i-robot
e67f83fdf1
!26983 fix FusedMulAddN code trace print
Merge pull request !26983 from yuchaojie/ir_fusion
4 years ago
i-robot
6daea6079b
!26906 Optimize DynamicRNNGrad fission when batch_size is multiple of 16 or t_size is 1
Merge pull request !26906 from yuchaojie/ir_fusion3
4 years ago
yuchaojie
d3d1d5e97b
fix FusedMulAddN code trace print
4 years ago
lby
618d4b7dfe
move key 'trace' to interface DumpSourceLines
4 years ago
yuchaojie
805541a3a9
Optimize DynamicRNNGrad fission when batch_size is multiple of 16 or t_size is 1
4 years ago
i-robot
900c9ef88b
!26812 add dump flag for transdata which is inserted for multi-output node
Merge pull request !26812 from yuchaojie/ir_fusion
4 years ago
i-robot
c87a50a174
!26763 add trace for exception log
Merge pull request !26763 from yuchaojie/ir_fusion3
4 years ago
yuchaojie
6a700483c5
add dump flag for transdata which is inserted for multi-output node
4 years ago
yuchaojie
c09501aad0
add trace for exception log
4 years ago
i-robot
b11c6d0591
!26727 log improvement
Merge pull request !26727 from laiyongqiang/adapter
4 years ago
LaiYongqiang
a42c936df6
log improvement
4 years ago
i-robot
fa5ea7b3a6
!26370 DynamicRNNGrad support `hidden_size not multiple of 16` scene
Merge pull request !26370 from yuchaojie/ir_fusion4
4 years ago
i-robot
6c0547106e
!26697 make trace source lines more accurate
Merge pull request !26697 from yuchaojie/ir_fusion
4 years ago
yuchaojie
c3eea22ab4
make trace source lines more accurate
4 years ago
He Wei
41dcac9c49
Replace std::unordered_map/set with robin-hood-hashing
Robin-hood-hashing (https://github.com/martinus/robin-hood-hashing )
is considered faster then std::unordered_map/set,
so we use it to improve mindspore performance.
1. robin_hood head file in `third_party/robin_hood/include`;
2. In `utils/hash_map.h` and `utils/hash_set.h`, we define:
- mindspore::HashMap as an alias of robin_hood::unordered_map;
- mindspore::HashSet as an alias of robin_hood::unordered_set;
3. Replace:
- `#include <unordered_map>` --> `#include "utils/hash_map.h"`;
- `#include <unordered_set>` --> `#include "utils/hash_set.h"`;
- `std::unordered_map` --> `mindspore::HashMap`;
- `std::unordered_set` --> `mindspore::HashSet`;
- `map.insert(std::pair(key, value))` --> `map.emplace(key, value)`;
- `[] (const std::pair<K, V> &p) {..} ` --> `[] (const auto &p) {..} `;
4. Fix issues found by switch to robin_hood:
- AnfNodeConfig hash and equal;
- Fix a bug in `Slice::operator==()`;
- Fix a bug in `CNode::HasPrimalAttr()`;
- Fix map.erase() usage bugs: `map.erase(iter++)` --> `iter = map.erase(iter)`;
- Fix some iterator invalidated problem;
5. Some std::unordered_map/set can not replace by robin_hood:
- As parameter of functions that exposed to python by pybind11;
- Use bad hash that cause robin_hood::map over_flow, such as AbstractBasePtrListHasher;
6. Update cpp unit tests;
7. Add build option '-F' to enable robin_hood, default on.
4 years ago
yuchaojie
b760eba23a
DynamicRNNGrad support `hidden_size not multiple of 16` scene
4 years ago
yuchaojie
c75077cd5a
add dump flag for inserted transdata&cast
4 years ago
i-robot
62202f8186
!26416 add MatmulAddFusion
Merge pull request !26416 from yuchaojie/ir_fusion
4 years ago
yuchaojie
3e2c952419
add MatmulAddFusion
4 years ago
yuchaojie
afc81cf8e4
add log print for memcpy_s errorno
4 years ago
i-robot
5233c73805
!25592 Reshape support shape is variable
Merge pull request !25592 from wangnan39/reshape_support_tensor
4 years ago
王南
1163cfe967
reshape support shape is tensor
4 years ago
ttudu
33ac1de062
fix bug
4 years ago
i-robot
f2e6dc286c
!26116 add NodeDebugInfo for fusion nodes
Merge pull request !26116 from yuchaojie/ir_fusion3
4 years ago
hwjiaorui
1f8773e00d
add sparse unify mindir pass
4 years ago
i-robot
10b63dffc0
!25744 IR fusion adapts dump flag
Merge pull request !25744 from yuchaojie/ir_fusion
4 years ago
yuchaojie
f2dd7c07fe
IR fusion adapts dump flag
4 years ago
yuchaojie
31c2435f66
add NodeDebugInfo for fusion nodes
4 years ago
i-robot
816de6f0ee
!26055 convert attr to value node
Merge pull request !26055 from yanzhenxiang2020/aicpu_random_seed_to_input
4 years ago
i-robot
8bf7e28fa6
!25410 add dump flag for fusion nodes
Merge pull request !25410 from yuchaojie/ir_fusion3
4 years ago
jjfeing
34b73e305d
convert attr to value node
4 years ago
dayschan
cbb84ff580
Move IsRealKernel and IsRealCNodeKernel from AnfAlgo to AnfUtils
the function IsOneOfPrimitive and IsOneOfPrimitiveCNode is useful,
we can move them into anf.cc
4 years ago
yuchaojie
0c90aecae4
add dump flag for fusion nodes
4 years ago
i-robot
0bb2d8a8cb
!25927 fix node suspend problem in InsertTransOp&InsertCast
Merge pull request !25927 from yuchaojie/op_select
4 years ago
yuchaojie
3640c495ad
fix node suspend problem in InsertTransOp&InsertCast
4 years ago
ttudu
e953c15cd2
NeighborExchangeV2 & Grad
4 years ago
i-robot
a6e077a7d7
!25733 fix transpose format
Merge pull request !25733 from yuchaojie/op_select
4 years ago
i-robot
2224f75b47
!25696 add pass switch
Merge pull request !25696 from zhoufeng/pass-switch
4 years ago
zhoufeng
fb17c448c2
add pass switch
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
4 years ago
yuchaojie
bee2797b3f
fix transpose format
4 years ago
王南
1ffc3ae58a
ops dynamic shape fix
4 years ago
lby
6872e67131
split compile ang gen kernel mod
4 years ago
jjfeing
c6520f878f
status record in vm
4 years ago
jjfeing
19235dd91f
insert when input is gettuple
4 years ago
jjfeing
6f5164f5e0
fix insert tensormove
4 years ago
zhoufeng
accce47fbe
delete some pass from lic
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
4 years ago
i-robot
499e0e6f26
!24682 code check clean
Merge pull request !24682 from yuchaojie/code-clean
4 years ago
yuchaojie
ed25dd2d21
code check clean
4 years ago
yuchaojie
b7e4098770
fix dynamic_rnn_grad_fission_v2 output shape
4 years ago
i-robot
e4d8e63e8c
!24606 add pass lic switch
Merge pull request !24606 from zhoufeng/add-lic-pass
4 years ago