42e13803c7
ZERO host downloads in any GpuTensor operation. All ops GPU-native: elementwise.rs (NEW): 7 CUDA kernels compiled via compile_ptx_for_device - elementwise_binary: add/sub/mul/div (parameterized op) - elementwise_unary: powf/sqr/floor/relu/clamp (parameterized op) - broadcast_scalar_binary: scalar broadcast with any op - broadcast_row_binary: row broadcast [1,N] op [M,N] - expand_broadcast: general N-D broadcast with GPU stride tables - transpose_2d: [rows,cols] → [cols,rows] - gather_select: index_select along any dimension gpu_tensor.rs: 18 methods rewritten from host-roundtrip to GPU-native - gpu_clone: cuMemcpyDtoDAsync (was DtoH+HtoD) - add/sub/mul: elementwise_binary kernel (was CPU zip) - broadcast_mul/div: broadcast kernels (was CPU map) - narrow(dim=0): DtoD slice view (was CPU slice) - narrow(dim>0): gather_select kernel - argmax: ReductionKernels (was CPU scan) - mean_all/sum_all: ReductionKernels (was CPU sum) - powf/sqr/floor/relu/clamp: elementwise_unary kernel - transpose: transpose_2d kernel - expand/broadcast_as: expand_broadcast kernel - index_select: gather_select kernel 3 reduction tests have init bug (max not -INF) — fix follows. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Foxhunt
Production HFT trading system in Rust.
Architecture
The workspace contains 32 crates organized as follows:
Core Libraries (16)
| Crate | Purpose |
|---|---|
trading_engine |
Order processing, FIX 4.4, IB TWS, SIMD, RDTSC timing |
risk |
VaR, Kelly, circuit breakers, kill switches, compliance |
risk-data |
Risk data types and shared structures |
trading-data |
Trading data types |
ml |
DQN Rainbow, PPO, TFT, Mamba2, ensemble inference |
ml-data |
ML data types and feature definitions |
data |
Market data ingestion and storage |
backtesting |
Replay engine, strategy tester |
adaptive-strategy |
Ensemble execution, microstructure analysis |
common |
Shared types, resilience, error handling |
storage |
S3 and local model storage |
model_loader |
Model serialization and loading |
market-data |
Market data feed handlers |
database |
PostgreSQL access layer (SQLx) |
config |
Configuration management |
tli |
CLI commands and tooling |
Services (8)
| Service | Purpose |
|---|---|
backtesting_service |
gRPC backtesting service |
broker_gateway_service |
FIX routing, broker connectivity |
trading_service |
Core trading operations |
ml_training_service |
Model training orchestration |
data_acquisition_service |
Market data acquisition |
trading_agent_service |
Autonomous trading agents |
api_gateway |
gRPC API gateway with auth |
web-gateway |
Axum REST + WebSocket gateway |
Frontend
web-dashboard/ -- React 19 + TypeScript + Vite + TradingView charts.
Building
# Check compilation (no PostgreSQL required)
SQLX_OFFLINE=true cargo check --workspace
# Run tests for a specific crate
SQLX_OFFLINE=true cargo test -p <crate> --lib
# Clippy
SQLX_OFFLINE=true cargo clippy --workspace
ML Models
Four production model architectures on Candle v0.9.1 with CUDA:
- DQN Rainbow -- Deep Q-Network with prioritized replay, dueling heads, noisy nets
- PPO -- Proximal Policy Optimization with GAE and LSTM policies
- TFT -- Temporal Fusion Transformer for multi-horizon forecasting
- Mamba2 -- State space model for sequence prediction
Each model has a standalone trainer and a UnifiedTrainable adapter for the hyperopt pipeline.
Infrastructure
- Git: Gitea at
git.fxhnt.ai(Tailscale-only), Scaleway DEV1-S - Observability: OpenTelemetry OTLP (env
OTEL_EXPORTER_OTLP_ENDPOINT) - Database: PostgreSQL with SQLx offline mode for CI
License
Proprietary. All rights reserved.