Struct EngineCoreReadyResponse 

pub struct EngineCoreReadyResponse {

    pub max_model_len: i64,
    pub num_gpu_blocks: i64,
    pub prefix_granularity: i64,
    pub dp_stats_address: Option<String>,
    pub dtype: String,
    pub multimodal_encoder_dtype: Option<String>,
    pub vllm_version: String,
    pub world_size: i64,
    pub data_parallel_size: i64,
    pub tensor_parallel_size: i64,
    pub pipeline_parallel_size: i64,
    pub decode_context_parallel_size: i64,
    pub data_parallel_rank: i64,
    pub max_num_seqs: i64,
    pub max_num_batched_tokens: i64,
    pub instance_id: String,
    pub kv_cache_size_tokens: Option<i64>,
    pub kv_cache_max_concurrency: Option<f64>,
    pub kv_events_config: Option<Dict<Value, Value>>,
}

The engine’s configuration, sent when it registers on the input socket.

Field names, not order, are the contract; some names are the frontend’s vocabulary rather than TokenSpeed’s.

Fields

max_model_len: i64

Longest supported sequence.

num_gpu_blocks: i64

KV pages on the device.

prefix_granularity: i64

Prefix-cache matching granularity in tokens.

dp_stats_address: Option<String>

DP stats address, unused by TokenSpeed.

dtype: String

Model dtype: bfloat16, float16 or float32.

multimodal_encoder_dtype: Option<String>

Multimodal encoder dtype, if any.

vllm_version: String

tokenspeed-<version>; the frontend’s field name.

world_size: i64

Ranks in the engine.

data_parallel_size: i64

Attention-DP ranks.

tensor_parallel_size: i64

Attention-TP ranks.

pipeline_parallel_size: i64

Pipeline stages.

decode_context_parallel_size: i64

KV parallelism (--kv-parallel-size); the frontend’s field name.

data_parallel_rank: i64

This engine’s attention-DP rank.

max_num_seqs: i64

Running-request limit.

max_num_batched_tokens: i64

Tokens per step limit; 0 is no cap.

instance_id: String

Served model name.

kv_cache_size_tokens: Option<i64>

KV cache capacity in tokens.

kv_cache_max_concurrency: Option<f64>

KV cache capacity in maximum-length requests.

kv_events_config: Option<Dict<Value, Value>>

KV event publisher configuration.

Trait Implementations

impl Clone for EngineCoreReadyResponse

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for EngineCoreReadyResponse

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for EngineCoreReadyResponse

fn default() -> Self

Returns the “default value” for a type.

impl PartialEq for EngineCoreReadyResponse

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for EngineCoreReadyResponse

impl Wire for EngineCoreReadyResponse

fn encode(&self, encoder: &mut Encoder)

Append this value to a message.

fn decode(decoder: &mut Decoder<'_>) -> Result<Self>

Read one value of this type.