No description
Find a file
Zachery Aaron Shores-Chmielewski 187c81498a feat: per-stage GGUF weight sharding and deploy hardening
Distribute only each stage's GGUF layer slice over HTTP, add sampler and weight-load health telemetry, and harden node-image build, orchestrator provisioning, and the VastAI lease/search path.

- gguf_shard (new): StageShardPlan and plan_stage_shard parse the GGUF directory and compute coalesced per-stage tensor byte ranges; materialize_stage_shard_http fetches only those ranges (plus the header) to build a stage-local GGUF, with planned_fetch_bytes accounting.
- orchestrator_app: build a BTreeMap<u32, StageShardPlan> from the run plan for HuggingFace sources, thread stage_shard_plan through StageProvisionWire and weight-load, emit stage_shard_plan summaries, and add liveness phases (prefetching/fetching_stage_shard, cache_ready, stage_shard_ready).
- worker_node: add a stage-shard-fetcher subcommand and materialize_stage_shard_with_process that spawns the fetcher, streams its stdout/stderr as stage_shard_fetch events (StageShardCacheReady/StageShardReady), caches under MVP_MODEL_CACHE_DIR, and feeds the local shard path into load_weights.
- worker_node: add NODE_SAMPLER_CHANNEL and SamplerHealth telemetry (gpu/cpu/net samplers emit started/waiting/ready/failed) plus structured helper stdout/stderr streaming (wait_for_helper_event/drain_worker_stderr).
- node_image: expand node-image build/push handling for the deploy path.
- tools/vastai: extend lease, search, and types and drop unused pricing code.

Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-07-26 13:01:18 +04:00
.cargo feat(mvp-system): working mvp-chat over edge transport 2026-07-22 11:50:53 +04:00
apps/mvp-node feat: successful 8 stage pipeline parallel run, more metrics 2026-07-25 20:04:57 +04:00
crates feat: per-stage GGUF weight sharding and deploy hardening 2026-07-26 13:01:18 +04:00
src feat: working two stage pipeline parallel over vastai 2026-07-24 12:59:32 +04:00
tests feat: actor direct control via runtime admin 2026-07-08 19:23:03 +04:00
tools/vastai feat: per-stage GGUF weight sharding and deploy hardening 2026-07-26 13:01:18 +04:00
xtask feat: per-stage GGUF weight sharding and deploy hardening 2026-07-26 13:01:18 +04:00
.dockerignore refactor: retire old pipeline-parallel app, restructure mvp-system 2026-07-07 14:40:02 +04:00
.gitignore feat: verification gate for N pipeline stage deployment 2026-07-25 11:42:29 +04:00
Cargo.lock feat: per-stage GGUF weight sharding and deploy hardening 2026-07-26 13:01:18 +04:00
Cargo.toml feat: add mvp provisioning subsystem and datastream transport 2026-06-25 16:30:18 +04:00
README.md feat: add mvp provisioning subsystem and datastream transport 2026-06-25 16:30:18 +04:00

swactor

Minimal actor runtime for Rust. One trait, one message type. Single-threaded (tick()) or multi-threaded (run()).

Description

Core runtime is src/. Actors implement ActorInterface (in actor.rs), interact through Ctx (in runtime.rs), and run on worker threads (worker.rs).

crates/ builds upward: std adds OTP patterns (supervision, monitoring, groups), distribution adds clustering, everything else composes from there.

Dev commands

cargo xtask --help for the basic test command.

Testing

cargo check --workspace
cargo xtask test