chenfei
fa7cb34c8a
set abstract of switch
4 years ago
Parastoo Ashtari
8d96ee8b46
Move DumpConstantData to DebugOnStepEnd for ascend MindRT and add check for kGraphOpRun
4 years ago
i-robot
c74e418f52
!29076 Disable internal output when using MemScheduler
Merge pull request !29076 from tanghuikang/sync_host_to_device
4 years ago
tanghuikang
477ddddb0c
Disable internal output when using MemScheduler
4 years ago
baihuawei
d97fa33c49
session decoup with device target
4 years ago
i-robot
7bfb901e03
!28552 code warning clean
Merge pull request !28552 from yuchaojie/code-clean
4 years ago
maning202007
c7633672f1
Update the rank_id issue for Ascend
4 years ago
yuchaojie
90798644e5
code warning clean
4 years ago
louie5
9e629e53c5
Update DumpGraphs method with RDR, recording graph and execute order.
4 years ago
zjun
939790e1e9
Fix sens reference
Signed-off-by: zjun <zhangjun0@huawei.com>
4 years ago
i-robot
3058a16213
!28144 using data type to get cube size
Merge pull request !28144 from liubuyu/SBB
4 years ago
i-robot
955fbb8913
!28251 fix dump final irs
Merge pull request !28251 from huanghui/fix-be-dump-final-ir
4 years ago
huanghui
79db0fb1c1
fix dump final ir
4 years ago
zjun
4d9a64058d
Fix forward release
Signed-off-by: zjun <zhangjun0@huawei.com>
4 years ago
lby
e6857a6553
using cube to get c0 value
4 years ago
i-robot
2685040012
!28228 log correction
Merge pull request !28228 from yuchaojie/code-clean
4 years ago
i-robot
60335a4ab9
!28237 Enable ms function run with old control flow
Merge pull request !28237 from JoyLvliang/enable_ms_function_run_with_old_control_flow
4 years ago
yuchaojie
3903184edb
log correction
4 years ago
i-robot
708f5559b3
!28211 Bugfix for PyNative Heterogeneous in MindRT
Merge pull request !28211 from caifubi/master-pynative-mindrt-heter
4 years ago
7347157+joylvliang@user.noreply.gitee.com
8f34dbc73f
enable_ms_function_run_with_old_control_flow
4 years ago
caifubi
a84600c15d
fix pynative mindrt heter bug
4 years ago
lby
e87b5dabc4
mindrt bug fix: out put error
4 years ago
zjun
84585b602b
Fix lazy build
Signed-off-by: zjun <zhangjun0@huawei.com>
4 years ago
i-robot
73b24da0a3
!28094 fix summary ops to collect flatten ops info
Merge pull request !28094 from jiangshuqiang/fix_summary_ops
4 years ago
jiangshuqiang
5aa522a923
fix summary ops to collect flatten ops info
4 years ago
zjun
64bb8d87aa
Fix bug
Signed-off-by: zjun <zhangjun0@huawei.com>
Update value node(c4e868f5d8)
Signed-off-by: zjun <zhangjun0@huawei.com>
fix mindrt (fafd9bd053)
Signed-off-by: zjun <zhangjun0@huawei.com>
4 years ago
i-robot
15cf13dd8e
!27678 hete mindrt
Merge pull request !27678 from baihuawei/hetert
4 years ago
baihuawei
e8f641e55b
fix heterogeneous with mindrt
4 years ago
caifubi
aaa2891bc3
Enable Lazy Build in PyNative Bprop
4 years ago
zjun
eb450dd31f
Optimize pynative device memory use
Add gradient to pynative unique
Signed-off-by: zjun <zhangjun0@huawei.com>
4 years ago
caifubi
db7fbc9224
PyNative Usability:Modify log
4 years ago
caifubi
ba5516335e
PyNative Ascend MindRT
4 years ago
zhaosida
13a1250a42
fix err log
4 years ago
chujinjin
d20c9d93d8
support pynative heterogeneous
4 years ago
lby
618d4b7dfe
move key 'trace' to interface DumpSourceLines
4 years ago
yuchaojie
c09501aad0
add trace for exception log
4 years ago
i-robot
a78b4fd60f
!26722 vm err log modify
Merge pull request !26722 from zhaosida/code_clean_master
4 years ago
zhaosida
8759ac41ad
fix err log
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
caifubi
38352c1ba8
PyNative MindRT Op Lazy Build
4 years ago
xiefangqi
24ac2382ff
mac compile
4 years ago
l00591931
9353bd10a4
Enable tuple as variable in optimizer
4 years ago
i-robot
1bfedfcb7f
!26007 Ascend Device Context for MindRT
Merge pull request !26007 from hwjiaorui/ascend_context
4 years ago
i-robot
87919b1f29
!26063 Disable cache internal parameter in control flow.
Merge pull request !26063 from gaoyong10/runtime_second9
4 years ago
gaoyong10
401e8ff9a3
Disable cache internel parameter in control flow.
4 years ago
hwjiaorui
d6b2a34a69
MindRT Ascend Devcie Context
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
tanghuikang
87c7f72cf1
Sync parameter output after execute kernel when using MemScheduler
4 years ago
i-robot
0a4cc28c9d
!25761 Fix switch layer recursive bug
Merge pull request !25761 from chenfei_mindspore/switch_layer_reursive_fix
4 years ago
chenfei
d714fd4695
fix cell list
4 years ago