Google Cloud A4X Max Bare Metal Instances Support 50k GPU Clusters, Network Bandwidth Doubles

robot
Abstract generation in progress
ME News Updates, April 19 (UTC+8), Google Cloud announced that its A4X Max bare-metal instance supports clusters of up to 50,000 GPUs, with network bandwidth twice that of the previous generation. This instance belongs to the Google Compute Engine Accelerator-Optimized Machine Series, which comes pre-installed with NVIDIA GPUs and is designed for AI, machine learning, high-performance computing, and graphics-intensive applications. The documentation details multiple machine series including A4X Max, A4X, A4, A3, A2, G4, and G2, and recommends specific series based on workload types such as pre-training, fine-tuning, inference, graphics, and high-performance computing. Additionally, the documentation explains pricing and consumption options (on-demand, Spot, Flex-start, reserved) based on pre-installed GPUs, vCPUs, memory, and local SSDs, as well as the maintenance experience for different machine types. (Source: InFoQ)
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 8
  • Repost
  • Share
Comment
Add a comment
Add a comment
MirrorBallRolling
· 1h ago
Spot instances for AI training—aren’t you worried about them being snatched up before the checkpoint is fully saved?
View OriginalReply0
BluePeonyPlan
· 2h ago
How to calculate the upgrade and migration costs for the third generation of A2 to A4X, with three generations living together?
View OriginalReply0
Lightning-FastComposure
· 2h ago
G2 is still selling, and this product line covers everything from inference to graphics rendering.
View OriginalReply0
SatsumaSignal
· 2h ago
What’s the new trick with Flex-start—an intermediate state between on-demand and reserved?
View OriginalReply0
LeverageWhisperer
· 2h ago
The better service is the more expensive one.
View OriginalReply0
Half-SectionedSucculent
· 2h ago
Cloud providers are going crazy, a 50k-card cluster—are they trying to take over the supercomputing center's business?
View OriginalReply0
LatencyLullaby
· 2h ago
Is the pricing based on pre-installed GPUs? Isn't the VRAM size hidden and not mentioned?
View OriginalReply0
SmallPosition,BigMouth
· 2h ago
Bare metal + 50k cards, what kind of divine network topology is this?
View OriginalReply0
  • Pinned