What makes it fast

• Hybrid level-batched growth — whole levels of sibling pairs are scored, synchronized and applied in one launch each instead of per split, turning a latency-bound loop into a throughput-bound one, with leaf-wise-identical trees.

• CUDA-graph level loops — the per-level launch sequence is captured once and replayed by a device-side controller, removing host round-trips.

• NVRTC runtime JIT — construct kernels are specialized at runtime to the actual data shape, self-tested against the ahead-of-time kernel, and promoted only if bit-identical.

• Per-tree compact column view — with any feature_fraction < 1, only the sampled columns are materialized for histogram construction; the win scales with the excluded fraction (~3.4× end-to-end at feature_fraction = 0.1 on wide, low-cardinality data).

• GPU-native dataset construction — dense binning, row-data build and EFB pre-checking run on the device; CuPy and __cuda_array_interface__ inputs are ingested without a host round-trip.

• Quantized training, two ways — quant_mode=stochastic is the speed end: 4-bin gradients with seeded stochastic rounding; quant_mode=fixedpoint is the near-lossless end: deterministic rounding with an internal outlier-robust gradient scale. Bin counts are overridable with quant_bins; both modes are bit-reproducible run to run.

• GPU inference via NVIDIA FIL — with cuML installed, Booster.predict() transparently runs on the Forest Inference Library; CuPy arrays stay on the device end to end.

• An execution planner — shape-conditional kernel choices are resolved once from the data and parameters (cuda_plan=auto), every decision guaranteed bit-identical and individually overridable.

Every optimization above is required to be bit-identical to the reference path, and that is enforced mechanically: a 38-cell regression lattice of (config × data-shape) training cells fingerprinted by model md5 runs on a real GPU on every commit, alongside plan-flip equality cells, validity assertions, metric floors and a perf gate.

Install

pip install falcata

That builds the CUDA library from source, so you need the CUDA toolkit (>= 11.0), CMake >= 3.28, a C++17 compiler and Python >= 3.10. Nothing else: the source distribution vendors every dependency, so no git clone and no submodule dance.

The build detects the GPU(s) in the machine and compiles only for those (15–20 minutes on typical hardware). When no GPU is visible at build time it targets every architecture the toolkit supports instead, which takes several times longer. To pick an architecture explicitly:

# RTX 5090 = 120, RTX 4090 = 89, A100 = 80, T4 = 75
pip install falcata --config-settings=cmake.define.CMAKE_CUDA_ARCHITECTURES=89

There is a CPU build, though it is not what this library is for:

pip install falcata --config-settings=cmake.define.USE_CUDA=OFF

Multi-GPU training additionally needs NCCL and its headers (libnccl-dev on Debian/Ubuntu):

pip install falcata \
  --config-settings=cmake.define.USE_NCCL=ON \
  --config-settings=cmake.define.BUILD_WITH_SHARED_NCCL=ON

Installing the wheel also installs a lightgbm import shim. If the target environment already has stock LightGBM, uninstall it first or use a fresh environment — pip will not report the collision.

Quick start

import falcata as flc

ds = flc.Dataset(X_train, label=y_train, params={"device_type": "cuda"})
model = flc.train(
    {
        "objective": "regression",
        "device_type": "cuda",
        "num_leaves": 255,
        "quant_mode": "stochastic",   # none | stochastic | fixedpoint
        "cuda_precision": "fp32",     # fp64 (default) | fp32
        "cuda_plan": "auto",          # the planner picks the kernels
    },
    ds,
    num_boost_round=1000,
)

Compatibility with LightGBM

Falcata began as a fork of LightGBM and deliberately stays interoperable at the data boundaries:

• Models written by Falcata load in stock LightGBM (and vice versa) with bit-identical predictions — verified in CI.

• Binary datasets (.dataset) interchange in both directions.

• Parameter names are unchanged; Falcata’s additions (quant_mode, cuda_precision, cuda_plan) are new names upstream simply ignores.

• The historical LGBM_* C API names remain as aliases for FLC_*. The Python package is import falcata only — code written against import lightgbm needs its import changed, nothing else.

License

MIT. Falcata derives from LightGBM (copyright Microsoft Corporation and the LightGBM developers, MIT); that copyright is retained. Falcata is not affiliated with, endorsed by, or supported by Microsoft or the LightGBM maintainers.