Skip to content

Implements blob compatibility check for NPU#430

Merged
preetha-intel merged 5 commits intoovep-develop-lnl-1.2from
sriram/blob_compatibility_check
Sep 4, 2024
Merged

Implements blob compatibility check for NPU#430
preetha-intel merged 5 commits intoovep-develop-lnl-1.2from
sriram/blob_compatibility_check

Conversation

@preetha-intel
Copy link

Description

  • OVEP catches the NPU driver exception and return failure status
  • NPU to CPU fallback is disabled when inferencing with blob

Motivation and Context

  • When the NPU driver version of blob creation and inference differs, it fails with NPU driver exception. Handled that exception and throw meaningful exception message
  • Restricted the NPU to CPU fallback when inferencing with blob

@sfatimar
Copy link

Please check other EP Codes to see how they handle failure in Compile call

* OVEP catches the NPU driver exception and return failure status

* NPU to CPU fallback is disabled when inferencing with blob
@vthaniel vthaniel force-pushed the sriram/blob_compatibility_check branch from 31b4e5e to 4c4a5b2 Compare August 30, 2024 07:19
@srirammaswamy-intel srirammaswamy-intel force-pushed the sriram/blob_compatibility_check branch 2 times, most recently from 539c7d1 to c80fd27 Compare August 30, 2024 10:03
@vthaniel vthaniel force-pushed the sriram/blob_compatibility_check branch from c80fd27 to 5429e55 Compare September 2, 2024 11:16
* Changes failure status code to exception (std::runtime_error)

* Capture all NPU related errors

* Throw minimal error message with error type and error code for Release
  builds
@vthaniel vthaniel force-pushed the sriram/blob_compatibility_check branch from 5429e55 to e4ff4ba Compare September 2, 2024 11:18
@ankitm3k
Copy link

ankitm3k commented Sep 2, 2024

LGTM

@sfatimar
Copy link

sfatimar commented Sep 4, 2024

@srirammaswamy-intel can you please close on the review comments by Preetha

@sfatimar
Copy link

sfatimar commented Sep 4, 2024

@preetha-intel can you please close on this.

@vthaniel vthaniel force-pushed the sriram/blob_compatibility_check branch from dc71179 to f5ca289 Compare September 4, 2024 06:00
@vthaniel vthaniel force-pushed the sriram/blob_compatibility_check branch from f5ca289 to 4203dcb Compare September 4, 2024 09:02
@preetha-intel
Copy link
Author

LGTM

@preetha-intel preetha-intel merged commit 86e25f0 into ovep-develop-lnl-1.2 Sep 4, 2024
@preetha-intel preetha-intel deleted the sriram/blob_compatibility_check branch September 4, 2024 09:32
sfatimar pushed a commit that referenced this pull request Sep 6, 2024
* Implements blob compatibility check for NPU

* OVEP catches the NPU driver exception and return failure status

* NPU to CPU fallback is disabled when inferencing with blob

* Update NPU device exception handling approach

* Changes failure status code to exception (std::runtime_error)

* Capture all NPU related errors

* Throw minimal error message with error type and error code for Release
  builds

* Fix lint issues

* Address review comments

* Address review comments

---------

Co-authored-by: Srirammaswamy <srirammaswamy.s@intel.com>
sfatimar pushed a commit that referenced this pull request Sep 6, 2024
* Implements blob compatibility check for NPU

* OVEP catches the NPU driver exception and return failure status

* NPU to CPU fallback is disabled when inferencing with blob

* Update NPU device exception handling approach

* Changes failure status code to exception (std::runtime_error)

* Capture all NPU related errors

* Throw minimal error message with error type and error code for Release
  builds

* Fix lint issues

* Address review comments

* Address review comments

---------

Co-authored-by: Srirammaswamy <srirammaswamy.s@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants