此语言仍有 24 条内容或界面任务待翻译;未完成部分显示中文原文。
可用量待供方确认GPU Cloud Host (Cloud Instance, Computing Type)

ZunCloud GPU Cloud Host Computing Type g4v.2xlarge.8: 1×NVIDIA V100 32GB Cloud Instance

尊云(广东尊云网络科技有限公司) · Nationwide 10 multiple availability zones, distributed across East China, North China, South China, Inner Mongolia and other regions

Content Correction
Reference unit priceProvided by SupplierPending Quote
单卡显存Provided by Supplier32GB
单卡算力Provided by SupplierCUDA FP32 performance 14 TFLOPS; Tensor FP16/FP32 performance 112 TFLOPS
Availability AgreementProvided by Supplier99.95% availability commitment

Configuration and Network

GPU
NVIDIA V100
CPU
Intel Xeon Gold 6248R@3.0GHz (1 units)
Host Memory
8C64G
Storage
SSD cloud disk 40G
Delivery method
GPU Cloud Host (Cloud Instance, Computing Type)
节点地区
Nationwide 10 multiple availability zones, distributed across East China, North China, South China, Inner Mongolia and other regions
Minimum rental period
From ¥/month

Deployment and Delivery Instructions

ZunCloud offers GPU cloud host computing type services based on the NVIDIA T4 and V100 models, primarily targeting scenarios such as machine learning, graphics rendering, 3D modeling, and floating-point computing. This listing corresponds to the g4v.2xlarge.8 instance: a single NVIDIA V100 (32GB VRAM, 900 GB/s bandwidth), Intel Xeon Gold 6248R CPU, 8C64G RAM, 40G SSD cloud disk, 10M bandwidth, listed from 3529 yuan/month. The page states that GPU products are launched in 10 multiple availability zones across East China, North China, South China, Inner Mongolia and other regions, with an 99.95% availability commitment. The full-machine monthly price cannot be converted to yuan/card-hour; specific availability zones and card-count inventory must be confirmed with the supplier.

Activation Process

  1. 提交用途、配置、卡数与租期
  2. 与供方确认资源、网络和正式报价
  3. 签订服务约定并确认交付
  4. 验收环境、访问权限与服务指标

Editorial Analysis Inferred from source information

g4v.2xlarge.8 was selected because it is the instance on the page that simultaneously provides instance identification, GPU card type, single-card VRAM (32GB), and selling region information. This configuration is suitable for AI training, inference, and HPC scenarios (as noted on the page); compared with the T4 model (16GB, 320 GB/s), the V100 has larger VRAM and approximately 3× VRAM bandwidth, making it suitable for workloads sensitive to VRAM bandwidth in large-model inference and training. The full-machine monthly billing means budget evaluations should be based on the whole machine rather than converted to card-hours; before deployment, verify with the supplier the actual rentable card count in the target availability zone, the minimum lease term, and whether hourly billing options are available.

Sources 尊云

Compilation Notes 内容由 AI 根据公开资料整理;事实与推断分开呈现。

Content Correction
Product Selection Assistant

让需求变成可修改的筛选条件

依据已录入参数生成条件,结果可查看和修改。帧率与性能需核对实测,不能仅凭算力数值保证。