随着 Apple 发布了更加强劲的 M2 pro 和 Max 芯片，以及 PyTorch / Tensorflow 对 Apple silicon 芯片的支持，在 Mac 上训练深度模型也变得越来越现实。翻了翻几篇比较详尽的报告，分别在 M1 / M1 Pro 和 M1 Max 上跑训练的比较。很期待这些实验在 M2 系列上表现。
Running PyTorch on the M1 GPU
https://sebastianraschka.com/blog/2022/pytorch-m1-gpu.html
Benchmarking the Apple M1 Max
https://tlkh.dev/benchmarking-the-apple-m1-max
Deep Learning on the M1 Pro with Apple Silicon
https://wandb.ai/tcapelle/apple_m1_pro/reports/Deep-Learning-on-the-M1-Pro-with-Apple-Silicon---VmlldzoxMjQ0NjY3
PyTorch Runs On the GPU of Apple M1 Macs Now!
https://wandb.ai/capecape/pytorch-M1Pro/reports/PyTorch-Runs-On-the-GPU-of-Apple-M1-Macs-Now-Announcement-With-Code-Samples---VmlldzoyMDMyNzMz
Can Apple’s M1 Help You Train Models Faster & Cheaper Than NVIDIA’s V100?
https://wandb.ai/vanpelt/m1-benchmark/reports/Can-Apple-s-M1-Help-You-Train-Models-Faster-Cheaper-Than-NVIDIA-s-V100---VmlldzozNTkyMzg
Post #2404
804
DPS Main 感觉 Apple silicon SoC 把整个笔记本的价格都往下拉了。 把 CPU 和 内存都选到顶配，12核 C PU + 38 核 GPU + 96G 内存，总价三万出点头。 https://www.apple.com.cn/macbook-pro-14-and-16/