paper-with-me

Papers

Ember: A Compiler for Efficient Embedding Operations on Decoupled Access-Execute Architectures

2025-04-14 · Marco Siracusa, Olivia Hsu, Victor Soria-Pardos, Joshua Randall, Arnaud Grasset, Eric Biscondi, Doug Joseph, Randy Allen, Fredrik Kjolstad, Miquel Moretó Planas, Adrià Armejach

Irregular embedding lookups are a critical bottleneck in recommender models, sparse large language models, and graph learning models. In this paper, we first demonstrate that, by offloading these lookups to specialized access units, Decoupled Access-Execute (DAE) processors achieve 2.6$\times$ higher performance and 6.4$\times$ higher performance/watt than GPUs on end-to-end models. Then, we propose the Ember compiler for automatically generating optimized DAE code from PyTorch and TensorFlow. Conversely from other DAE compilers, Ember features multiple intermediate representations specifically designed for different optimization levels. In this way, Ember can implement all optimizations to match the performance of hand-written code, unlocking the full potential of DAE architectures at scale.

📄 PDF Abstract BibTeX arXiv:2504.09870

Code (0)

등록된 구현이 없습니다.

Tasks

Graph Learning

Similar Papers 제목 키워드 기반

TapirXLA: Embedding Fork-Join Parallelism into the XLA Compiler in TensorFlow Using Tapir

2019-08-29 · Tao B. Schardl, Siddharth Samsi

This work introduces TapirXLA, a replacement for TensorFlow's XLA compiler that embeds recursive fork-join parallelism into XLA's low-level representation of code. Machine-learning applications rely on efficient parallel…

BIG-bench Machine LearningCPU

PowerFusion: A Tensor Compiler with Explicit Data Movement Description and Instruction-level Graph IR

2023-07-11 · Zixuan Ma, Haojie Wang, Jingze Xing, Liyan Zheng 외

Deep neural networks (DNNs) are of critical use in different domains. To accelerate DNN computation, tensor compilers are proposed to generate efficient code on different domain-specific accelerators. Existing tensor com…

GPU

oneDNN Graph Compiler: A Hybrid Approach for High-Performance Deep Learning Compilation

2023-01-03 · Jianhui Li, Zhennan Qin, Yijie Mei, Jingze Cui 외

With the rapid development of deep learning models and hardware support for dense computing, the deep learning workload characteristics changed significantly from a few hot spots on compute-intensive operations to a broa…

Code GenerationCompiler OptimizationDeep Learning

ML-Triton, A Multi-Level Compilation and Language Extension to Triton GPU Programming

2025-03-19 · Dewei Wang, Wei Zhu, Liyang Ling, Ettore Tiotto 외

In the era of LLMs, dense operations such as GEMM and MHA are critical components. These operations are well-suited for parallel execution using a tilebased approach. While traditional GPU programming often relies on low…

GPU

Automatic Compiler Based FPGA Accelerator for CNN Training

2019-08-15 · Shreyas Kolala Venkataramanaiah, Yufei Ma, Shihui Yin, Eriko Nurvithadhi 외

Training of convolutional neural networks (CNNs)on embedded platforms to support on-device learning is earning vital importance in recent days. Designing flexible training hard-ware is much more challenging than inferenc…