Developers
August 4, 2020

Google Cloud Compute Engine: How To Choose The One Suits Your Specific Workloads

Running virtual machines in Google Cloud. Choose between E2, N2, N2D, and N1.

Today we will talk about the different Compute Engine machine families. If you are new to cloud computing, the recommendations given by Google Cloud can help you optimize your Compute Engine usage.  

If you are an organization, and you wish to run a virtual machine in Google Cloud, then Compute Engine might be the suitable option you are looking for. Compute Engine offers multiple machine families you can choose from. Each family is suited for specific workloads and applications.

General-purpose are machines that balance price and performance. They are suitable for almost every workload, including databases, development and testing of environments, development of web applications, and mobile gaming.  

Compute-optimized are machines that provide the highest performance per core on Compute Engine. They are optimized for compute-intensive workloads. Including high-performance computing, game servers, and latency-sensitive API serving.

Memory-optimized are machines that offer the highest of memory configurations. It works across all the VM families, and support up to 12 TB for a single instance. They are a good fit for any memory-intensive workload, for example, large in-memory database or data analytics workloads.

Accelerator-optimized These machines work on the NVIDIA Ampere a100 Tensor Core GPU. They support up to 16 GPUs in a single VM. These machines are suitable for demanding workloads like machine learning training. 

These machines provide a good balance of price and performance. They are suitable for many uses cases, most commonly common workloads. You can choose from four general-purpose types. E2, N2, N2D, and N1.

E2 offers the lowest total cost of ownership (TCO)  on the entire Google Cloud ecosystem. You can save up to 31% compared to the first generation, the N1. E2 runs across multiple platforms, Intel and AMD. If offers up to 32 vCPUs and 128GB of memory per node. E2 is also known for dynamic resource management, offering economic benefits for workloads, prioritizing savings.

N2 is the second generation of Intel Xeon Scalable Processors for Compute Engine´s general-purpose family. It offers a 20% price-performance improvement compared to the first generation, the N1 machines. It supports up to 25% more memory per CPU.

  • N2D is built on the second generation AMD EPYC CPU. It supports the highest core count and memory of any general-purpose Compute Engine VM. N2D provides you the same features as N2 VMs including local SSD and custom machine types.
  • N1s are the first generation of general-purpose VMs. It offers up to 96 vCPUSs and 624GB of memory. In most cases, it is recommended to choose the latest generations, (all of the above).

General-purpose machines come predefined, meaning that they have a predefined number of vCPUs and memory. They can also be configured as custom machine types. This customization allows you to independently configure CPU and memory so that you can find the right balance for your application, meaning that you only pay for what you need. Following next, we will describe the general-purpose machine family taking a closer look.

E2 

E2 VMs use dynamic resource management technologies. They are developed for Google's services so that it improves the hardware resources. It puts down costs, allowing you to save compared to other providers.

If you currently have workloads such as databases or application development that are running on N1, you should consider moving them to E2.  E2 VMs offer a 31% improvement in price-performance.

N2 

N2 machines run at a 2.8GHz base frequency and up to 3.4GHz. If offers up to 80 vCPUs and 640GB of memory. It makes them a great fit for many general-purpose workloads, benefiting from increased performance per core. It is widely used for web applications, servers, enterprise applications, gaming servers, content, collaboration systems, and databases.

If you are running a business based database or an interactive web application, N2 offers the ability to get a 30% higher performance from your VMs, shortening your computing processes.

N2 performs 2.82 times faster than N1. With that alone, you have the reason to move to N2.  It performs well on AI inference and Deep models. For this, it uses Intel-optimized TensorFlow.

In conclusion, If you want to run a virtual machine in Google Cloud, then a Compute Engine might be a suitable option you are looking for. Compute Engine offers multiple machine families you can choose from. Each family is suited for specific workloads and applications. The machines provide a good balance of price and performance. You can use them for a variety of cases. You can choose from four general-purpose types, with two being the most used. E2, N2, N2D, and N1. If you are running your services on N1, it is recommended you migrate them to N2 as there is more efficiency and you get a 30% performance improvement.

These machines provide a good balance of price and performance. They are suitable for many uses cases, most commonly common workloads. You can choose from four general-purpose types. E2, N2, N2D, and N1.

TagsTensorFlowCompute EngineGCP
Lucas Bonder
Technical Writer
Lucas is an Entrepreneur, Web Developer, and Article Writer about Technology.

