Last updated: 2026-06-12

Inference Runtime

Bench Hot Decode

All API Sections

Hot-path decode kernel microbenchmarks.

Measures per-dispatch GPU latency, memory bandwidth, and VRAM utilisation for individual compute kernels (DMMV, SSM delta-net) in isolation. Run via `zig build hot-bench -Doptimize=ReleaseFast`.

1 exports 0 methods src/bench_hot_decode.zig

1 exports shown

function

main

#
pub fn main() !void

Run the hot-decode microbenchmark suite against the selected model and kernels.

src/bench_hot_decode.zig:768