Skip to content

Dynamic pipeline mlir deps #32203

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Draft
wants to merge 42 commits into
base: master
Choose a base branch
from
Draft

Dynamic pipeline mlir deps #32203

wants to merge 42 commits into from
+5,724 −55

Conversation

XinWangIntel
Copy link
Contributor

Details:

  • item1
  • ...

Tickets:

  • ticket-id

XinWangIntel and others added 30 commits September 11, 2025 09:31
@XinWangIntel
Add dynamic-pipeline
0f286d8 
Use point instread of MemRef
Wrap deps to zero pipeline
Refactor code
Move tensor allocator to pipeline
Try to allocate tensor on pipeline
Allocate tensor in dynamic inferrequest
Clear changes on inferrequest
Clean code
add free
Not release mem buffer to solve segment issue
Use StridedMemRefType
Clear zero_dynamic_infer_request
Remove zero dynamic_infer_request
Move ir_graph to compiler_adapter
Fix copy issue
Add TODO
Reuse engine in dynamic pipeline
Support windows
Update to latest commit
Fix stream status issue
Update mlir lib
Fix windows build and link issue
* can not use llvm cache package that built with /MT
* need to solve warnings, suppress now

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix nomral build
118bad8 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix min func issue on win and lin
b4e3c3a 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@insoow @XinWangIntel
dynamic shape
d476faa 
Signed-off-by: Insoo Woo <insoo.woo@intel.com>
@sshumihi @XinWangIntel
Fix linux build compilation
907610d 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@insoow @XinWangIntel
add NPU_LLVM_BACKEND
8f0492b 
Signed-off-by: Insoo Woo <insoo.woo@intel.com>
@insoow @XinWangIntel
support multiple pipeline (batch)
4da34ca 
Signed-off-by: Insoo Woo <insoo.woo@intel.com>
@insoow @XinWangIntel
Impelemtn updateMutableCmdList
7f105ef 
Signed-off-by: Insoo Woo <insoo.woo@intel.com>
@insoow @XinWangIntel
Fix compilation errors after rebasing the code
8b0eba9 
Signed-off-by: Insoo Woo <insoo.woo@intel.com>
@insoow @XinWangIntel
fix a compilation issue w/ NPU_LLVM_BACKEND disabled
491641c 
Signed-off-by: Insoo Woo <insoo.woo@intel.com>
@XinWangIntel
Fix accuracy check
cdb9fce 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Update level_zero_wrapper, mlir lib
fe8998c 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Use different tensor for different inferrequest in benchmark
ce2d683 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Update windows jit libs
34872d3 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix build issue
39f9091 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Add set_argument_property to update data, strides, shapes
4e13704 
* Fix xml compilation issue for benchmark
* Add new element type for benchmark dump

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Change update_graph_arguments to pass strides and shape
7ab9ffb 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Solve conflict
35922a4 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Blob will not be released in IRGraph
0b2ebff 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix style
d6e99eb 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix buffer issue
3a80d33 
* If user tensor can not be used directly, create local L0 tensor with
  user info
* If user tensor can not be used directly and old local L0 tensor is
  larger, reuse old tensor
* Add HACK_OUTPUT_SHAPE env which can use set HACK_OUTPUT_SHAPE=1*2*3*4
  to hack local L0 tensor shape for output
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Update stride in MemRefType to be element based
0c6ed60 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix style
0d1c661 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Move debug log into logger
92ac70b 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Update arguments if we reuse L0 tensor on dynamic pipeline
e9f4073 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Use plugin Metadata to show blob type
274d75f 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Solve build issue
cca7e8f 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Solve build issue
bfa4774 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix build
bb37814 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Add BlobType
5b5a96c 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Use zero_dynamic_infer_request to manage buffer for dynamic pipeline
777af5d 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix style
63ee9b0 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix metadata test and code style
8c215d0 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Code refactor
df54360 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Add InferRequestDynamicShapeTest
97d44fd 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Fix test
11e82e4 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Add test and fix engine init issue
461bbfc 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Only disable test for min size and unprocessed model
046eae5 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Add test for remote tensor and host tensor
0bbc86c 
Signed-off-by: Xin Wang <xin1.wang@intel.com>

Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel
Rebase and solve conflict
5ebc6aa 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@github-actions github-actions bot added category: build OpenVINO cmake script / infra category: samples OpenVINO Runtime Samples category: NPU OpenVINO NPU plugin labels Sep 25, 2025
@XinWangIntel
Move mlir deps to npu mlir runtime
c6c891f 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
@XinWangIntel XinWangIntel force-pushed the dynamic-pipeline-mlir-deps branch from 4dbbe8c to c6c891f Compare September 29, 2025 06:46
@XinWangIntel
Add metadata to return shapeFromIRModel
4ddb31b 
Signed-off-by: Xin Wang <xin1.wang@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
category: build OpenVINO cmake script / infra category: NPU OpenVINO NPU plugin category: samples OpenVINO Runtime Samples
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@XinWangIntel @insoow @sshumihi