LecturaInvestigaciónRadarMarco de inversión
Iniciar sesión / Registrarse
← Volver a la lectura
NoticiasTema · TorchTPU开源项目Material2 min de lecturaEditorial de DisconfirmAI

Anyscale and Google Cloud Announce Native Ray Support for PyTorch on TPUs

Developers can now run PyTorch tasks directly on Google TPUs via the Ray framework without rewriting code, enabling efficient distributed training and inference.

Verificado 2026-10-21 07:26 GMT+8

On October 20, 2026, compute platform company Anyscale and Google Cloud jointly announced that the open-source distributed computing framework Ray now natively supports running PyTorch workloads on Google Tensor Processing Units (TPUs). This integration builds on Google's TorchTPU project introduced last year and KubeRay on Google Kubernetes Engine (GKE), aiming to provide a unified control plane for platform engineering and machine learning teams to schedule and manage distributed training, vLLM inference, and reinforcement learning tasks across GPUs and TPUs.

Historically, using TPUs with Ray required adapting to a Single-Program, Multiple Data (SPMD) execution model designed primarily for JAX, where a single Ray worker process per host VM managed all attached TPU chips. This update introduces a native PyTorch experience with one worker per TPU chip: each Ray worker requests a single logical TPU device, similar to setting num_gpus=1 on GPUs. Ray automatically isolates and assigns individual TPU devices to workers via TPU_VISIBLE_CHIPS. Developers can migrate existing PyTorch code by simply changing the device string from "cuda" to "tpu" and switching the communication backend from "nccl" to "tpu_dist".

For large-scale training scenarios, Ray Train's TorchTrainer now supports automatic configuration of TPU slices. By specifying use_tpu=True along with the target topology (e.g., "4x4") and accelerator type (e.g., "TPU-V6E"), the system automatically reserves atomic TPU slices, ensuring all training workers are gang-scheduled onto contiguous physical TPU slices over high-speed interconnects, preventing resource fragmentation. Additionally, combined with the vLLM-TorchTPU plugin, developers can deploy OpenAI-compatible online inference services on TPU slices using the same declarative API.

Iniciar sesión / Registrarse
LecturaInvestigaciónRadarMarco de inversión

Fuentes

  • anyscale.com↗
Tema

TorchTPU开源项目

谷歌于2026年10月20日在PyTorch大会上正式开源TorchTPU,旨在打破英伟达(Nvidia)在AI芯片领域的软件垄断。该工具允许开发者通过简单的配置更改,将现有的PyTorch应用原生迁移至谷歌云TPU上运行,无需重写代码。此…

4 pasos3 piezasActualizado 2026-10-22 09:56

Cómo llegó hasta aquí

  1. 2026-10-18亚马逊AWS发布TorchNeuron,支持PyTorch工作负载在其Trainium芯片上运行,并承诺向主PyTorch项目贡献解耦代码。techcrunch.com ↗
  2. 2026-10-20谷歌与Meta联合开源TorchTPU,提供原生PyTorch后端以支持Google Cloud TPU,并与Hugging Face、vLLM等主流工具集成。ithome.com.tw ↗
  3. 2026-10-20Anyscale宣布Ray框架原生支持在Google TPU上运行PyTorch任务,简化分布式训练与推理流程。anyscale.com ↗

Lo que aún está pendiente

  • 开发者实际迁移到TPU的比例是否会显著侵蚀英伟达的市场份额?
  • TorchTPU在复杂模型上的性能表现能否真正匹敌经过深度优化的CUDA环境?
  • 其他硬件厂商(如AMD、Intel)是否会推出类似的专用PyTorch后端以应对竞争?

También en este hilo

  • Google and Meta Open-Source TorchTPU to Enable Native PyTorch on TPUs, Challenging Nvidia's Dominance2026-10-21
  • Intel and AMD Announce Collaboration on oneDNN Optimizations for PyTorch2026-10-21
Leer el tema completo →
Continuar investigando →