[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-08-18。"],[],[],null,["# Verify Confidential Computing is enabled on GPUs\n\nAfter completing\n[the configuration steps](/confidential-computing/confidential-vm/docs/create-a-confidential-vm-instance-with-gpu),\nverify that your GPU is correctly operating in confidential computing mode:\n\n1. Verify that the `nvidia-persistenced` service is running with\n `uvm-persistence-mode` enabled:\n\n ps aux | grep nvidia-persistenced | grep -v grep\n\n The output is similar to the following: \n\n nvidia-+ 729 6.2 0.0 5468 1664 ? Ss 12:02 0:10\n /usr/bin/nvidia-persistenced \\\n --user nvidia-persistenced \\\n --uvm-persistence-mode \\\n --verbose\n\n | **Caution:** Don't restart the `nvidia-persistenced` service after enabling `uvm-persistence-mode`, as this can disrupt the secure connection and require a VM reboot.\n2. Confirm that the GPU is running in confidential computing mode:\n\n sudo nvidia-smi conf-compute -f\n\n The output is similar to the following: \n\n CC status: ON\n\n3. Set GPU to ready state after each reboot.\n\n sudo nvidia-smi conf-compute -srs 1\n\n4. Verify Confidential Computing ready state.\n\n sudo nvidia-smi conf-compute -grs\n\n If you transition a GPU out of the ready state using\n `nvidia-smi conf-compute -srs 0`, reboot the VM instance to return the GPU\n to the ready state.\n\n The output is similar to the following: \n\n Confidential Compute GPUs Ready state: ready"]]