Related Articles

Back
DevelopersAugust 4, 2020
Google Cloud Compute Engine: How To Choose The One Suits Your Specific Workloads
Running virtual machines in Google Cloud. Choose between E2, N2, N2D, and N1.

Today we will talk about the different Compute Engine machine families. If you are new to cloud computing, the recommendations given by Google Cloud can help you optimize your Compute Engine usage.  

If you are an organization, and you wish to run a virtual machine in Google Cloud, then Compute Engine might be the suitable option you are looking for. Compute Engine offers multiple machine families you can choose from. Each family is suited for specific workloads and applications.

General-purpose are machines that balance price and performance. They are suitable for almost every workload, including databases, development and testing of environments, development of web applications, and mobile gaming.  

Compute-optimized are machines that provide the highest performance per core on Compute Engine. They are optimized for compute-intensive workloads. Including high-performance computing, game servers, and latency-sensitive API serving.

Memory-optimized are machines that offer the highest of memory configurations. It works across all the VM families, and support up to 12 TB for a single instance. They are a good fit for any memory-intensive workload, for example, large in-memory database or data analytics workloads.

Accelerator-optimized These machines work on the NVIDIA Ampere a100 Tensor Core GPU. They support up to 16 GPUs in a single VM. These machines are suitable for demanding workloads like machine learning training. 

These machines provide a good balance of price and performance. They are suitable for many uses cases, most commonly common workloads. You can choose from four general-purpose types. E2, N2, N2D, and N1.

E2 offers the lowest total cost of ownership (TCO)  on the entire Google Cloud ecosystem. You can save up to 31% compared to the first generation, the N1. E2 runs across multiple platforms, Intel and AMD. If offers up to 32 vCPUs and 128GB of memory per node. E2 is also known for dynamic resource management, offering economic benefits for workloads, prioritizing savings.

N2 is the second generation of Intel Xeon Scalable Processors for Compute Engine´s general-purpose family. It offers a 20% price-performance improvement compared to the first generation, the N1 machines. It supports up to 25% more memory per CPU.

  • N2D is built on the second generation AMD EPYC CPU. It supports the highest core count and memory of any general-purpose Compute Engine VM. N2D provides you the same features as N2 VMs including local SSD and custom machine types.
  • N1s are the first generation of general-purpose VMs. It offers up to 96 vCPUSs and 624GB of memory. In most cases, it is recommended to choose the latest generations, (all of the above).

General-purpose machines come predefined, meaning that they have a predefined number of vCPUs and memory. They can also be configured as custom machine types. This customization allows you to independently configure CPU and memory so that you can find the right balance for your application, meaning that you only pay for what you need. Following next, we will describe the general-purpose machine family taking a closer look.

E2 

E2 VMs use dynamic resource management technologies. They are developed for Google's services so that it improves the hardware resources. It puts down costs, allowing you to save compared to other providers.

If you currently have workloads such as databases or application development that are running on N1, you should consider moving them to E2.  E2 VMs offer a 31% improvement in price-performance.

N2 

N2 machines run at a 2.8GHz base frequency and up to 3.4GHz. If offers up to 80 vCPUs and 640GB of memory. It makes them a great fit for many general-purpose workloads, benefiting from increased performance per core. It is widely used for web applications, servers, enterprise applications, gaming servers, content, collaboration systems, and databases.

If you are running a business based database or an interactive web application, N2 offers the ability to get a 30% higher performance from your VMs, shortening your computing processes.

N2 performs 2.82 times faster than N1. With that alone, you have the reason to move to N2.  It performs well on AI inference and Deep models. For this, it uses Intel-optimized TensorFlow.

In conclusion, If you want to run a virtual machine in Google Cloud, then a Compute Engine might be a suitable option you are looking for. Compute Engine offers multiple machine families you can choose from. Each family is suited for specific workloads and applications. The machines provide a good balance of price and performance. You can use them for a variety of cases. You can choose from four general-purpose types, with two being the most used. E2, N2, N2D, and N1. If you are running your services on N1, it is recommended you migrate them to N2 as there is more efficiency and you get a 30% performance improvement.

These machines provide a good balance of price and performance. They are suitable for many uses cases, most commonly common workloads. You can choose from four general-purpose types. E2, N2, N2D, and N1.

TensorFlow
Compute Engine
GCP
About the author
Lucas Bonder -Technical Writer
Lucas is an Entrepreneur, Web Developer, and Article Writer about Technology.

Related Articles