TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Ask HN: How come my Ryzen 3750H is faster than my RTX 2060 in TensorFlow?

7 点作者 s-p-n大约 5 年前
I thought the RTX would do better here, but the Ryzen 3750h performs better on training a model in tensorflow. I&#x27;m using this guide to get things going, and I&#x27;m usng tf-nightly with Ubuntu 20.04.<p>https:&#x2F;&#x2F;www.tensorflow.org&#x2F;tutorials&#x2F;quickstart&#x2F;beginner<p>I&#x27;m wondering if using nvidia 440 drivers is the reason? I also have 3 versions of CUDA installed (10.0, 10.1, and 10.2). The tensorflow-js test I did uses CUDA 10.0, and the Ryzen performs roughly 50% faster. The Python test I did uses CUDA 10.1, and the CPU is less than 10% faster.<p>Even so, shouldn&#x27;t the GPU perform an order of magnitude better here? I feel like I must be doing something wrong :&#x2F;

3 条评论

Ragib_Zaman大约 5 年前
The network that you train in that tutorial is quite small. Your networks have to get quite large before the overhead of transferring the network onto the GPU is negligible to the cost of training it. You&#x27;ll see the benefit of the GPU with larger models.
navjack27大约 5 年前
I think my calculations are correct here<p><pre><code> Model Zen+ 2060 Cores 4 1920 Threads 8 1920 Cycles 2.3 1.365 FP64 GFLOPS 2944 163 FP32 GFLOPS 5888 5242 FP16 GFLOPS 11776 10483 </code></pre> So it would kind of make sense yeah. 2060 isn&#x27;t that strong.<p>But note: I&#x27;m not totally sure on the Zen+ theoretical speeds
评论 #22613234 未加载
deathtrader666大约 5 年前
Didn&#x27;t know that TensorFlow works on AMD GPUs..
评论 #22614800 未加载