RTX 4080 sampled-k1 PG systems workload

This workload measures runtime and semantic conformance for verl-opd-v0.8-single-gpu-pg-k1-v1. The source profile pins verl v0.8.0 at 7aed6b230776f963fa09509c10d9c3a767d1102c.

Item	Measured value
Hardware	one NVIDIA RTX 4080, 15.992 GiB
Models	Qwen3-0.6B actor / Qwen3-1.7B teacher, pinned NF4 snapshots
Workload	32 distinct prompts, response bound 64, 8 strict updates
Target	sampled-token teacher log-probability; k1 + vanilla policy loss
Peak CUDA	2.3032 GiB allocated / 3.1914 GiB reserved
Time to first update	30.0902 s
Steady rollout	13.4693 generated tokens/s
Steady teacher scoring	893.86 sampled positions/s
Steady actor update	221.0899 positions/s

The matched run stopped after update 4, reconstructed the trainer from its checkpoint, and finished at update 8 with byte-identical trajectories, adapter and optimizer tensors and identical training-state fields (apart from the deliberately run-specific resolved-config digest). Every update recorded rollout policy version == parameter version before update; each successful update incremented the parameter version once.

The standard PEFT adapter exported with tokenizer metadata. A separate checksummed scale-out bundle preserved the sampled-k1 fields without a top-k requirement; exact student and teacher snapshots were materialized transactionally, the pinned upstream config merge and sequential CPU model/data load smoke passed, and the local bundle became launchable: true. No distributed verl job ran, and algorithm-wide parity is not claimed.

The complete machine-readable record is evidence/rtx4080-verl-pg-k1-v1.json (SHA-256 b549905b2b62b5e51d721ba2932befb7b910284ff1b4016095d341ec232a725f). The estimator/scalar/gradient contract is documented in ADR 0010.

Scope

The endpoints are local runtime, recovery and the documented scalar/gradient contract. Task quality and algorithm comparisons were not evaluated.