Lightning ⚡ Intel Habana

Habana® Gaudi® AI Processor (HPU) training processors are built on a heterogeneous architecture with a cluster of fully programmable Tensor Processing Cores (TPC) along with its associated development tools and libraries, and a configurable Matrix Math engine.

The TPC core is a VLIW SIMD processor with an instruction set and hardware tailored to serve training workloads efficiently. The Gaudi memory architecture includes on-die SRAM and local memories in each TPC and, Gaudi is the first DL training processor that has integrated RDMA over Converged Ethernet (RoCE v2) engines on-chip.

On the software side, the PyTorch Habana bridge interfaces between the framework and SynapseAI software stack to enable the execution of deep learning models on the Habana Gaudi device.

Gaudi offers a substantial price/performance advantage -- so you get to do more deep learning training while spending less.

For more information, check out Gaudi Architecture and Gaudi Developer Docs.

Installation

pip install -U lightning-habana

Usage

To enable PyTorch Lightning to utilize the HPU accelerator, simply provide accelerator="hpu" parameter to the Trainer class.

from lightning import Trainer

# run on as many Gaudi devices as available by default
trainer = Trainer(accelerator="auto", devices="auto", strategy="auto")
# equivalent to
trainer = Trainer()

# run on one Gaudi device
trainer = Trainer(accelerator="hpu", devices=1)
# run on multiple Gaudi devices
trainer = Trainer(accelerator="hpu", devices=8)
# choose the number of devices automatically
trainer = Trainer(accelerator="hpu", devices="auto")

The devices>1 parameter with HPUs enables the Habana accelerator for distributed training. It uses HPUParallelStrategy internally which is based on DDP strategy with the addition of Habana's collective communication library (HCCL) to support scale-up within a node and scale-out across multiple nodes.