Preprint
Aug 2026
Rollplex: Cross-Phase GPU Spatial Sharing for Vision Language Model Post-Training
Rollplex is presented, a runtime that decomposes the reference and training phase and moves the prefix computation into the rollout decode window and achieves speedup over serial colocation and disaggregation under the same GPU budget, while preserving the synchronous RL update.
Han-Feng Lu, Tian-Yu Feng, Su-Yi Li et al.
· 1 citation