Skip to content

[NPUW]Enable NPUW_LLM_GENERATE_PYRAMID by default.#36583

Merged
dmatveev merged 3 commits into
openvinotoolkit:masterfrom
intelgaoxiong:xiong/turn_on_pyramid_for_2nd_token
Jun 30, 2026
Merged

[NPUW]Enable NPUW_LLM_GENERATE_PYRAMID by default.#36583
dmatveev merged 3 commits into
openvinotoolkit:masterfrom
intelgaoxiong:xiong/turn_on_pyramid_for_2nd_token

Conversation

@intelgaoxiong

@intelgaoxiong intelgaoxiong commented Jun 26, 2026

Copy link
Copy Markdown
Contributor

Details:

Enable NPUW_LLM_GENERATE_PYRAMID by default.

Tickets:

AI Assistance:

  • AI assistance used: no / yes
  • If yes, summarize how AI was used and what human validation was performed (build/tests/manual checks).

Signed-off-by: intelgaoxiong <xiong.gao@intel.com>
@intelgaoxiong intelgaoxiong marked this pull request as ready for review June 26, 2026 06:42
@intelgaoxiong intelgaoxiong requested review from a team as code owners June 26, 2026 06:42
@github-actions github-actions Bot added the category: NPU OpenVINO NPU plugin label Jun 26, 2026
Signed-off-by: intelgaoxiong <xiong.gao@intel.com>
#include "openvino/runtime/intel_npu/properties.hpp"
#include "whisper/prepare_whisper_model.hpp"
#include "openvino/pass/stateful_to_stateless.hpp"
#include "openvino/runtime/intel_npu/properties.hpp"

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Only changed below test cases:
LLMCompiledModelFactoryOptionsTest.CacheRopeEnabledRoundsTripThroughCompiledModel
LLMCompiledModelFactoryOptionsTest.CacheRopeDisabledRoundsTripThroughCompiledModel

Others are changed by clang-format

@dmatveev dmatveev added this to the 2026.3 milestone Jun 30, 2026
Comment on lines +355 to +363
{"NPUW_LLM_SHARED_HEAD", "YES"}, {"NPU_USE_NPUW", "YES"},
{"NPUW_DEVICES", "CPU,NPU"}, {"NPUW_SUBMODEL_DEVICE", "0:CPU,last:NPU"},
{"NPUW_WEIGHTS_BANK_ALLOC", "CPU"}, {"NPUW_CACHE_DIR", "/tmp/npuw-cache"},
{"NPUW_PARALLEL_COMPILE", "YES"}, {"NPUW_FUNCALL_ASYNC", "NO"},
{"NPUW_UNFOLD_IREQS", "YES"}, {"NPUW_FALLBACK_EXEC", "YES"},
{"NPUW_ACC_CHECK", "YES"}, {"NPUW_ACC_THRESH", "0.25"},
{"NPUW_ACC_DEVICE", "CPU"}, {"NPUW_DUMP_FULL", "YES"},
{"NPUW_DUMP_SUBS", "YES"}, {"NPUW_DUMP_SUBS_DIR", "/tmp/npuw-dumps"},
{"NPUW_DUMP_SUBS_ON_FAIL", "last"}, {"NPUW_DUMP_IO", "0,last"},

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why does it aways change this thing

@dmatveev dmatveev added this pull request to the merge queue Jun 30, 2026
Merged via the queue into openvinotoolkit:master with commit 6082b1c Jun 30, 2026
184 checks passed
@dmatveev dmatveev deleted the xiong/turn_on_pyramid_for_2nd_token branch June 30, 2026 14:35
praasz pushed a commit to praasz/openvino that referenced this pull request Jul 1, 2026
…6583)

### Details:
Enable NPUW_LLM_GENERATE_PYRAMID by default.

### Tickets:
 - *[EISW-222699](https://jira.devtools.intel.com/browse/EISW-222699)*

### AI Assistance:
 - *AI assistance used: no / yes*
- *If yes, summarize how AI was used and what human validation was
performed (build/tests/manual checks).*

---------

Signed-off-by: intelgaoxiong <xiong.gao@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

category: NPU OpenVINO NPU plugin

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants