arXiv subjects

Avinash Nayak

Publications and source records attributed to Avinash Nayak.

3 recordsLinked to original sources

arXiv2023

Software-Hardware Co-design for Fast and Scalable Training of Deep Learning Recommendation Models

Deep learning recommendation models (DLRMs) are used across many business-critical services at Facebook and are the single largest AI application in terms of infrastructure demand in its data-centers. In this paper we discuss the SW/HW co-designed solution for high-performance distributed training of large-scale DLRMs. We introduce a high-performance scalable software stack based on PyTorch and pair it with the new evolution of Zion platform, namely ZionEX. We demonstrate the capability to train very large DLRMs with up to 12 Trillion parameters and show that we can attain 40X speedup in terms of time to solution over previous systems. We achieve this by (i) designing the ZionEX platform with dedicated scale-out network, provisioned with high bandwidth, optimal topology and efficient transport (ii) implementing an optimized PyTorch-based training stack supporting both model and data parallelism (iii) developing sharding algorithms capable of hierarchical partitioning of the embedding tables along row, column dimensions and load balancing them across multiple workers; (iv) adding high-performance core operators while retaining flexibility to support optimizers with fully deterministic updates (v) leveraging reduced precision communications, multi-level memory hierarchy (HBM+DDR+SSD) and pipelining. Furthermore, we develop and briefly comment on distributed data ingestion and other supporting services that are required for the robust and efficient end-to-end training in production environments.

cs.DC↗

arXiv2021

First-Generation Inference Accelerator Deployment at Facebook

In this paper, we provide a deep dive into the deployment of inference accelerators at Facebook. Many of our ML workloads have unique characteristics, such as sparse memory accesses, large model sizes, as well as high compute, memory and network bandwidth requirements. We co-designed a high-performance, energy-efficient inference accelerator platform based on these requirements. We describe the inference accelerator platform ecosystem we developed and deployed at Facebook: both hardware, through Open Compute Platform (OCP), and software framework and tooling, through Pytorch/Caffe2/Glow. A characteristic of this ecosystem from the start is its openness to enable a variety of AI accelerators from different vendors. This platform, with six low-power accelerator cards alongside a single-socket host CPU, allows us to serve models of high complexity that cannot be easily or efficiently run on CPUs. We describe various performance optimizations, at both platform and accelerator level, which enables this platform to serve production traffic at Facebook. We also share deployment challenges, lessons learned during performance optimization, as well as provide guidance for future inference hardware co-design.

cs.AR↗

arXiv2015

Thermal Oxidation of WSe2 Nano-sheets Adhered on SiO2/Si Substrates

Due to the drastically different intralayer versus interlayer bonding strengths, the mechanical, thermal, and electrical properties of two-dimensional (2D) materials are highly anisotropic between the in-plane and out-of-plane directions. The structural anisotropy may also play a role in chemical reactions, such as oxidation, reduction, and etching. Here, the composition, structure, and electrical properties of mechanically exfoliated WSe2 nano- sheets on SiO2/Si substrates were studied as a function of the extent of thermal oxidation. A major component of the oxidation, as indicated from optical and Raman data, starts from the nano-sheet edges and propagates laterally towards the center. Partial oxidation also occurs in certain areas at the surface of the flakes, which are shown to be highly conductive by microwave impedance microscopy. Using secondary ion mass spectroscopy, we also observed extensive oxidation at the WSe2/SiO2 interface. The combination of multiple microcopy methods can thus provide vital information on the spatial evolution of chemical reactions on 2D materials and the nanoscale electrical properties of the reaction products.

cond-mat.mtrl-sci↗