Configure XMRig Log Files

What this guide covers

This page addresses configuration syntax and option scope for XMRig on Windows. Its scope is deliberately narrow so it does not compete with the general Windows overview, the download page or the full configuration reference.

Before making changes

Use hardware you own or are authorized to administer. Record the XMRig version, Windows build, processor or graphics device, pool endpoint and current configuration. Preserve a copy of the last working config.json. When a downloaded archive is involved, compare its SHA-256 value with the checksum published by the original project before running it.

Step-by-step method

1. Reproduce the current behavior and save the complete startup log.
2. Establish a baseline with default or previously verified settings.
3. Change only the option directly connected to this guide's problem.
4. Allow the miner and hardware to reach a steady state before measuring.
5. Confirm the result after a restart and retain the configuration used.

Use the 15-minute local hashrate and a longer pool-side window for performance comparisons. A short 10-second figure is useful for spotting an immediate failure but is too noisy for judging a small improvement. Record accepted and rejected shares separately.

How to interpret the result

A successful test should be reproducible and should not trade a small hashrate increase for errors, rejected shares, excessive temperature or loss of Windows responsiveness. Pool-reported hashrate naturally lags and varies because it is estimated from submitted shares. Compare equivalent time windows rather than expecting the pool and local console to match moment by moment.

For RandomX, fast mode needs a shared dataset slightly above 2 GB, with additional allocation possible on multi-NUMA systems. It does not require about 2.5 GB separately for every mining thread. Thread selection is influenced heavily by processor cache and topology.

Common mistakes

• Editing several unrelated options before retesting.
• Judging performance from the first minute after launch.
• Using a configuration copied from a different processor or pool.
• Ignoring temperatures, clocks or rejected shares.
• Disabling security controls broadly instead of investigating a specific warning.

Testing method and sources

The XMRigSoftware Editorial Team reviews instructions on Windows 10 and Windows 11 using the current public XMRig release, a clean configuration, a warm-up period and repeat measurements. Primary references are the XMRig source repository, the original documentation, Microsoft Windows documentation, and processor information from AMD and Intel.

Written and reviewed by: XMRigSoftware Editorial Team
Last reviewed: September 22, 2026
Test scope: Windows 10, Windows 11, x64 and Windows ARM64 where supported.