{
  "execution_mode": "CPU algorithm with controlled inputs; any latency/capacity coefficients are assumptions",
  "pool_config": {
    "total_token_capacity": 4096,
    "max_seq_len": 512,
    "block_size": 16,
    "submitted_requests": 32
  },
  "contiguous": {
    "admitted_requests": 8,
    "rejected_requests": 24,
    "active_sequences": 8,
    "reserved_tokens": 4096,
    "used_tokens": 758,
    "wasted_tokens": 3338,
    "waste_ratio": 0.8149,
    "utilization_of_total_pool": 0.1851
  },
  "paged": {
    "admitted_requests": 32,
    "rejected_requests": 0,
    "active_sequences": 32,
    "allocated_blocks": 199,
    "free_blocks": 57,
    "reserved_tokens": 3184,
    "physical_stored_tokens": 2942,
    "logical_tokens_served": 2942,
    "internal_fragmentation_tokens": 242,
    "waste_ratio": 0.076,
    "cow_copies_performed": 0
  },
  "concurrency_multiplier": 4.0,
  "copy_on_write_demo": {
    "prompt_tokens": 40,
    "shared_beams": 3,
    "blocks_before_decode_append": 3,
    "active_sequences": 3,
    "allocated_blocks": 5,
    "free_blocks": 27,
    "reserved_tokens": 80,
    "physical_stored_tokens": 59,
    "logical_tokens_served": 123,
    "internal_fragmentation_tokens": 21,
    "waste_ratio": 0.2625,
    "cow_copies_performed": 2,
    "beam_tails": {
      "beam_0_last_token": 9000,
      "beam_1_last_token": 9001,
      "beam_2_last_token": 9002
    }
  }
}
