Troubleshooting RTX 4000 SFF Ada Passthrough GPU Reset Failure on ESXi 8 with Minisforum MS-02

A GPU passthrough checklist for RTX 4000 SFF Ada cards on ESXi 8 and Minisforum MS-02 hosts, covering firmware, VM settings, and validation.

Quick Read

Symptom: A GPU passthrough checklist for RTX 4000 SFF Ada cards on ESXi 8 and Minisforum MS-02 hosts, covering firmware, VM settings, and validation.
Check first: Confirm the ESXi build, host BIOS version, VM firmware mode, and whether the RTX 4000 SFF Ada appears in the same IOMMU group as other devices.
Risk: Changes system state

Symptoms

VM fails to power on with error code 88 due to GPU reset failure when using RTX 4000 SFF Ada passthrough on ESXi 8.

Environment

VMware ESXi 8, RTX 4000 SFF Ada GPU, Minisforum MS-02 hardware.

Most Likely Causes

The GPU reset failure may be caused by improper configuration of passthrough settings, BIOS settings, or compatibility issues between the ESXi version and the GPU.

What to Check First

Confirm the ESXi build, host BIOS version, VM firmware mode, and whether the RTX 4000 SFF Ada appears in the same IOMMU group as other devices.
Capture the exact ESXi power-on error, vmkernel timestamp, VMX log timestamp, and whether the failure occurs after a host reboot, guest reboot, or only after a prior VM crash.
Verify whether other PCI passthrough devices on the Minisforum MS-02 work correctly before changing GPU-specific settings.

Fix Steps

Verify GPU Compatibility
Ensure that the RTX 4000 SFF Ada is compatible with ESXi 8 for passthrough.
Check BIOS Settings
Access the BIOS of the Minisforum MS-02 and ensure that the following settings are configured correctly:
Example pattern only. Adjust for your environment before running.
```
Enable 'Above 4G Decoding'
Enable 'IOMMU' or 'VT-d' (if available)
Disable 'Secure Boot' (if applicable)
```
Configure ESXi for Passthrough
Log into the ESXi host and configure the RTX 4000 SFF Ada for passthrough.
Example pattern only. Adjust for your environment before running.
```
Navigate to 'Host' > 'Manage' > 'Hardware' > 'PCI Devices'
Locate the RTX 4000 SFF Ada GPU and click 'Toggle passthrough'
Reboot the ESXi host to apply changes
```
Create or Edit the Virtual Machine
Ensure the virtual machine is configured to use the passthrough GPU.
Example pattern only. Adjust for your environment before running.
```
Edit the VM settings
Add a new PCI device and select the RTX 4000 SFF Ada GPU
Ensure the VM is set to use UEFI firmware
```

Adjust VM Settings for GPU

Modify the VM settings to optimize for GPU usage.

Example pattern only. Adjust for your environment before running.

Set 'Guest OS' to a compatible version (e.g., Windows 10/11 or Linux with NVIDIA drivers)
Allocate sufficient CPU and RAM resources to the VM
Enable 3D acceleration if required

Install NVIDIA Drivers in the VM
Ensure that the appropriate NVIDIA drivers are installed in the guest operating system.
Example pattern only. Adjust for your environment before running.
```
Download the latest NVIDIA drivers from the official NVIDIA website
Install the drivers following the installation instructions provided
```
Test VM Power-On
Attempt to power on the VM to check if the issue is resolved.
Example pattern only. Adjust for your environment before running.
```
Select the VM and click 'Power On'
Monitor for any error messages during the power-on process
```

Validation

After each change, the VM powers on with the passthrough GPU attached and the ESXi Recent Tasks pane shows the task completed without error code 88.
The guest operating system detects the RTX 4000 SFF Ada and loads the expected NVIDIA driver without Code 43, reset, or device-start errors.
A host reboot followed by a cold VM power-on still works, which proves the fix survives the reset path that originally failed.

Logs to Check

VMware VMX log in the virtual machine folder for PCI passthrough and device reset errors.
ESXi vmkernel log around the failed power-on timestamp for PCIe reset, IOMMU, or passthrough messages.
Host hardware event logs or IPMI/BIOS event history if the GPU disappears after warm reboot or power-state changes.

Rollback and Escalation

Record the original BIOS, passthrough, and VM hardware settings before changing them.
Remove the passthrough PCI device from the VM before reverting host passthrough or BIOS settings.
If a change prevents the host or VM from booting cleanly, disable GPU passthrough, reboot the ESXi host, and return the VM to console-only access before retesting.

Escalate When

Escalate if the GPU fails reset after a host cold boot and ESXi logs show PCIe reset errors for the same device.
Escalate if the GPU shares an IOMMU group with storage, network, or chipset devices that cannot safely be isolated on the MS-02 platform.

Edge Cases

If the VM fails to power on again, check for any additional error codes in the ESXi logs.
Verify that no other devices are conflicting with the passthrough settings.

Notes from the Field

Small-form-factor hosts can expose platform-specific PCIe reset behavior. Treat every BIOS and passthrough change as a host-impacting change.
Avoid testing GPU passthrough fixes on the only management path to the ESXi host. Keep out-of-band or local console access available.