mirror of
https://github.com/hanzoai/engine.git
synced 2026-11-20 14:17:52 +00:00
- Demand-sized unified-memory KV cache: size to max_seq_len*max_batch_size (context*concurrency) bounded by the post-model/OS-reserve ceiling, instead of grabbing all free RAM. Fixes the ROCm/APU thrash where an ~88 GB grab on a 128 GB unified box never returned; 30B-A3B now 88 GB -> 24 GB, elastic with concurrency. (hanzo-engine/src/paged_attention/mod.rs) - Vulkan paged-attention wired (is_vulkan branch, opt-in via --paged-attn on): runs and eliminates the naive-Sdpa layout-copy churn, but the v1 scalar kernel regresses decode on this APU, so default OFF until v2 + MoE-glue fusion land. i64/i32 paged-attn index metadata now binds directly (backend coerces to u32). - hanzo-quant: Vulkan/Wgpu storage arms in the distributed pipeline. - Drop the temporary [patch.crates-io] hanzo-ml redirect; pin hanzo-ml 0.11.55. - Manifest hygiene: version on path deps (enso, ffi, server-core, zen3-serving).
23 lines
821 B
TOML
23 lines
821 B
TOML
[package]
|
|
name = "hanzo-engine-ffi"
|
|
version = "0.1.0"
|
|
edition = "2021"
|
|
publish = false
|
|
description = "C ABI over the Hanzo native inference engine, for cgo/FFI callers (Go EVM precompiles)."
|
|
|
|
[lib]
|
|
name = "hanzo_engine_ffi"
|
|
# cdylib for cgo dynamic linking; staticlib for fully-static embedding.
|
|
crate-type = ["cdylib", "staticlib"]
|
|
|
|
[dependencies]
|
|
hanzo-engine = { version = "1.7.41", path = "../hanzo-engine" }
|
|
hanzo-server-core = { version = "1.7.41", path = "../hanzo-server-core", default-features = false }
|
|
tokio = { workspace = true }
|
|
|
|
[features]
|
|
# Forward GPU/CPU backends to the engine. The Go bench builds with `accelerate`.
|
|
accelerate = ["hanzo-engine/accelerate", "hanzo-server-core/accelerate"]
|
|
metal = ["hanzo-engine/metal", "hanzo-server-core/metal"]
|
|
cuda = ["hanzo-engine/cuda", "hanzo-server-core/cuda"]
|