G-Power：基于已知GPU聚合知识基础的架构级GPU功耗建模
G-Power: Architecture-level GPU Power Modeling with Aggregated Knowledge Foundations from Known GPUs
浏览论文内容
中文总结 AI 辅助
本文针对现有架构级GPU功耗模型因采用微基准测试训练导致准确率低的问题，提出G-Power框架，通过利用已知GPU聚合知识的三阶段算法建模，在四款NVIDIA GPU上实现比AccelWattch更优的功耗预测性能。
中文摘要 AI 辅助
图形处理器（GPU）已成为大规模并行计算的关键计算资源。随着芯片复杂度不断提升，能效已成为现代GPU的重要设计目标。GPU功耗优化依赖快速功耗评估，需要架构级GPU功耗模型。然而，由于功耗标签采集耗时，现有模型仅采用简单微基准测试作为训练数据。微基准测试作为训练数据的局限性导致AccelWattch等现有架构级GPU功耗模型准确率较低。为解决微基准测试作为训练数据的局限性，本文提出G-Power，一种架构级GPU功耗建模框架，利用额外已知GPU芯片提供附加知识。G-Power利用额外已知GPU芯片的聚合知识基础，再对目标GPU进行微调。为提供额外已知GPU芯片的基础并捕捉相似性以利用这些基础进行微调，G-Power采用三阶段算法：1）用额外已知芯片进行预训练；2）类注意力聚合；3）对目标GPU进行微调。我们在四款现代NVIDIA GPU上评估G-Power，结果显示其准确率较高。G-Power平均可实现14%的低平均绝对百分比误差（MAPE）和0.88的高相关系数（R），相比AccelWattch，MAPE降低22%，R提高0.36。
英文摘要
Graphics Processing Units (GPUs) have been serving as critical computation resources for large-scale parallel computations. With increasing chip complexity, power efficiency has become an important design objective for modern GPUs. GPU power optimization relies on fast power evaluation, requiring architecture-level GPU power model. However, because of the time-consuming power label collection, only simple microbenchmarks are adopted for training. The limitation of microbenchmarks as training data incurs low accuracy for existing architecture-level GPU power models like AccelWattch. To address the limitation of microbenchmarks as training data, we propose G-Power, an architecture-level GPU power modeling framework that utilizes additional known GPU chips to provide additional knowledge. G-Power utilizes the aggregated knowledge foundation from additional known GPU chips and then performs fine-tuning on our target GPU. To provide foundations with additional known GPU chips and capture the similarity to utilize these foundations for fine-tuning, G-Power adopts a three-phase algorithm consisting of 1) pre-training with additional known chips, 2) attention-inspired aggregation, and 3) fine-tuning on our target GPU. We evaluate G-Power on four modern NVIDIA GPUs, demonstrating high accuracy. G-Power can achieve a low MAPE of 14% and a high correlation coefficient R of 0.88 on average, which are 22% lower MAPE and 0.36 higher R than AccelWattch.
