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.

This paper introduces two extensions to the popular PyTorch machine learning framework, TorchDynamo and TorchInductor, which implement the torch.compile feature released in PyTorch 2. TorchDynamo is a Python-level just-in-time (JIT) compiler that enables graph compilation in PyTorch programs without sacrificing the flexibility of Python. It achieves this by dynamically modifying Python bytecode before execution and extracting sequences of PyTorch operations into an FX graph, which is then JIT compiled using one of many extensible backends. TorchInductor is the default compiler backend for TorchDynamo, which translates PyTorch programs into OpenAI's Triton for GPUs and C++ for CPUs. Results show that TorchDynamo is able to capture graphs more robustly than prior approaches while adding minimal overhead, and TorchInductor is able to provide a 2.27× inference and 1.41× training geometric mean speedup on an NVIDIA A100 GPU across 180+ real-world models, which outperforms six other compilers. These extensions provide a new way to apply optimizations through compilers in eager mode frameworks like PyTorch.

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.

It is widely acknowledged that large models have the potential to deliver superior performance across a broad range of domains. Despite the remarkable progress made in the field of machine learning systems research, which has enabled the development and exploration of large models, such abilities remain confined to a small group of advanced users and industry leaders, resulting in an implicit technical barrier for the wider community to access and leverage these technologies. In this paper, we introduce PyTorch Fully Sharded Data Parallel (FSDP) as an industry-grade solution for large model training. FSDP has been closely co-designed with several key PyTorch core components including Tensor implementation, dispatcher system, and CUDA memory caching allocator, to provide non-intrusive user experiences and high training efficiency. Additionally, FSDP natively incorporates a range of techniques and settings to optimize resource utilization across a variety of hardware configurations. The experimental results demonstrate that FSDP is capable of achieving comparable performance to Distributed Data Parallel while providing support for significantly larger models with near-linear scalability in terms of TFLOPS.

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.