top of page
Edition

Scaletorch Powered GPU Servers

Scaletorch Powered GPU Servers (STPGS) are a combination of GPU servers with offloader machines connected via a high speed network (dual 40gbit or dual 100gbit) using previous generation GPUs.

STPGS.png

Use Case.

All size of models supported for CV, video and audio workloads.

Edge AI Use Cases.

Group 390.png

Models where a large number of iterations need to be carried out to quantize/prune models

Models where a experimentation with a number of different backbones needs to be carried out.

Types of GPUs Available

A100

V100

P100

RTX A6000 Ada

RTXA4000 Ada

No. of GPUs per Machine

4

6

8

Scaletorch Hosted Private Cloud
stpgs_diagram.png

Full bare metal access to the Scaletorch Powered GPU Servers​

Peer the machine with customer VPC(s)

Scaletorch Powered GPU Servers is situated at the nearby datacenter

Encrypted cache on offloader

(Optional)

Supported Cloud Connectivity

SKUs

GPU Servers

P100

8 x P100 (16GB)
4 x 40Gbit interconnect
2 x 10GbE ethernet

V100

8 x V100 (16GB)
4 x 40Gbit or 4 x 100Gbit interconnect
2 x 10GbE ethernet

A100

8 x V100 (16GB)
4 x 40Gbit or 4 x 100Gbit interconnect
2 x 10GbE ethernet

Offload Servers

128 Core

4 x 40Gbit interconnect
2 x 10GbE

192 Core

4 x 40Gbit or 2 x 100Gbit interconnect
2 x 10GbE

256 Core

4 x 40Gbit or 2 x 100Gbit interconnect
2 x 10GbE

Pricing Plans

What's the price?

Thanks for submitting! 🙌

Have a look at our another products:

bottom of page