More articles

Want to bridge the cloud cost transparency gap between Management and Engineering?

Get in touch with us, we're here to help.
Zoltán Guth

CTO

Balázs Molnár

CEO

Boost performance and cut costs with AMD-based machines on Google Cloud

Rabbit Team

3 min read

Did you know that AMD-based machine series on Google Cloud are not only more cost-effective but also offer better performance? Read on for the details of why you should consider making the switch.

Cost Savings and Performance Boost

Savings: AMD-based machines are 13% cheaper compared to their Intel counterparts.

Performance: AMD machines provide up to 20% better performance.

Comparing Intel and AMD machines

All the following examples come from the US-Central-1 Region.

First, let’s see the vCPUs:

Example 1: N2 Series

  • Intel (n2-standard-4)
    • vCPUs: 4
    • RAM: 16 GB
    • Cost: $141.79/month
  • AMD (n2d-standard-4)
    • vCPUs: 4
    • RAM: 16 GB
    • Cost: $123.36/month (-13%)

Example 2: C3 Series

  • Intel (c3-standard-4)
    • vCPUs: 4
    • RAM: 16 GB
    • Cost: $152.43/month
  • AMD (c3d-standard-4)
    • vCPUs: 4
    • RAM: 16 GB
    • Cost: $132.57/month (-13%)

In both cases, the cost is 13% less for AMD machines than for the Intel ones.
The 13% difference also applies if you have CUDs. (Source)

As for performance, let’s compare their CoreMarks.
CoreMark is a benchmark that measures the performance of CPUs, higher is better. (Source)

Example 1: N2 Series

  • Intel (n2-standard-4)
    • CoreMark Score: 66,884
  • AMD (n2d-standard-4)
    • CoreMark Score: 80,098 (+20%)

Example 2: C3 Series

  • Intel (c3-standard-4)
    • CoreMark Score: 80,609
  • AMD (c3d-standard-4)
    • CoreMark Score: 94,572 (+17%)

So, in both cases, AMD machines can give a significantly better performance: +17% and +20%.

Real-World Impact

One customer using Rabbit switched their GKE node pool from N2 to N2D, resulting in:

  • 26.5% total savings
    • 13.5% reduction in the number of nodes
    • 13% decrease in node costs

Considerations Before Switching

  • Active Resource-Based Commitments: Wait until they expire to switch.
  • Nested Virtualization: Not supported on AMD-based machines.
  • Hyperdisk Extreme: Not compatible with AMD.

Next steps

Switching to AMD-based machines can lead to significant cost savings and performance improvements. Check your specific environment and see how much you can save with AMD on Google Cloud.

For more details, refer to the Google Cloud pricing page and CoreMark scores. Or if you want to see for yourself how much you can save in your GCP environment, see below how Rabbit shows multiple options for machine types:

Machine type recommendations in Rabbit, comparing AMD and Intel machines

Sign up for a free trial, or contact us to try Rabbit on demo data.

More articles

3 min read
The ultimate guide for GCP cost optimization - part 1 (GKE, GCE and CUD)

Zoltán Guth, CTO of Rabbit shares the team’s findings and experiences with our customers on how to identify cost-saving opportunities at the service level.

Read more
7 min read
Cloud engineers, care about cloud cost - sooner rather than later!

Our team attended the first three stops of the Google Cloud Summits DACH. Thank you for stopping by our booth and for the amazing conversations!

Read more

Want to bridge the cloud cost transparency gap between Management and Engineering?

Get in touch with us, we're here to help.
Zoltán Guth

CTO

Balázs Molnár

CEO