-bot
b82926138d
!2905 two level reduce sparse gradient
Merge pull request !2905 from kisnwang/parallel-reduce-sparse-gradient
5 years ago
kswang
a719a9fe3a
add two level reduce sparse gradient
5 years ago
liuxiao93
2b0ecfd2b1
Add TBE op UnsortedSegmentProd for VM.
5 years ago
mindspore-ci-bot
f92c4a5312
!2862 reselect the kernel after rectify the build info of domask
Merge pull request !2862 from lianliguang/reselect-the-kernel-info-after-rectify-domask-kernel
5 years ago
WilliamLian
60e3849178
reselect the domask's child node after rectify the node domask
5 years ago
mindspore-ci-bot
99d6430b3f
!2835 fix cpu gatherV2 index out of bounds
Merge pull request !2835 from baihuawei/gatherV2
5 years ago
baihuawei
b1d6ef0e88
fix GatherV2 index out of bounds
5 years ago