Refactor GPU backend planner#117

Merged

orionpapadakis merged 18 commits into

mainfrom

refactor/tornadovm-planning

Jun 9, 2026

Commits on May 28, 2026

[prf/dec]Implement prefill-decode for Llama Q8_0
orionpapadakis
committed
Reorganize TornadoVM execution planning and improve naming conventions
orionpapadakis
committed
Update naming from `ActivationGraph` to `ActivationTaskGraph` across TornadoVM components
orionpapadakis
committed
Rename `AbstractFFNLayers` to `AbstractTransformerLayerTaskGraphs` and `AbstractLogitsLayer` to `AbstractLogitsTaskGraph`, updating all references to improve clarity and align with naming conventions.
orionpapadakis
committed
Refactor FFN layer comments to `transformer-layer task graphs`, aligning with updated naming conventions.
orionpapadakis
committed
[ci] Add workflows for Llama-3.2-1B-Instruct Q8_0 inference with prefill-decode and CUDA-graph variants
orionpapadakis
committed

Commits on Jun 5, 2026

Commits on Jun 8, 2026

[prf/dec] Conform to formatter
orionpapadakis
committed

Commits on Jun 9, 2026

Merge remote-tracking branch 'upstream/main' into refactor/tornadovm-planning
orionpapadakis
committed