# Confidential Computing **URL:** https://forums.developer.nvidia.com/c/accelerated-computing/confidential-computing/663.md [Latest](https://forums.developer.nvidia.com/latest.md) · [Categories](https://forums.developer.nvidia.com/categories.md) · [Tags](https://forums.developer.nvidia.com/tags.md) --- ## [Welcome & General Information!](https://forums.developer.nvidia.com/t/welcome-general-information/260554)
**Author:** [@rnertney](https://forums.developer.nvidia.com/u/rnertney)\ **Replies:** 0\ **Last updated:** [July 20, 2023, 4:53pm UTC](https://forums.developer.nvidia.com/t/welcome-general-information/260554 "2023-07-20T16:53:38Z")
Welcome to the NVIDIA Confidential Computing Forum! Please feel free to provide feedback, ask questions, leave comments, or report issues here. For Getting Started Information, please check out our GitHub repository. Th… --- ## [Cryptographic binding between NVIDIA GPU attestation and actual CUDA execution context](https://forums.developer.nvidia.com/t/cryptographic-binding-between-nvidia-gpu-attestation-and-actual-cuda-execution-context/385119)
**Author:** [@partners21](https://forums.developer.nvidia.com/u/partners21)\ **Replies:** 4\ **Last updated:** [October 13, 2026, 7:48pm UTC](https://forums.developer.nvidia.com/t/cryptographic-binding-between-nvidia-gpu-attestation-and-actual-cuda-execution-context/385119 "2026-10-13T19:48:06Z")
We are designing a fail-closed GPU workload verification system using NVIDIA GPU Attestation / NVAT and CUDA. We need to determine whether NVIDIA provides a supported mechanism that cryptographically binds verified GPU … --- ## [NVSwitch attestation from inside a TDX guest: nvattest 1.2.0 fails with NSCQ\_RC\_WARNING: RDT init failure (Code 1)](https://forums.developer.nvidia.com/t/nvswitch-attestation-from-inside-a-tdx-guest-nvattest-1-2-0-fails-with-nscq-rc-warning-rdt-init-failure-code-1/383536)
**Author:** [@voltagegpu](https://forums.developer.nvidia.com/u/voltagegpu)\ **Replies:** 0\ **Last updated:** [September 17, 2026, 12:40pm UTC](https://forums.developer.nvidia.com/t/nvswitch-attestation-from-inside-a-tdx-guest-nvattest-1-2-0-fails-with-nscq-rc-warning-rdt-init-failure-code-1/383536 "2026-09-17T12:40:19Z")
Setup HGX node with 8x H100 80GB HBM3, driver 595.71.05, tenant VM is an Intel TDX guest (Ubuntu 24.04 image, kernel 6.8.0-110-generic, /dev/tdx\_guest present). GPUs run in multi-GPU Protected PCIe mode (nvidia-smi con… --- ## [RIM Bundle Not Found on H200 Attestation](https://forums.developer.nvidia.com/t/rim-bundle-not-found-on-h200-attestation/382686)
**Author:** [@cameronpwilding](https://forums.developer.nvidia.com/u/cameronpwilding)\ **Replies:** 0\ **Last updated:** [September 8, 2026, 9:57am UTC](https://forums.developer.nvidia.com/t/rim-bundle-not-found-on-h200-attestation/382686 "2026-09-08T09:57:51Z")
I am attempting to test nvattest on an H200 GPU. However, whenever I run nvattest attest --device gpu --verifier local as is suggested in the docs, I get back 2026-09-08 09:50:51.323 \[/home/cw/attestation-sdk/nv-atte… --- ## [RTX PRO 4500 Blackwell fails in TDX CC-On (0x62:0x65:2119) — which TRD release will add support?](https://forums.developer.nvidia.com/t/rtx-pro-4500-blackwell-fails-in-tdx-cc-on-0x622119-which-trd-release-will-add-support/381959)
**Author:** [@doohkim71](https://forums.developer.nvidia.com/u/doohkim71)\ **Replies:** 1\ **Last updated:** [September 1, 2026, 4:44pm UTC](https://forums.developer.nvidia.com/t/rtx-pro-4500-blackwell-fails-in-tdx-cc-on-0x622119-which-trd-release-will-add-support/381959 "2026-09-01T16:44:34Z")
We are testing single-GPU passthrough Confidential Computing with an Intel TDX confidential VM and two NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs. Both physical cards reproduce the same behavior: PCI ID: 10de:2… --- ## [Is Secure Boot / firmware authentication enabled by default on non-Confidential-Computing GPUs such as A100?](https://forums.developer.nvidia.com/t/is-secure-boot-firmware-authentication-enabled-by-default-on-non-confidential-computing-gpus-such-as-a100/376214)
