Error: Purple Screen Of Death PSOD

Products

• VMware vSphere ESXi

Issue/Introduction

• ESXi host experienced PSOD
• You can see error messages in /var/run/log/vmkernel.log

vmkernel: cpu20:######)NVRM: GPU at PCI:######: GPU-###### In(182) vmkernel: cpu20:######)NVRM: GPU Board Serial Number: ######
In(182) vmkernel: cpu20:######)NVRM: Xid (PCI:######): ######, pid=######, name=, Timeout after 4s of waiting for RPC response from GPU0 GSP! 
Expected function 76 (GSP_RM_CONTROL) (############).
2025-05-05T08:31:37.702Z In(182) vmkernel: cpu20:######)NVRM: GPU0 GSP RPC buffer contains function 76 (######) and data ######
2025-05-05T08:31:37.702Z In(182) vmkernel: cpu20:######)NVRM: GPU0 RPC history (CPU -> GSP):
2025-05-05T08:31:37.702Z In(182) vmkernel: cpu20:######)NVRM:     entry function                   data0              data1              ts_start           ts_end             duration actively_polling2

05T08:31:37.702Z In(182) vmkernel: cpu20:######)NVRM:      0    ###### GSP_SEND_USER_SHARED_ ############      
2025-05-05T08:31:37.702Z In(182) vmkernel: cpu2######)NVRM:     -1    ######

Reset for device failed with Failure
2025-05-05T09:14:09.190Z In(182) vmkernel: cpu59:######)PCIPassthru: ######: pcipdevInfo: ###### (######), state 0, destroyed
2025-05-05T09:14:09.191Z Wa(180) vmkwarning: cpu48:######1)WARNING: PCI: 745: Dev ###### is unresponsive after reset

 

Environment

VMware vSphere 8.x

VMware vSphere 7.x

Cause

NVIDIA drivers 

Resolution

Please engage NVIDIA to address the issue of unresponsive device