Struct TargetDefaults 

pub struct TargetDefaults {

    pub hw: &'static str,
    pub lm_head_batchm_max: u32,
    pub ssm_batched_recurrent: bool,
    pub gdn_prefill_tc: bool,
    pub ssm_ba_gates_hopper: bool,
    pub fp8_act_quant_hopper: bool,
    pub decode_split_silu: bool,
    pub attn_decode_splitk: &'static str,
    pub ffn_m16_tc: bool,
    pub attn_m16_tc: bool,
    pub lm_head_m16_tc: bool,
    pub attn_ncol_gemv: bool,
    pub ffn_gateup_fused: bool,
    pub w8a8_prefill_max_m_widening: u32,
    pub w8a8_prefill_max_m_narrowing: u32,
}

2026-09-25: One compiled target’s serving defaults. A const that build.rs writes into OUT_DIR/target_ptx.rs and lib.rs include!s, so crate::TARGET_DEFAULTS is known at compile time.

Fields

hw: &'static str

2026-09-25: The kernels/<hw> directory this binary was built for (METRALE_TARGET_HW, else build_diagnose::DEFAULT_HW). It is set even when that tree has no HARDWARE.toml: read_defaults then returns baseline(hw).

lm_head_batchm_max: u32

2026-09-25: Upper edge of the BF16 decode head’s batched-GEMV band (metrale-model-engine model/trait_impl/lm_head_batched.rs). The resolver clamps it to DENSE_GEMV_BATCHM_MAX_M (resolve_batchm_max). hopper declares 16; gb10, b200 and b300 declare the baseline 8.

ssm_batched_recurrent: bool

2026-09-25: Batched recurrent launch on the GDN decode path (metrale-model-layers layers/qwen3_ssm/gdn_flags.rs). True on hopper.

gdn_prefill_tc: bool

2026-09-25: The tensor-core GDN chunked-prefill family, including the gated_delta_rule_chunk_delta_h_tcfuse_x2 state spine (kernels/hopper/common/gated_delta_rule_chunk_tc.cu), selected in metrale-model-layers layers/ops/ssm_gdn_a3.rs. True on hopper.

ssm_ba_gates_hopper: bool

2026-09-25: dense_gemm_ba_gates_prefill_hopper, one CTA per token, for the SSM BA projection and GDN gate transforms (metrale-model-layers layers/ops/ssm_ba_gates_hopper.rs). True on hopper. Its source exists only under kernels/hopper, so the row has no effect elsewhere. SSM-BA-GATES-ATTRIBUTION.md holds the measurements.

fp8_act_quant_hopper: bool

2026-09-25: per_token_group_quant_fp8_hopper (kernels/hopper/common/fp8_act_quant_hopper.cu) for the per-token FP8 activation quantizer. True on hopper. The twin takes a launch only when its grid reaches the CTA floor in metrale-model-layers layers/ops/fp8_act_quant_floor.rs. Its source exists only under kernels/hopper, so the row has no effect elsewhere. FP8-ACT-QUANT-ATTRIBUTION.md holds the measurements.

decode_split_silu: bool

2026-09-25: Split SiLU+down on the decode path (ModelLevers::decode_split_silu). True on every declaring target; METRALE_NO_DECODE_SPLIT_SILU turns it off.

attn_decode_splitk: &'static str

2026-09-25: How the paged-decode attention path picks its KV split count: legacy, auto or a pinned count, parsed by crate::attn_splitk::parse. hopper declares auto; every other target declares or defaults to legacy.

ffn_m16_tc: bool

2026-09-25: The w8a16_gemm_m16 tensor-core tier on the dense-FFN decode arm (metrale-model-layers layers/dense_ffn_m16_tc.rs). False on every target, hopper included; the hopper HARDWARE.toml records why.

attn_m16_tc: bool

2026-09-25: The w8a16_gemm_m16{,_strided} tensor-core tiers on the decode Q/K/V and o_proj projections. True on hopper.

lm_head_m16_tc: bool

2026-09-25: The dense_gemm_m16_bf16 tensor-core arm on the BF16 decode head (metrale-model-engine model/trait_impl/lm_head_batched.rs). True on hopper. It reassociates the K reduction relative to dense_gemv_bf16, so a near-tie argmax can change.

attn_ncol_gemv: bool

2026-09-25: w8a16_gemv_batch16_ncol{2,4} on the decode attention projections (metrale-model-layers layers/qwen3_attention/attn_ncol_gemv.rs). False on every target.

ffn_gateup_fused: bool

2026-09-25: One fused [gate | up] GEMM at N = 2 * intermediate instead of two at N = intermediate (metrale-model-layers layers/dense_ffn_gateup_fused.rs). True only on hopper; its strided-SiLU consumer (kernels/hopper/common/silu_mul_strided.cu) exists only in the hopper tree.

w8a8_prefill_max_m_widening: u32

2026-09-25: Upper M for the W8A8 block-scaled dense-FFN prefill on a widening projection (n > k: gate/up), read in metrale-model-layers layers/dense_ffn_w8a8_prefill.rs. u32::MAX means no cap. gb10 declares 64; its HARDWARE.toml records the measurement.

w8a8_prefill_max_m_narrowing: u32

2026-09-25: Upper M for the same path on a narrowing projection (n <= k: down). gb10 declares 384. See Self::w8a8_prefill_max_m_widening.

Trait Implementations

impl Clone for TargetDefaults

fn clone(&self) -> TargetDefaults

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TargetDefaults

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

Formats the value using the given formatter.

impl PartialEq for TargetDefaults

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for TargetDefaults

impl Eq for TargetDefaults

impl StructuralPartialEq for TargetDefaults