The rapidly evolving landscape of products, surfaces, policies, and regulations poses significant challenges for deploying state-of-the-art recommendation models at industry scale, primarily due to data fragmentation across domains and escalating infrastructure costs that hinder sustained quality improvements.

This paper presents a comprehensive evaluation of microscaling (MX) quantization in the pre-training of large language models (LLMs), investigating its potential to enhance the computation and memory efficiencies. We systematically examine the effects of key design parameters - including data types, rounding modes, scaling strategies, granularity, and organization - on numerical accuracy and training stability. Our extensive experimental study on Llama3 models reveals critical insights into the challenges of 4-bit training for LLMs and identifies optimal configurations with mixed precisions of 4-bit and 6-bit MX formats that significantly enhance training quality, bridging the gap with higher-precision formats. This research provides valuable guidance on the benefits and limitations of MX quantization, laying the groundwork for future innovations in low-precision LLM training.

The rapid growth of AI workloads at Meta has motivated our inhouse development of AI chips, aiming to significantly reduce the total cost of ownership and mitigate risks posed by unpredictable GPU supplies.At ISCA'23, we presented Meta's first-generation AI chip, MTIA 1.This paper describes its successor, MTIA 2i, now deployed at scale and serving billions of users.MTIA 2i significantly improves upon MTIA 1, reducing total cost of ownership by 44% compared to GPUs while delivering competitive performance per watt.A key differentiator is its memory hierarchy: instead of costly HBM, it uses large SRAM alongside LPDDR.Although there has been a proliferation of publications on AI chips, they often focus on architectural design and overlook three critical aspects:(1) co-designing and optimizing ML models to work effectively with the AI chip; (2) demonstrating sufficient flexibility to support a wide range of models; and (3) during the productionization process, addressing challenges unanticipated or decisions deferred at design time, such as dealing with memory errors, safe overclocking, reducing provisioned power, and implementing real-time firmware updates to mitigate silicon design defects.A key contribution of this paper is sharing our experience with these aspects, based on our journey of productionizing MTIA 2i at scale.

We present DCPerf, the first open-source performance benchmark suite actively used to inform procurement decisions for millions of CPU in hyperscale datacenters.Although numerous benchmarks exist, our evaluation reveals that they inaccurately project server performance for datacenter workloads or fail to scale to resemble production workloads on modern many-core servers.DCPerf distinguishes itself in two aspects: (1) it faithfully models essential software architectures and features of datacenter applications, such as microservice architecture and highly optimized multi-process or multi-thread concurrency; and (2) it strives to align its performance characteristics with those of production workloads, at both the system level and microarchitecture level.Both are made possible by our direct access to the source code and hyperscale production deployments of datacenter workloads.Additionally, we share real-world examples of using DCPerf in critical decision-making, such as selecting future CPU SKUs and guiding CPU vendors in optimizing their designs.Our evaluation demonstrates that DCPerf accurately projects the performance of representative production workloads within a 3.3% error margin across four generations of production servers introduced over a span of six years, with core counts varying widely from 36 to 176.

Scaling laws play an instrumental role in the sustainable improvement in model quality. Unfortunately, recommendation models to date do not exhibit such laws similar to those observed in the domain of large language models, due to the inefficiencies of their upscaling mechanisms. This limitation poses significant challenges in adapting these models to increasingly more complex real-world datasets. In this paper, we propose an effective network architecture based purely on stacked factorization machines, and a synergistic upscaling strategy, collectively dubbed Wukong, to establish a scaling law in the domain of recommendation. Wukong’s unique design makes it possible to capture diverse, any-order of interactions simply through taller and wider layers. We conducted extensive evaluations on six public datasets, and our results demonstrate that Wukong consistently outperforms state-of-the-art models quality-wise. Further, we assessed Wukong’s scalability on an internal, large-scale dataset. The results show that Wukong retains its superiority in quality over state-of-the-art models, while holding the scaling law across two orders of magnitude in model complexity, extending beyond 100 GFLOP/example, where prior arts fall short.

Meta has traditionally relied on using CPU-based servers for running inference workloads, specifically Deep Learning Recommendation Models (DLRM), but the increasing compute and memory requirements of these models have pushed the company towards using specialized solutions such as GPUs or other hardware accelerators. This paper describes the company's effort in constructing its first silicon specifically designed for recommendation systems; it describes the accelerator architecture and platform design, the software stack for enabling and optimizing PyTorch-based models and provides an initial performance evaluation. With our emerging software stack, we have made significant progress towards reaching the same or higher efficiency as the GPU: We averaged 0.9x perf/W across various DLRMs, and benchmarks show operators such as GEMMs reaching 2x perf/W. Finally, the paper describes the lessons we learned during this journey which can improve the performance and programmability of future generations of architecture.

This paper introduces Block Data Representations (BDR), a framework for exploring and evaluating a wide spectrum of narrow-precision formats for deep learning. It enables comparison of popular quantization standards, and through BDR, new formats based on shared microexponents (MX) are identified, which outperform other state-of-the-art quantization approaches, including narrow-precision floating-point and block floating-point. MX utilizes multiple levels of quantization scaling with ultra-fine scaling factors based on shared microexponents in the hardware. The effectiveness of MX is demonstrated on real-world models including large-scale generative pretraining and inferencing, and production-scale recommendation systems.

