Skip to content

[SYMPP-PERF] Start verified Windows MCP release directly - #569

Merged
Pimpmuckl merged 2 commits into
mainfrom
perf/mcp-startup-followup
Aug 8, 2026
Merged

[SYMPP-PERF] Start verified Windows MCP release directly#569
Pimpmuckl merged 2 commits into
mainfrom
perf/mcp-startup-followup

Conversation

@Pimpmuckl

Copy link
Copy Markdown
Owner

Context

The verified Windows artifact path nested PowerShell before the release launcher, adding about 1.3 seconds before readiness.

TL;DR

  • Start stock Windows artifacts through their packaged release batch file and expose cold-start phase timings.

Summary

  • Launch the stock release directly while preserving artifact, workflow, log, port, and dashboard environment.
  • Keep the wrapper for custom manifest arguments and avoid parsing artifact-root characters through cmd.exe.
  • Report backend-start, runtime-ready, and tools/list timings in the exact-command benchmark.
  • Reduce exact cold initialize from 11.43 seconds to 9.19 seconds in repeated isolated measurements.

Alternatives

  • Pre-stage artifacts or move waiting off the measured path; rejected because that shifts work instead of reducing it.
  • Replace the launcher or add a service; rejected because the packaged release already provides the required entrypoint.

Test Plan

  • make -C elixir all
  • make -C elixir ci-full when release/full validation is required
  • pwsh -NoProfile -File plugins/symphony-plus-plus-mcp/tests/launcher/run-launcher-tests.ps1
  • pwsh -NoProfile -File scripts/benchmarks/sympp-mcp/run-performance-gate.ps1 -SelfTest
  • pwsh -NoProfile -File scripts/benchmarks/sympp-mcp/run-performance-gate.ps1
  • Review Suite fast rvw_1e1aae6a

ci-full was not run because this changes only the PowerShell launcher and benchmark. The full packaged performance gate passed twice.

Pimpmuckl and others added 2 commits August 3, 2026 03:10
Summary:
- launch verified Windows artifacts through the packaged release entrypoint
- preserve artifact acknowledgement, server, and isolated release temp settings
- expose exact-command phase timings and cover direct command selection

Rationale:
- the nested PowerShell startup wrapper added about 1.3 seconds before backend readiness
- the release batch entrypoint already performs the required packaged-runtime launch

Tests:
- make -C elixir all
- pwsh -NoProfile -File plugins/symphony-plus-plus-mcp/tests/launcher/run-launcher-tests.ps1
- pwsh -NoProfile -File scripts/benchmarks/sympp-mcp/run-performance-gate.ps1
- pwsh -NoProfile -File scripts/benchmarks/sympp-mcp/run-performance-gate.ps1 -SelfTest

Co-authored-by: Codex <codex@openai.com>
Summary:
- merge the current main branch into the startup performance PR
- preserve both process-runtime and beta-launcher regression coverage

Rationale:
- validate the startup optimization against current production behavior
- keep newer launcher tests while resolving the only textual conflict

Tests:
- launcher run-launcher-tests.ps1
- make -C elixir all

Co-authored-by: Codex <codex@openai.com>
@Pimpmuckl
Pimpmuckl merged commit 4df20c2 into main Aug 8, 2026
22 checks passed
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.

1 participant