**Author:** [@qing18](https://forums.developer.nvidia.com/u/qing18)\ **Replies:** 1\ **Last updated:** [August 31, 2026, 5:07pm UTC](https://forums.developer.nvidia.com/t/is-secure-boot-firmware-authentication-enabled-by-default-on-non-confidential-computing-gpus-such-as-a100/376214 "2026-08-31T17:07:55Z")
My understanding is that NVIDIA GPUs have supported some form of firmware protection for several generations. For example, Maxwell-generation GPUs appear to have introduced anti-tamper mechanisms, where the hardware star… --- ## [NVENC nvEncInitializeEncoder returns NV\_ENC\_ERR\_OUT\_OF\_MEMORY (10) on RTX PRO 6000 Blackwell Server Edition in SPT CC; NVDEC works;](https://forums.developer.nvidia.com/t/nvenc-nvencinitializeencoder-returns-nv-enc-err-out-of-memory-10-on-rtx-pro-6000-blackwell-server-edition-in-spt-cc-nvdec-works/381553)
**Author:** [@1224533208](https://forums.developer.nvidia.com/u/1224533208)\ **Replies:** 1\ **Last updated:** [August 31, 2026, 5:06am UTC](https://forums.developer.nvidia.com/t/nvenc-nvencinitializeencoder-returns-nv-enc-err-out-of-memory-10-on-rtx-pro-6000-blackwell-server-edition-in-spt-cc-nvdec-works/381553 "2026-08-31T05:06:24Z")
Environment GPU: 1× NVIDIA RTX PRO 6000 Blackwell Server Edition (96 GB), passthrough to one CVM GPU CC: SPT (CC State = ON). Not Hopper PPCIe. Not Workstation / Max-Q SKU. CPU TEE: Intel TDX guest OS: Ubuntu 24.04 (ke… --- ## [Enabling Multi-GPU CC with H100](https://forums.developer.nvidia.com/t/enabling-multi-gpu-cc-with-h100/381535)
**Author:** [@leejongh](https://forums.developer.nvidia.com/u/leejongh)\ **Replies:** 2\ **Last updated:** [August 27, 2026, 10:18pm UTC](https://forums.developer.nvidia.com/t/enabling-multi-gpu-cc-with-h100/381535 "2026-08-27T22:18:03Z")
Hi. I have a server machine with Intel ER and 2 H100 GPUs. Reading the deployment guide, you guys now allow Multi-GPU CC with B200 GPUs. I was wondering, is there any way to work around enabling this multi-GPU CC enviro… --- ## [I couldn't find a Jetson to book in RunPod so I created RunSnack](https://forums.developer.nvidia.com/t/i-couldnt-find-a-jetson-to-book-in-runpod-so-i-created-runsnack/379749)
**Author:** [@smartcityjaen](https://forums.developer.nvidia.com/u/smartcityjaen)\ **Replies:** 3\ **Last updated:** [August 26, 2026, 7:10am UTC](https://forums.developer.nvidia.com/t/i-couldnt-find-a-jetson-to-book-in-runpod-so-i-created-runsnack/379749 "2026-08-26T07:10:47Z")
Hey everyone, quick story. I was looking for a Jetson Nano to test some edge AI stuff and realized none of the cloud providers have them. RunPod doesn’t offer it, Vast.ai doesn’t either. So I ended up building my own sol… --- ## [H100 CC: Is there a GSP timeout between SPDM KEY\_EXCHANGE\_RSP and FINISH?](https://forums.developer.nvidia.com/t/h100-cc-is-there-a-gsp-timeout-between-spdm-key-exchange-rsp-and-finish/380421)
**Author:** [@eliot583](https://forums.developer.nvidia.com/u/eliot583)\ **Replies:** 0\ **Last updated:** [August 17, 2026, 12:33pm UTC](https://forums.developer.nvidia.com/t/h100-cc-is-there-a-gsp-timeout-between-spdm-key-exchange-rsp-and-finish/380421 "2026-08-17T12:33:17Z")
I am testing SPDM session establishment on an H100 in Confidential Computing mode and noticed that delaying the FINISH message after receiving KEY\_EXCHANGE\_RSP appears to fail with a timeout somewhere between approximate… --- ## [Can H100/H200 CC attestation give us info about which specific process ran on the enclave?](https://forums.developer.nvidia.com/t/can-h100-h200-cc-attestation-give-us-info-about-which-specific-process-ran-on-the-enclave/380183)
**Author:** [@hagrid175bil](https://forums.developer.nvidia.com/u/hagrid175bil)\ **Replies:** 0\ **Last updated:** [August 14, 2026, 8:56am UTC](https://forums.developer.nvidia.com/t/can-h100-h200-cc-attestation-give-us-info-about-which-specific-process-ran-on-the-enclave/380183 "2026-08-14T08:56:45Z")
My understanding of this is that the GPU attestation (via NVIDIA’s attestation SDK / NRAS) would report measurements of the GPU’s firmware, driver, and VBIOS against known-good reference values, plus confirms the confi… --- ## [Request for CC-enabled VBIOS for NVIDIA B200 HGX (40 GPUs)](https://forums.developer.nvidia.com/t/request-for-cc-enabled-vbios-for-nvidia-b200-hgx-40-gpus/374384)
