Why does XMRig say "failed to allocate RandomX dataset using 1GB pages"?

XMRig shows “failed to allocate RandomX dataset using 1GB pages” when it cannot reserve the special large memory pages you enabled for Monero mining. This usually means a performance option failed, not that mining has stopped: XMRig can normally fall back to smaller pages.

What XMRig is trying to do

RandomX is Monero’s mining algorithm—the set of calculations your computer performs to help secure the network. In its fast mode, it uses a dataset of about 2,080 MB stored in RAM.

Your operating system divides memory into chunks called pages. XMRig can use larger chunks to reduce the work involved in accessing that dataset.

There are two important sizes to distinguish:

• 2 MB huge pages: The usual large-page optimization for XMRig.
• 1 GB huge pages: A separate, optional optimization supported by XMRig on Linux.

The warning specifically concerns 1 GB pages. It does not automatically mean that ordinary 2 MB huge pages also failed.

Because the dataset is slightly larger than 2 GB, XMRig generally needs three 1 GB pages per dataset. On systems with multiple memory regions, called NUMA nodes, it may create additional datasets.

Why that allocation fails

The most common cause is that your system has no free 1 GB huge pages reserved. Having plenty of ordinary free RAM is not enough: Linux maintains a separate pool for these pages.

Other possible causes include:

• Missing support: Your processor or Linux kernel does not support 1 GB huge pages.
• Fragmented memory: Enough memory is free overall, but Linux cannot find sufficiently large, continuous sections to reserve.
• Access restrictions: Your account, service, or container cannot allocate the reserved pages.
• NUMA placement: Pages are available on one memory node, but XMRig needs them on another.
• Another application: Something else has already taken the reserved pages.

This is not normally a problem with your wallet address, mining pool, or internet connection.

Check whether mining continued

Read the lines immediately after the warning. Look for successful dataset initialization and mining activity.

XMRig’s reported hashrate—the number of mining calculations performed each second—shows whether your processor is working. Accepted shares, which are pieces of work acknowledged by your pool, help confirm that mining is functioning.

If XMRig continues mining, you can usually leave it running. The benefit of 1 GB pages over working 2 MB huge pages is generally modest and depends on your system.

If dataset initialization fails completely or XMRig exits, investigate the later error messages too. You may have a broader memory shortage rather than just an unavailable optimization.

Fix it, or disable the optional feature

If you do not specifically need 1 GB pages, remove the --randomx-1gb-pages command-line option or set "1gb-pages": false inside the "randomx" section of your configuration. You can still use ordinary huge pages.

If you want to keep the feature:

• Check that your Linux system supports 1 GB pages.
• Reserve enough pages for each dataset XMRig creates.
• Prefer boot-time reservation if allocating them after startup fails.
• Check allocation permissions and any container limits.

Reserved huge pages reduce RAM available to other applications. Do not reserve most of your memory or run an untrusted “fix” script.

The bottom line

This warning usually means XMRig could not use an optional Linux memory optimization. Check whether mining continued before changing anything; working 2 MB huge pages are often the simpler choice.