Skip to content
Projects
Groups
Snippets
Help
Loading...
Sign in / Register
Toggle navigation
N
ngraph
Project
Project
Details
Activity
Cycle Analytics
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Charts
Issues
0
Issues
0
List
Board
Labels
Milestones
Merge Requests
0
Merge Requests
0
CI / CD
CI / CD
Pipelines
Jobs
Schedules
Charts
Packages
Packages
Wiki
Wiki
Snippets
Snippets
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Charts
Create a new issue
Jobs
Commits
Issue Boards
Open sidebar
submodule
ngraph
Commits
99245594
Commit
99245594
authored
6 years ago
by
Sergey Shalnov
Committed by
Sang Ik Lee
6 years ago
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
IntelGPU backend: Broadcast operation fix (#2461)
parent
780ef08c
Show whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
18 additions
and
5 deletions
+18
-5
external_cldnn.cmake
cmake/external_cldnn.cmake
+1
-1
intelgpu_backend.cpp
src/ngraph/runtime/intelgpu/intelgpu_backend.cpp
+17
-4
No files found.
cmake/external_cldnn.cmake
View file @
99245594
...
...
@@ -22,7 +22,7 @@ include(ExternalProject)
#------------------------------------------------------------------------------
set
(
CLDNN_GIT_REPO_URL https://github.com/intel/clDNN.git
)
set
(
CLDNN_GIT_LABEL
f91d7d83d8f121e4e159776b108e316f2f08bdf5
)
set
(
CLDNN_GIT_LABEL
a34a1614af20dafa2864e2be0806a8d80b4eb356
)
set
(
BOOST_VERSION 1.64.0
)
set
(
OUT_DIR
${
EXTERNAL_PROJECTS_ROOT
}
/cldnn/out
)
...
...
This diff is collapsed.
Click to expand it.
src/ngraph/runtime/intelgpu/intelgpu_backend.cpp
View file @
99245594
...
...
@@ -844,14 +844,27 @@ shared_ptr<runtime::Executable>
{
do_equal_propagation
(
topology
,
get_input_name
(
op
),
get_output_name
(
op
));
}
else
if
(
get_input_type
(
op
)
!=
element
::
i32
&&
get_input_type
(
op
)
!=
element
::
i64
&&
((
get_input_shape
(
op
).
size
()
==
1
&&
get_input_shape
(
op
).
at
(
0
)
==
1
)
||
get_input_shape
(
op
).
empty
()))
else
if
((
get_output_shape
(
op
).
size
()
<=
4
)
&&
((
get_input_type
(
op
)
==
element
::
f32
)
||
(
get_input_type
(
op
)
==
element
::
i32
)))
{
const
size_t
shift
=
4
-
get_output_shape
(
op
).
size
();
vector
<
uint16_t
>
fixed_b_axes
;
for
(
uint16_t
i
=
0
;
i
<
shift
;
++
i
)
{
fixed_b_axes
.
push_back
(
i
);
}
for
(
auto
it
=
axis
.
cbegin
();
it
!=
axis
.
cend
();
++
it
)
{
fixed_b_axes
.
push_back
(
*
it
+
shift
);
}
const
cldnn
::
tensor
output_tensor_size
=
intelgpu_space
::
create_cldnn_tensor
(
get_output_shape
(
op
));
const
cldnn
::
broadcast
cldnn_broadcast
(
get_output_name
(
op
),
get_input_name
(
op
),
output_tensor_size
);
get_output_name
(
op
),
get_input_name
(
op
),
output_tensor_size
,
fixed_b_axes
);
topology
.
add
(
cldnn_broadcast
);
}
else
...
...
This diff is collapsed.
Click to expand it.
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment