Skip to content

Commit 9f0119b

Browse files
Copilotpelikhan
Copilot
and
pelikhan
authored
feat(qmd): use aw-gpu-runner-T4 as default indexing job runner
Change the default runner for the qmd indexing job from ubuntu-latest to aw-gpu-runner-T4, enabling GPU acceleration for node-llama-cpp during the embedding phase. Users can still override with runs-on:. Add DefaultQmdIndexingRunnerImage constant in pkg/constants/constants.go. Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com> Agent-Logs-Url: https://github.com/github/gh-aw/sessions/78657957-9d39-4cf3-8b27-2f585b36aeb4
1 parent 203269b commit 9f0119b

4 files changed

Lines changed: 8 additions & 4 deletions

File tree

‎.github/workflows/dev.lock.yml‎

Lines changed: 1 addition & 1 deletion
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

‎.github/workflows/smoke-codex.lock.yml‎

Lines changed: 1 addition & 1 deletion
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

‎pkg/constants/constants.go‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -433,6 +433,10 @@ const DefaultPlaywrightMCPVersion Version = "0.0.68"
433433
// DefaultQmdVersion is the default version of the @tobilu/qmd npm package
434434
const DefaultQmdVersion Version = "2.0.1"
435435

436+
// DefaultQmdIndexingRunnerImage is the default runner image for the qmd indexing job.
437+
// Uses the GPU-enabled T4 runner image so node-llama-cpp can leverage the GPU during embedding.
438+
const DefaultQmdIndexingRunnerImage = "aw-gpu-runner-T4"
439+
436440
// DefaultPlaywrightBrowserVersion is the default version of the Playwright browser Docker image
437441
const DefaultPlaywrightBrowserVersion Version = "v1.58.2"
438442

‎pkg/workflow/qmd.go‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -598,8 +598,8 @@ func (c *Compiler) buildQmdIndexingJob(data *WorkflowData) (*Job, error) {
598598
})
599599

600600
// Determine the runner for the indexing job.
601-
// Default to ubuntu-latest; user can override via qmd.runs-on.
602-
indexingRunsOn := "runs-on: ubuntu-latest"
601+
// Default to aw-gpu-runner-T4 for GPU-accelerated embedding; user can override via qmd.runs-on.
602+
indexingRunsOn := "runs-on: " + constants.DefaultQmdIndexingRunnerImage
603603
if data.QmdConfig.RunsOn != "" {
604604
indexingRunsOn = "runs-on: " + data.QmdConfig.RunsOn
605605
}

0 commit comments

Comments
 (0)