Deep Learning Recommendation Models (DLRM) are widespread, account for a considerable data center footprint, and grow by more than 1.5x per year. With model size soon to be in terabytes range, leveraging Storage Class Memory (SCM) for inference enables lower power consumption. This paper evaluates the major challenges in extending the memory hierarchy to SCM for DLRM, and presents different techniques to improve performance through a Software Defined Memory. We show how underlying technologies such as Nand Flash and 3DXP differentiate, and relate to real world scenarios, enabling from 5% to 29% power savings.

Deep learning recommendation models (DLRMs) have been used across many business-critical services at Meta and are the single largest AI application in terms of infrastructure demand in its data-centers. In this paper, we present Neo, a software-hardware co-designed system for high-performance distributed training of large-scale DLRMs. Neo employs a novel 4D parallelism strategy that combines table-wise, row-wise, column-wise, and data parallelism for training massive embedding operators in DLRMs. In addition, Neo enables extremely high-performance and memory-efficient embedding computations using a variety of critical systems optimizations, including hybrid kernel fusion, software-managed caching, and quality-preserving compression. Finally, Neo is paired with ZionEX, a new hardware platform co-designed with Neo's 4D parallelism for optimizing communications for large-scale DLRM training. Our evaluation on 128 GPUs using 16 ZionEX nodes shows that Neo outperforms existing systems by up to 40× for training 12-trillion-parameter DLRM models deployed in production.

Embedding representations power machine intelligence in many applications, including recommendation systems, but they are space intensive - potentially occupying hundreds of gigabytes in large-scale settings. To help manage this outsized memory consumption, we explore mixed dimension embeddings, an embedding layer architecture in which a particular embedding vector's dimension scales with its query frequency. Through theoretical analysis and systematic experiments, we demonstrate that using mixed dimensions can drastically reduce the memory usage, while maintaining and even improving the ML performance. Empirically, we show that the proposed mixed dimension layers improve accuracy by 0.1 % using half as many parameters or maintain it using 16 x fewer parameters for click-through rate prediction on the Criteo Kaggle dataset. They also train over 2x faster on a GPU. A full version of this paper is accessible at: https://arxiv.org/abs/1909.11810

The widespread application of deep learning has changed the landscape of computation in data centers. In particular, personalized recommendation for content ranking is now largely accomplished using deep neural networks. However, despite their importance and the amount of compute cycles they consume, relatively little research attention has been devoted to recommendation systems. To facilitate research and advance the understanding of these workloads, this paper presents a set of real-world, production-scale DNNs for personalized recommendation coupled with relevant performance metrics for evaluation. In addition to releasing a set of open-source workloads, we conduct in-depth analysis that underpins future system design and optimization for at-scale recommendation: Inference latency varies by 60% across three Intel server generations, batching and co-location of inference jobs can drastically improve latency-bounded throughput, and diversity across recommendation models leads to different optimization strategies.

Personalized recommendation systems leverage deep learning models and account for the majority of data center AI cycles. Their performance is dominated by memory-bound sparse embedding operations with unique irregular memory access patterns that pose a fundamental challenge to accelerate. This paper proposes a lightweight, commodity DRAM compliant, near-memory processing solution to accelerate personalized recommendation inference. The in-depth characterization of production-grade recommendation models shows that embedding operations with high model-, operator and data-level parallelism lead to memory bandwidth saturation, limiting recommendation inference performance. We propose RecNMP which provides a scalable solution to improve system throughput, supporting a broad range of sparse embedding models. RecNMP is specifically tailored to production environments with heavy co-location of operators on a single server. Several hardware/software cooptimization techniques such as memory-side caching, tableaware packet scheduling, and hot entry profiling are studied, providing up to 9.8× memory latency speedup over a highly-optimized baseline. Overall, RecNMP offers 4.2× throughput improvement and 45.8% memory energy savings.

In this tutorial we show how to build deep learning recommendation systems and resolve the associated interpretability, integrity and privacy challenges. We start with an overview of the PyTorch framework, features that it offers and a brief review of the evolution of recommendation models. We delineate their typical components and build a proxy deep learning recommendation model (DLRM) in PyTorch. Then, we discuss how to interpret recommendation system results as well as how to address the corresponding integrity and quality challenges.

Modern deep learning-based recommendation systems exploit hundreds to thousands of different categorical features, each with millions of different categories ranging from clicks to posts. To respect the natural diversity within the categorical data, embeddings map each category to a unique dense representation within an embedded space. Since each categorical feature could take on as many as tens of millions of different possible categories, the embedding tables form the primary memory bottleneck during both training and inference. We propose a novel approach for reducing the embedding size in an end-to-end fashion by exploiting complementary partitions of the category set to produce a unique embedding vector for each category without explicit definition. By storing multiple smaller embedding tables based on each complementary partition and combining embeddings from each table, we define a unique embedding for each category at smaller cost. This approach may be interpreted as using a specific fixed codebook to ensure uniqueness of each category's representation. Our experimental results demonstrate the effectiveness of our approach over the hashing trick for reducing the size of the embedding tables in terms of model loss and accuracy, while retaining a similar reduction in the number of parameters.