Tightly coupled high performance computing (HPC) workloads often use the Message
Passing Interface (MPI) to communicate between processes and Compute Engine
instances. However, the process of building your own OS image that is tuned for
optimal MPI performance requires systems expertise, Google Cloud knowledge,
and extra time for maintenance. To quickly set up an optimized environment for
your HPC workloads, use an *HPC VM image*.

An HPC VM image is based on Rocky Linux 8 or Rocky Linux 9, and is
optimized for tightly-coupled HPC workloads. The image includes pre-configured
kernel and network tuning parameters required to create compute instances that
achieve optimal MPI performance on Google Cloud.

For best results, deploy the HPC VM image on a machine type from the
[compute-optimized machine series](https://docs.cloud.google.com/compute/docs/compute-optimized-machines).
These machine types are designed for
tightly-coupled HPC applications. To use Cloud RDMA with H4D,
you must use version 20250917 or later of the HPC VM Rocky 8 image.

You can create an HPC VM instance by using any of the methods available to
create a compute instance. You can also use
[SchedMD's Slurm workload manager](https://docs.cloud.google.com/cluster-toolkit/docs/quickstarts/slurm-cluster),
to deploy an HPC cluster. Slurm support isn't yet available for HPC VM images
based on Rocky Linux 9.

## Benefits

The HPC VM image provides the following benefits:

1. **Compute instances pre-configured for HPC workloads**. There is no need to manually tune performance, manage compute instance reboots, or stay up to date with the latest Google Cloud updates for tightly-coupled HPC workloads.
2. **Networking optimizations for tightly-coupled workloads**. Optimizations that reduce latency for small messages are included. These optimizations benefit applications that are heavily dependent on point-to-point and collective communications. If using the H4D machine series, the HPC VM Rocky 8 image contains the necessary Cloud RDMA drivers.
3. **Compute optimizations for HPC workloads**. Optimizations that reduce system jitter are included, which makes single-node high performance more predictable.
4. **Consistent, reproducible performance**. OS image standardization gives you consistent, reproducible application-level performance.
5. **Improved application compatibility** . Alignment with the node-level requirements of the [Intel HPC platform specification](https://www.intel.com/content/www/us/en/high-performance-computing/hpc-platform-specification.html) enables a high degree of interoperability between systems.

## HPC VM image features

The HPC VM image offers several features that are designed to optimize
performance for high performance computing (HPC) workloads:

- Automatic updates disabled
- MPI collective tunings
- Pre-installed RPMs

### Automatic updates disabled

Automatic updates can have a negative impact on the performance of HPC
workloads. Disable automatic updates for HPC VM images by setting the
`google_disable_automatic_updates` metadata entry to `TRUE` during instance
creation. How you set this metadata entry during instance
creation depends on the tool that you use to create the instance.

For example, when you use the
[`gcloud compute instances create`](https://docs.cloud.google.com/sdk/gcloud/reference/compute/instances/create)
command to create a compute instance, provide the
[`--metadata`](https://docs.cloud.google.com/sdk/gcloud/reference/compute/instances/create#--metadata)
argument. For more information, see [About VM metadata](https://docs.cloud.google.com/compute/docs/metadata/overview).

The `google_disable_automatic_updates` metadata entry is available as follows:

- **HPC VM Rocky 9 images**: all versions.
- **HPC VM Rocky 8 images** : only in images built on or after July 12, 2024. The build date (formatted as `YYYYMMDD`) is used as the version number at the end of image names. For example, an HPC VM Rocky 8 image built on the 21st of July, 2025 has the name `hpc-rocky-linux-8-v20250721`.

> [!IMPORTANT]
> **Important:** If you disable automatic updates, then VM Manager can cause performance of HPC workloads to degrade. VM Manager consumes CPU and network resources at regular intervals when it checks for updates. A workaround is to [disable the OS Config agent](https://docs.cloud.google.com/compute/vm-manager/docs/setup#disable-agent).

If you disable automatic updates on your H4D instances, then you should
regularly run the `dnf update` command on the instance to keep the
Cloud RDMA driver up to date.

### Intel MPI library

Google recommends that you use the `google_install_intelmpi` script to install
the message passing interface (MPI) library before running MPI jobs on
Google Cloud. For more information, see
[Use Intel MPI 2021](https://docs.cloud.google.com/compute/docs/instances/create-hpc-vm#use-mpi-2021).

### Pre-installed RPMs

The HPC VM image comes with the following RPM packages pre-installed:

- `gcc-gfortran`
- `gcc-toolset-12`
- `Lmod`
- `dkms`
- `htop`
- `hwloc`
- `hwloc-devel`
- `infiniband-diags`
- `kernel-devel`
- `kmod-idpf-irdma`
- `libfabric`
- `librdmacm-utils`
- `libibverbs-utils`
- `libXt`
- `ltrace`
- `nfs-utils`
- `numactl`
- `numactl-devel`
- `papi`
- `pciutils`
- `pdsh`
- `perf`
- `perftest`
- `rdma-core`
- `redhat-lsb-core` (HPC VM Rocky 8 only)
- `redhat-lsb-cxx` (HPC VM Rocky 8 only)
- `rsh`
- `screen`
- `strace`
- `wget`
- `zsh`
- "Development Tools" package group

## Pricing

The HPC VM image is available at no additional cost. Because the HPC VM image
runs on Compute Engine, you might incur charges for
Compute Engine resources such as vCPUs, disks, and memory.
To learn more, see
[Compute Engine pricing](https://cloud.google.com/compute/vm-instance-pricing).

## What's next

- Learn how to [Create an HPC-ready instance](https://docs.cloud.google.com/compute/docs/instances/create-hpc-vm).
- Learn about [Advanced Compute Images](https://docs.cloud.google.com/compute/docs/images/workload/aci-images) for HPC workloads.
- Review the [best practices for running HPC workloads](https://docs.cloud.google.com/cluster-toolkit/docs/tutorials/best-practices).
- Learn how to [set up and scale MPI applications on H4D VMs with Cloud RDMA](https://docs.cloud.google.com/cluster-toolkit/docs/setup/configure-mpi-application) on Google Cloud.
- Learn how to [create H4D HPC clusters with enhanced cluster management capabilities](https://docs.cloud.google.com/compute/docs/hpc/overview-hpc-clusters).
- If you have feedback or require support, then send an email at [hpc-image-feedback@google.com](mailto:hpc-image-feedback@google.com).