(All-cores)
(Single-core)
Specifications
Server Metadata
Vendor ID | gcp |
Server ID | 700095 |
Name | ct3p-hightpu-4t |
Description | 48 vCPUs, 340 GB RAM, 1 vNUMA nodes |
Family | ct3p |
Hw Virt | - |
Status | active |
Observed At | 2026-10-26T11:40:28.917944 |
Availability
| REGION / ID | SPOT | ONDEMAND |
|---|
Processor
vCPUs | 48 |
CPU Allocation | Dedicated |
CPU Architecture | x86_64 |
System Resources and Accelerators
| MEMORY | |
|---|---|
Memory Amount | 340 GiB |
| ACCELERATORS | |
|---|---|
Accelerator Count | 4 |
Accelerator Memory Min | 32 GiB |
Accelerator Memory Total | 128 GiB |
Accelerator Type | TPU |
Accelerator Manufacturer | Google |
Accelerator Family | TPU |
Accelerator Model | v3 |
Accelerators |
|
| STORAGE | |
|---|---|
Storage Size | 0 GB |
Storages |
| NETWORK | |
|---|---|
Inbound Traffic | 0 GB/month |
Outbound Traffic | 0 GB/month |
IPv4 | 0 |
Server Description
A dedicated accelerator platform featuring four integrated tensor processing units and high-capacity memory for complex machine learning workloads.
Google Cloud Platform ct3p-hightpu-4t is a specialized, accelerator-optimized server configured with 48 dedicated x86_64 vCPUs and 340 GB of system memory. Designed primarily for machine learning workloads, the instance integrates four Google TPU v3 accelerators providing a total of 128 GB of high-bandwidth VRAM. It operates within a single vNUMA node and does not include local storage, requiring external network storage solutions for data persistence. With 7.08 GB of RAM per core, the hardware profile is tailored to handle large datasets during model training and execution. This server is engineered for deep learning workloads, including neural network training, natural language processing, and high-throughput tensor inference, offering a highly dense and specialized compute environment for AI development.
Economics
Alternatives
Servers of the Same Family
| INSTANCE | vCPUs | MEMORY | Accelerators |
|---|
Similar Servers
| INSTANCE | VENDOR | vCPUs | MEMORY | Accelerators |
|---|