[GPU] Add support for i16, u16, and u32 element types in remote tensors #27573
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jade-cho
requested review from
sshlyapn,
isanghao,
vladimir-paramuzov,
sungeunk and
byungilm
+ Removed host memory data transformation for user input/output tensors with allocation type cl_mem and data types i16, u16, or u32. + User tensors can now be directly used as plugin tensors without additional data conversion overhead.
jade-cho
force-pushed
the
remote_tensor_support_other_dtypes
branch
from
a0a194b to
f72547e
Compare
+ Do not add a reorder when user node is eltwise. + Because eltwise already supports i16, u16 and u32 data types.
src/plugins/intel_gpu/src/graph/graph_optimizer/add_required_reorders.cpp
Outdated
Show resolved
Hide resolved
+ Removes unnecessary code in add_required_reorders. + Changes the graph of the added func test.
| @@ -30,7 +30,8 @@ static void CreateResultOp(ProgramBuilder& p, const std::shared_ptr<ov::op::v0:: | |||
| auto out_format = cldnn::format::get_default_format(out_rank); | |||
|
|
|||
| auto out_primitive_name = layer_type_name_ID(op); | |||
| auto out_data_type = cldnn::element_type_to_data_type(convert_to_supported_device_type(op->get_input_element_type(0))); | |||
| auto out_data_type = convert_to_supported_device_type(op->get_input_element_type(0)); | |||
| out_data_type = out_data_type == ov::element::boolean ? ov::element::u8 : out_data_type; | |||
There was a problem hiding this comment.
nit: can it be embedded into convert_to_supported_device_type?
vladimir-paramuzov
approved these changes
Nov 19, 2024
jade-cho
added a commit
to jade-cho/openvino
that referenced
this pull request
Nov 21, 2024
…rs (openvinotoolkit#27573) ### Details: - *Removed host memory data converting for user input/output tensors with data types i16, u16, or u32.* - *User tensors can now be directly used as plugin tensors without additional data conversion overhead.* ### Tickets: - *156709*
NishantPrabhuFujitsu
pushed a commit
to NishantPrabhuFujitsu/openvino
that referenced
this pull request
Nov 26, 2024
…rs (openvinotoolkit#27573) ### Details: - *Removed host memory data converting for user input/output tensors with data types i16, u16, or u32.* - *User tensors can now be directly used as plugin tensors without additional data conversion overhead.* ### Tickets: - *156709*
jade-cho
added a commit
to jade-cho/openvino
that referenced
this pull request
Dec 4, 2024
+ Remove unecessary memory copy.
jade-cho
added a commit
to jade-cho/openvino
that referenced
this pull request
Dec 4, 2024
+ Remove unecessary memory copy.
jade-cho
added a commit
to jade-cho/openvino
that referenced
this pull request
Dec 4, 2024
+ Remove unecessary memory copy.
jade-cho
added a commit
to jade-cho/openvino
that referenced
this pull request
Dec 4, 2024
+ Remove unecessary memory copy.
jade-cho
added a commit
to jade-cho/openvino
that referenced
this pull request
Dec 5, 2024
+ Remove unecessary memory copy.
peterchen-intel
added a commit
to peterchen-intel/openvino
that referenced
this pull request
Jan 6, 2025
…envinotoolkit#27898)" This reverts commit ea72f30.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Details:
Tickets: