Gate weights cache on runtime option instead of compile-time macro (#19603)#19603
Open
hboyraz wants to merge 2 commits into
Open
Gate weights cache on runtime option instead of compile-time macro (#19603)#19603hboyraz wants to merge 2 commits into
hboyraz wants to merge 2 commits into
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19603
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ⏳ No Failures, 138 PendingAs of commit 8601d3b with merge base 09a7cbe ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
Contributor
|
@hboyraz has exported this pull request. If you are a Meta employee, you can view the originating Diff in D105123995. |
This PR needs a
|
JacobSzwejbka
approved these changes
May 14, 2026
Author
|
@pytorchbot label "release notes: bug fix" |
|
Didn't find following labels among repository labels: release notes: bug fix |
GregoryComer
approved these changes
May 14, 2026
meta-codesync
Bot
changed the title
hboyraz
pushed a commit
to hboyraz/executorch
that referenced
this pull request
May 15, 2026
…ytorch#19603) Summary: Replaces the compile-time `#ifdef ENABLE_XNNPACK_WEIGHTS_CACHE` gate in XNNCompiler.cpp with a runtime boolean plumbed from `XnnpackBackendOptions::resolve_weight_cache(context)` through `XNNPACKBackend::init` to `XNNCompiler::compileModel`. This fixes a silent-disable bug: previously, runtime opt-in via `set_option(weight_cache_option_key, true)` was silently a no-op unless the build also set `-c executorch.xnnpack_weights_cache=1`, because the cache pointer handed to `xnn_create_runtime_v4` was hardcoded to nullptr when the macro was undefined. Multimethod LoRA models re-packed the entire backbone for every method load, costing hundreds of MB of resident memory. The runtime path now keys all three cache-relevant code regions (unpacked-data load, cache pointer handoff to xnn_create_runtime_v4, and finalize_for_runtime) on `bool use_weight_cache` resolved per-init from the BackendInitContext. The `Result<vector<string>>` declaration in compileModel was reshaped to plain `vector<string>` since `Result<>` is non-assignable, which is required for the new runtime branch. Reviewed By: GregoryComer Differential Revision: D105123995
hboyraz
force-pushed
the
export-D105123995
branch
from
e33c3f9 to
fcbc108
Compare
…ytorch#19603) Summary: Replaces the compile-time `#ifdef ENABLE_XNNPACK_WEIGHTS_CACHE` gate in XNNCompiler.cpp with a runtime boolean plumbed from `XnnpackBackendOptions::resolve_weight_cache(context)` through `XNNPACKBackend::init` to `XNNCompiler::compileModel`. This fixes a silent-disable bug: previously, runtime opt-in via `set_option(weight_cache_option_key, true)` was silently a no-op unless the build also set `-c executorch.xnnpack_weights_cache=1`, because the cache pointer handed to `xnn_create_runtime_v4` was hardcoded to nullptr when the macro was undefined. Multimethod LoRA models re-packed the entire backbone for every method load, costing hundreds of MB of resident memory. The runtime path now keys all three cache-relevant code regions (unpacked-data load, cache pointer handoff to xnn_create_runtime_v4, and finalize_for_runtime) on `bool use_weight_cache` resolved per-init from the BackendInitContext. The `Result<vector<string>>` declaration in compileModel was reshaped to plain `vector<string>` since `Result<>` is non-assignable, which is required for the new runtime branch. Reviewed By: GregoryComer Differential Revision: D105123995
hboyraz
force-pushed
the
export-D105123995
branch
from
fcbc108 to
bbf2b17
Compare
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Summary:
Replaces the compile-time
#ifdef ENABLE_XNNPACK_WEIGHTS_CACHEgate inXNNCompiler.cpp with a runtime boolean plumbed from
XnnpackBackendOptions::resolve_weight_cache(context)throughXNNPACKBackend::inittoXNNCompiler::compileModel.This fixes a silent-disable bug: previously, runtime opt-in via
set_option(weight_cache_option_key, true)was silently a no-op unlessthe build also set
-c executorch.xnnpack_weights_cache=1, because thecache pointer handed to
xnn_create_runtime_v4was hardcoded to nullptrwhen the macro was undefined. Multimethod LoRA models re-packed the entire backbone for every method load, costing
hundreds of MB of resident memory.
The runtime path now keys all three cache-relevant code regions
(unpacked-data load, cache pointer handoff to xnn_create_runtime_v4, and
finalize_for_runtime) on
bool use_weight_cacheresolved per-init fromthe BackendInitContext.
The
Result<vector<string>>declaration in compileModel was reshaped toplain
vector<string>sinceResult<>is non-assignable, which isrequired for the new runtime branch.
Reviewed By: GregoryComer
Differential Revision: D105123995