一句话：Atalanta：为定点量化 DNN 设计的无损张量压缩，在片外内存控制器前透明编解码，以极低开销接近算术编码效率。
定点量化深度神经网络推理受限于片上/片外存储容量与带宽。Atalanta 是一种无损、软硬件协同设计的张量压缩技术：在片外内存控制器前透明地压缩/解压权重与激活，兼容任意 ML 加速器。其编解码器各用不到 50B 内部存储，编码效率接近算术编码，却避免其复杂度和带宽限制。在 65nm 实现中，8-bit 量化模型上权重与激活数据占用平均降至 60% 和 48%，可与多种量化方法互补；与 Tensorcore 加速器集成后速度与能效分别提升 1.44× 和 1.37×，还能压缩训练中暂存的激活以支持定点推理。
Atalanta is a lossless, hardware/software co-designed compression technique for the tensors of fixed-point quantized deep neural networks. Atalanta increases effective memory capacity, reduces off-die traffic, and/or helps to achieve the desired performance/energy targets while using smaller off-die memories during inference. Atalanta is architected to deliver nearly identical coding efficiency compared to Arithmetic Coding while avoiding its complexity, overhead, and bandwidth limitations. Indicatively, the Atalanta decoder and encoder units each use less than 50B of internal storage. In hardware, Atalanta is implemented as an assist over any machine learning accelerator transparently compressing/decompressing tensors just before the off-die memory controller. This work shows the performance and energy efficiency of Atalanta when implemented in a 65nm technology node. Atalanta reduces data footprint of weights and activations to 60% and 48% respectively on average over a wide set of 8-bit quantized models and complements a wide range of quantization methods. Integrated with a Tensorcore-based accelerator, Atalanta boosts the speedup and energy efficiency to 1.44× and 1.37×, respectively. Atalanta is effective at compressing the stashed activations during training for fixed-point inference.
@inproceedings{LascorzMZ0IGAM24,
title = {Atalanta: A Bit is Worth a “Thousand” Tensor Values},
author = {Alberto Delmas Lascorz and Mostafa Mahmoud and Ali Hadi Zadeh and Milos Nikolic and Kareem Ibrahim and Christina Giannoula and Ameer Abdelhadi and Andreas Moshovos},
booktitle = {ASPLOS},
year = {2024},
doi = {10.1145/3620665.3640356}
}