Loading...
Loading...
For any HPE ProLiant Compute XD685 configured with any supported GPUs, system fans must operate at 100 percent to ensure safe thermal conditions for all GPU configurations.IMPORTANT:Due to current software and firmware limitations, any modification to the fan speed settings may result in insufficient cooling. As a result, any installed GPUs may experience severe hardware damage. Refer to the Resolution section for additional information.HPE strongly recommends the following:Do Not modify fan speed settings or thermal profiles.Ensure system management tools maintain fan operation at 100%.Monitor system health and GPU temperatures regularly.
In the scenario described above, any HPE ProLiant Compute XD685 configured with any supported GPUs.
A future software/firmware version will avoid this issue. This advisory will be updated when additional solutions become available.Troubleshooting TipsFor any fan speed or GPU thermal issues, perform the following steps:1. Verify Fan Speed Settings:Use iLO or a system management tool to confirm fans are running at 100%.Check BIOS/UEFI settings for any overrides or custom profiles.2. Monitor GPU Temperatures:Use HPE Insight Diagnostics or third-party GPU monitoring tools.Look for signs of thermal throttling or unexpected shutdowns.3. Check Firmware Versions:Ensure the following versions of firmware are running for the system board, iLO, and GPU:iLO6 1.71 (or later)BIOS A59_2.52_10_03_2025 (or later)HPC NVIDIA H200GPU Air-Cooled I2C NON MCTP Firmware Pack Version 1.7.0 (or later)HPC NVIDIA H200GPU Direct Liquid Cooling NON MCTP Firmware Pack Version 1.7.0 (or later)HPC NVIDIA B200 Direct Liquid Cooling USB Firmware Pack Version 1.2.0 (or later)Avoid using beta or custom firmware unless directed by HPE support.For the latest firmware downloads, documentation, and support tools, visit the HPE ProLiant Compute XD685 support page:https://support.hpe.com/connect/s/product?language=en_US&kmpmoid=1014862906&tab=knowledgeArticles4. Review System Logs:Look for thermal-related warnings or errors in iLO logs or OS event logs.Pay attention to any fan failure or GPU fault messages.5. Contact Support:If abnormal behavior occurs or thermal damage is suspected, contact HPE Customer Support.Provide system logs and configuration details to expedite troubleshooting.Frequently Asked Questions (FAQ)Q1: Can I reduce the fan speed to lower noise levels?A: No. Reducing fan speed may cause thermal damage to GPUs. Fans must remain at 100% until a validated solution is released.Q2: Does this advisory apply to all GPU types?A: Yes. This applies to all GPU models supported on the HPE ProLiant Compute XD685 platform.Q3: Is there a workaround available?A: No workaround is currently available. HPE is actively working on a solution.Q4: How will I know when the fix is available?A: HPE will update this advisory and issue an HPE Support Alert through support channels. Subscribe toHPE Support Alertsor contact your HPE representative.Q5: What risks are involved if I ignore this advisory?A: Ignoring this advisory may result in overheating, GPU failure, and potential system damage.
Operating Systems Affected:Not Applicable
Click on a version to see all relevant bugs
Hewlett Packard Enterprise Integration
Learn more about where this data comes from
BugZero Plan
Streamline upgrades with automated vendor bug scrubs
BugZero Prevent
Wish you caught this bug sooner? Get proactive today.