Enhancement: Hardware-Accelerated "Perception" via Metal #1008

Closed
opened 2026-03-22 23:04:54 +00:00 by gemini · 1 comment
Collaborator

Objective

Optimize the agent's perception stack by leveraging the M3 Max's GPU via Metal for image processing and UI state classification.

Scope

  • Refactor the integration to use Metal shaders for frame preprocessing (resizing, grayscale, contrast enhancement).
  • Ensure the Core ML classifier is explicitly pinned to the GPU/Neural Engine.
  • Benchmark perception latency before and after optimization.
  • Target: <2ms total perception latency for L1/L2 tasks.
## Objective Optimize the agent's perception stack by leveraging the M3 Max's GPU via Metal for image processing and UI state classification. ## Scope - Refactor the integration to use Metal shaders for frame preprocessing (resizing, grayscale, contrast enhancement). - Ensure the Core ML classifier is explicitly pinned to the GPU/Neural Engine. - Benchmark perception latency before and after optimization. - Target: <2ms total perception latency for L1/L2 tasks.
gemini was assigned by Rockachopa 2026-03-22 23:30:23 +00:00
claude added the gemini-review label 2026-03-23 13:51:31 +00:00
Collaborator

Kept open — aligns with harness strategy

Aligns with Perception Pipeline P0 — hardware-accelerated vision on Apple Silicon

Ref: Deep Backlog Triage #1076.

✅ **Kept open — aligns with harness strategy** Aligns with Perception Pipeline P0 — hardware-accelerated vision on Apple Silicon Ref: Deep Backlog Triage #1076.
gemini was unassigned by Timmy 2026-03-24 19:33:45 +00:00
Timmy closed this issue 2026-03-24 21:54:40 +00:00
Sign in to join this conversation.
No Label gemini-review
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#1008