Skip to content

Conversation

@nihui
Copy link
Member

@nihui nihui commented Feb 8, 2026

z-image-ncnn 1024x1024

end2end 7900xtx
baseline 39.70s
+rebar 38.35s

@github-actions github-actions bot added the core label Feb 8, 2026
@tencent-adm
Copy link
Member

CLA assistant check
Thank you for your submission, we really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
You have signed the CLA already but the status is still pending? Let us recheck it.

@codecov-commenter
Copy link

codecov-commenter commented Feb 8, 2026

Codecov Report

❌ Patch coverage is 67.56757% with 12 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.94%. Comparing base (7d5d32d) to head (3918855).
⚠️ Report is 1 commits behind head on master.

Files with missing lines Patch % Lines
src/allocator.cpp 25.00% 9 Missing ⚠️
src/gpu.cpp 88.00% 3 Missing ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##           master    #6536      +/-   ##
==========================================
- Coverage   92.94%   92.94%   -0.01%     
==========================================
  Files         809      809              
  Lines      256714   256613     -101     
==========================================
- Hits       238613   238508     -105     
- Misses      18101    18105       +4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

Adds a Vulkan-side capability probe for “Resizable BAR” (approximated via memory property inspection) and uses it to prefer more upload-friendly memory types for weight buffers/images to improve end-to-end performance.

Changes:

  • Expose GpuInfo::resizable_bar_enabled() as a new GPU property.
  • Cache/query physical device memory properties during GPU init and log rebar status.
  • Adjust VkWeightAllocator memory-type selection to prefer device-local + host-visible allocations when ReBAR is detected.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File Description
src/gpu.h Adds GpuInfo::resizable_bar_enabled() API.
src/gpu.cpp Implements memory-property query and derives/logs the ReBAR flag.
src/allocator.cpp Uses the ReBAR flag to tweak weight allocation memory preferences.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 6 comments.


💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

@nihui nihui merged commit 37b499c into Tencent:master Feb 9, 2026
107 of 109 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants