baihuawei
9f6a6058f5
fix cpu type excption
5 years ago
yanghaoran
7a2a714a11
run dynamic shape in existing context
5 years ago
mindspore-ci-bot
07c8a6114e
!7637 raise reduce precision bug fix
Merge pull request !7637 from liubuyu/op_support
5 years ago
mindspore-ci-bot
5461dd056b
!7452 Check absolute dump path
Merge pull request !7452 from caifubi/dump
5 years ago
liubuyu
f4446c30c8
raise reduce precision bug fix
5 years ago
mindspore-ci-bot
d54ea04df3
!7630 dynamic shape bug fix
Merge pull request !7630 from liubuyu/master
5 years ago
mindspore-ci-bot
55a277d1de
!7572 GPU reduce precision fix
Merge pull request !7572 from VectorSL/fix-reduce-pre
5 years ago
liubuyu
2b14316e40
dynamic shape bug fix: output shape is not updated
5 years ago
caifubi
980bfafc64
Check dump absolute path
5 years ago
mindspore-ci-bot
93c03ca4df
!7565 Decouple GraphKernel ops from akg op register
Merge pull request !7565 from DeshiChen/1012_gen_json_from_anf
5 years ago
mindspore-ci-bot
56bcaa5039
!7410 enable async run graph
Merge pull request !7410 from kisnwang/enable-async-run-graph
5 years ago
mindspore-ci-bot
08dad79529
!7533 fix bug of gpu‘s kernel setter
Merge pull request !7533 from lianliguang/master
5 years ago
VectorSL
bbcdd81d1b
fix reduce precision: deal tuplegetitem and param
5 years ago
dayschan
f429fe7888
Decouple GraphKernel ops from akg op register
1. Extract op info from AnfNode when generating composite operators.
2. Ignore the op register info when selecting kernel.
3. Update akg submodule.
4. Remove akg gpu op registers that added for GraphKernel.
5. Add an interface `IsNodeInGraphKernel` into `AnfRuntimeAlgorithm`, next to the `IsGraphKernel`.
6. Sort included fils for the clang-format rule is changed.
TODO: Remove the json generator in GraphKernelExpander, it's ok to call this interface.
5 years ago
mindspore-ci-bot
89cd882077
!7520 Critical path performance optimization
Merge pull request !7520 from chenweifeng/resnet50-profiling
5 years ago
wilfChen
b8e1c03cdc
Critical path performance optimization
5 years ago
mindspore-ci-bot
40b4844b76
!7263 Remove debugger argument from runtime functions
Merge pull request !7263 from Harshvardhan Gupta/add-dbg-runtime
5 years ago
Lianliguang
61f3c134c0
fix bug of gpu refresh parameter & valuenode's format info when kernel selecting && do not refresh ZN_LSTM format for valuenode
5 years ago
mindspore-ci-bot
84f66ef5b8
!7217 reduce or raise precision restructure
Merge pull request !7217 from liubuyu/op_support
5 years ago
gukecai
b9b4a86e5f
fix atomicaddrclean bug
5 years ago
liubuyu
020c71e72c
reduce or raise precision restructure
5 years ago
kswang
ece27f313e
enable async run
5 years ago
Harshvardhan Gupta
744355a005
remove dbg args from runtime and remove needless argument from IsWatchpoint
5 years ago
mindspore-ci-bot
687e9126aa
!7267 dump kernel inputs/outputs based on the input_output field of the config file.
Merge pull request !7267 from AdelShafiei/kernels_dump
5 years ago
mindspore-ci-bot
4424873e7e
!7256 GPU add reduce precision
Merge pull request !7256 from VectorSL/reduceprecision
5 years ago
laiyongqiang
b76fa31cc5
no need to add atomic for communication op's output
5 years ago
mindspore-ci-bot
283ab84884
!7269 dump constant values for each step
Merge pull request !7269 from lichen_101010/dump_constant_values_new
5 years ago
Adel Shafiei
e4ebdc599e
Added input/output kernel dump support based on the config file
5 years ago
lichen_101010
678962d718
implementation for dumping const values for each step
fix no output address error
address clang-format error
address some comments
clang_format check
5 years ago
VectorSL
ccab6f88d5
gpu add reduce precision:int64->int32
5 years ago
caifubi
9b76f4ed57
get physical device for gpu
5 years ago
mindspore-ci-bot
d9b7b65120
!7237 reverse reduce precision support int64 to int32
Merge pull request !7237 from liubuyu/bug_fix
5 years ago
liubuyu
9f7921a663
reverse reduce precision support int64 to int32
5 years ago
laiyongqiang
60fc029307
add atomic clean for all type communication input
5 years ago
mindspore-ci-bot
39bc43e674
!5990 fix parameter in multi-graph
Merge pull request !5990 from lianliguang/GAN
5 years ago
mindspore-ci-bot
618a876e0c
!7179 fix exception not exit
Merge pull request !7179 from kisnwang/fix-exception-not-exit
5 years ago
kswang
db985ab909
add ms exception
5 years ago
caifubi
fa6293d905
fix hccl dynamic kernel finalize bug
5 years ago
caifubi
d3b978147f
Ascend Dynamic Shape
5 years ago
mindspore-ci-bot
c69382c870
!5621 enable async run graph
Merge pull request !5621 from kisnwang/enable-async-run-graph
5 years ago
mindspore-ci-bot
9e48d6527a
!7131 Change directory to read the overflow dumps from
Merge pull request !7131 from Harshvardhan Gupta/fix-overflow
5 years ago
mindspore-ci-bot
f90529a4a8
!6996 In debugger, Add output filter for BatchNorm to avoid loading dirty output values
Merge pull request !6996 from lichen_101010/batch_norm
5 years ago
kswang
11989b5e30
enable async run
5 years ago
mindspore-ci-bot
14d1c4390b
!7086 Only dump the specified kernels in dump config file, GPU mode
Merge pull request !7086 from AdelShafiei/kernels_dump
5 years ago
Adel Shafiei
aa33ae12fb
Only dump the specified kernel in GPU mode
5 years ago
mindspore-ci-bot
d3bd1cb5b8
!6668 [Data Dump] [master] Fix asynchronous data dump failure
Merge pull request !6668 from caifubi/dump
5 years ago
mindspore-ci-bot
21c5607fca
!6971 cudnn inplace optimizer
Merge pull request !6971 from chenweifeng/tensoradd_inplace
5 years ago
Harshvardhan Gupta
e77e816f1d
fix overflow dump path
5 years ago
Harshvardhan Gupta
7c5e0541ba
load inputs before suspending execution in dbg
5 years ago
wilfChen
b420b6cda7
cudnn inplace optimizer
5 years ago