Skip to content

Tier 4: VM and GPU kernel capacity limits#34

Merged
kavanaghpatrick merged 6 commits intomainfrom
feat/tier4-capacity
Mar 8, 2026
Merged

Tier 4: VM and GPU kernel capacity limits#34
kavanaghpatrick merged 6 commits intomainfrom
feat/tier4-capacity

Conversation

@kavanaghpatrick
Copy link
Copy Markdown
Owner

Summary

  • GPU memory cells: 32 → 128 (supports complex transpiled programs with many variables)
  • GPU call stack depth: 16 → 32 (supports deeper recursion like fib(20))
  • GPU stack depth: 128 → 256 (supports save/restore around recursive calls with many locals)
  • Variable emoji pool: 50 → 258 (supports programs with 200+ unique variables)
  • Function emoji pool: 21 → 51 (supports programs with 50+ function definitions)
  • Added collision validation at module load time
  • 866 tests passing (7 new tests added)

Closes #30

Test plan

  • GPU kernel has NUM_MEMORY_CELLS = 128
  • GPU kernel has CALL_STACK_DEPTH = 32
  • Default stack depth is 256
  • Variable pool has 258 unique emoji, no duplicates
  • Function pool has 51 unique emoji, no cross-collision with variable pool
  • Program with 100+ variables transpiles and runs correctly
  • Program with 30+ functions transpiles and runs correctly
  • Full regression: 866 tests, 0 failures

🤖 Generated with Claude Code

claude added 6 commits March 8, 2026 17:19
Spec artifacts:
- research.md: feasibility analysis and codebase exploration
- requirements.md: user stories and acceptance criteria
- design.md: architecture and technical decisions
- tasks.md: POC-first implementation plan

Ready for implementation.
…idation

Update bytecode.py docstrings to reflect 256 stack depth (was 128).
Add _validate_emoji_pools() in transpiler.py that checks for duplicates
and cross-pool collisions at module load time.

Co-Authored-By: Claude Opus 4.6 <[email protected]>
@kavanaghpatrick kavanaghpatrick merged commit a631cbd into main Mar 8, 2026
4 checks passed
@kavanaghpatrick kavanaghpatrick deleted the feat/tier4-capacity branch March 8, 2026 09:31
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.

Tier 4: VM and GPU kernel capacity limits

2 participants