Ladder-residual: parallelism-aware architecture for accelerating large model inference with communication overlapping

TL;DR

This paper introduces Ladder Residual, an architectural modification that enables communication and computation overlap in large model inference, significantly improving speed by decoupling communication bottlenecks.

Contribution

It proposes Ladder Residual, a novel architectural approach applicable to residual models, that decouples communication from computation to accelerate large model inference, especially in tensor parallelism.

Findings

29% speedup in inference for 70B parameter Transformer with Ladder Residual

Comparable performance of Ladder Transformer to standard models at 1B and 3B scales

Minimal accuracy loss when converting parts of Llama-3.1 8B to Ladder Residual

Abstract

Large language model inference is both memory-intensive and time-consuming, often requiring distributed algorithms to efficiently scale. Various model parallelism strategies are used in multi-gpu training and inference to partition computation across multiple devices, reducing memory load and computation time. However, using model parallelism necessitates communication of information between GPUs, which has been a major bottleneck and limits the gains obtained by scaling up the number of devices. We introduce Ladder Residual, a simple architectural modification applicable to all residual-based models that enables straightforward overlapping that effectively hides the latency of communication. Our insight is that in addition to systems optimization, one can also redesign the model architecture to decouple communication from computation. While Ladder Residual can allow communication-computation decoupling in conventional parallelism patterns, we focus on Tensor Parallelism in this paper, which is particularly bottlenecked by its heavy communication. For a Transformer model with 70B parameters, applying Ladder Residual to all its layers can achieve 29% end-to-end wall clock speed up at inference time with TP sharding over 8 devices. We refer the resulting Transformer model as the Ladder Transformer. We train a 1B and 3B Ladder Transformer from scratch and observe comparable performance to a standard dense transformer baseline. We also show that it is possible to convert parts of the Llama-3.1 8B model to our Ladder Residual architecture with minimal accuracy degradation by only retraining for 3B tokens. We release our code for training and inference for easier replication of experiments.