How has the recent turmoil within the OpenAI offices changed your plans to use GPT in a business process or product in 2024?
Increased uncertainty means we are more likely to evaluate alternative AI chatbots and LLMs.
No change in plans, though we will keep an eye on the situation.
With Sam Altman back in charge, we are more likely to go all-in with GPT and LLMs.
What recent turmoil?
Cloud Services

Google Compute Engine Now Offers Custom Virtual Machine Sizing

Feb 17th, 2016 9:00am by
Featued image for: Google Compute Engine Now Offers Custom Virtual Machine Sizing

No longer will Google Compute Engine users will be forced to choose among preset vCPU speeds. The company has formally launched Custom Machine Types, allowing users to create virtual machines with custom vCPU and memory configurations.

The approach could save users up to 50 percent on compute costs, Google reckoned. The company first announced the service, in beta form, in November.

“Since our Beta launch, we have seen customers create virtual machines with novel vCPU and memory ratios that aren’t available from any of the major cloud providers,” wrote Sami Iqram, who is a product manager, for the Google Cloud Platform, in a blog post.

Custom Machine Types provides a way to configure the VMs for the specific workload at hand, balancing the amount of memory and CPU power you need against the price you are willing to pay.

When users sign onto the compute section of the Google Cloud Platform, they are now presented with the option of creating a basic or a “custom” machine type.

A virtual machine can have as little as a single vCPU and as many as 32 vCPUs. The memory can also be set as well, up to 6.5 Gibibytes.

Customers who have tried this approach have saved an average of 19 percent, compared to using standard virtual off-the-shelf vCPU configurations, the company boasted.

For instance, hosting provider Wix has saved 18 percent whilst using the variable vCPU option for its media platform. Marketing platform provider Lytics is saving between 20 and 50 percent and movie service iRewind enjoyed a 20 percent savings on its workloads, Google reported.

The pricing is based on a per-minute usage model. The base price is  $0.03492 per hour for vCPU and $0.00468 per GB of memory, and varies linearly by the size of configuration: an 8 vCPU 20 GiB memory VM will cost twice the amount of a 4 vCPU, 10 GiB of memory VM. Discounts based on sustained usage and preemptible workloads are are also available.

These virtual machine types can be used with CentOS, CoreOS, Debian, OpenSUSE, Ubuntu and — new with this release — Red Hat and Windows operating systems. You can also roll your own Linux distro and run it on these variable vCPUs as well. In addition to GCE, Custom Machine Types are recognized by the Google Container Engine and Deployment Manager.

Creating a custom sized VM through the gcloud command-line tool and through our API can be done thusly:


Group Created with Sketch.
THE NEW STACK UPDATE A newsletter digest of the week’s most important stories & analyses.