**Author:** [@izulfiya1986](https://forums.developer.nvidia.com/u/izulfiya1986)\ **Replies:** 1\ **Last updated:** [July 3, 2026, 11:56am UTC](https://forums.developer.nvidia.com/t/request-for-cc-enabled-vbios-for-nvidia-b200-hgx-40-gpus/374384 "2026-07-03T11:56:09Z")
Hello NVIDIA Team, We are deploying Confidential Computing workloads and require CC-enabled VBIOS for our NVIDIA B200 HGX GPUs. Deployment Details: GPU Model: NVIDIA B200 HGX (5 nodes x 8 GPUs = 40 GPUs total) Host E… --- ## [CC VBIOS request for NVIDIA B200 HGX](https://forums.developer.nvidia.com/t/cc-vbios-request-for-nvidia-b200-hgx/374221)
**Author:** [@ejcpqk](https://forums.developer.nvidia.com/u/ejcpqk)\ **Replies:** 0\ **Last updated:** [June 23, 2026, 9:51am UTC](https://forums.developer.nvidia.com/t/cc-vbios-request-for-nvidia-b200-hgx/374221 "2026-06-23T09:51:00Z")
Hello NVIDIA team, We need CC-enabled VBIOS for NVIDIA B200 HGX GPUs to deploy confidential computing workloads GPU Model: NVIDIA B200 HGX (8GPUs) Serial numbers: A936564X5C17285 A936564X5C17286 A936564X5C17288 … --- ## [From Governance Runtime Assurance to Human-Directed Intelligence](https://forums.developer.nvidia.com/t/from-governance-runtime-assurance-to-human-directed-intelligence/373803)
**Author:** [@bpm\_red\_academy](https://forums.developer.nvidia.com/u/bpm_red_academy)\ **Replies:** 7\ **Last updated:** [June 19, 2026, 2:12pm UTC](https://forums.developer.nvidia.com/t/from-governance-runtime-assurance-to-human-directed-intelligence/373803 "2026-06-19T14:12:19Z")
From Governance Runtime Assurance to Human-Directed IntelligenceWorking note from BPM RED Academy / HumAI MightHub development In my previous posts, I explored the progression from deterministic inference toward governan… --- ## [Looking for archived DU-12302-001\_v5.0 Confidential Computing Deployment Guide](https://forums.developer.nvidia.com/t/looking-for-archived-du-12302-001-v5-0-confidential-computing-deployment-guide/373017)
**Author:** [@zhan31220](https://forums.developer.nvidia.com/u/zhan31220)\ **Replies:** 0\ **Last updated:** [June 11, 2026, 2:42pm UTC](https://forums.developer.nvidia.com/t/looking-for-archived-du-12302-001-v5-0-confidential-computing-deployment-guide/373017 "2026-06-11T14:42:08Z")
Hi everyone, I am looking for an archived copy of the NVIDIA Confidential Computing Deployment Guide. DU-12302-001\_v5.0 Date: February 25, 2025 In the documentation history of the current DU-12302-001\_v7.1 guide, ver… --- ## [Is the Blackwell GPU RAM encrypted?](https://forums.developer.nvidia.com/t/is-the-blackwell-gpu-ram-encrypted/372747)
**Author:** [@mdb326](https://forums.developer.nvidia.com/u/mdb326)\ **Replies:** 0\ **Last updated:** [June 9, 2026, 8:29pm UTC](https://forums.developer.nvidia.com/t/is-the-blackwell-gpu-ram-encrypted/372747 "2026-06-09T20:29:39Z")
I am fairly confident the answer is no but have not found any source to verify this and have found one (fairly non-reputable) source claiming that it is. --- ## [Seeking Confidential Computing Configuration Guidance for Blackwell Pro 6000 Workstation Edition](https://forums.developer.nvidia.com/t/seeking-confidential-computing-configuration-guidance-for-blackwell-pro-6000-workstation-edition/346649)
**Author:** [@keke](https://forums.developer.nvidia.com/u/keke)\ **Replies:** 2\ **Last updated:** [May 25, 2026, 8:47am UTC](https://forums.developer.nvidia.com/t/seeking-confidential-computing-configuration-guidance-for-blackwell-pro-6000-workstation-edition/346649 "2026-05-25T08:47:49Z")
I have attempted to deploy Confidential Computing on the Blackwell Pro 6000 Workstation Edition by following the instructions in the NVIDIA Confidential Computing Deployment Guide (https://docs.nvidia.com/cc-deployment-g… --- ## [Cluster Networking with Nvidia CC](https://forums.developer.nvidia.com/t/cluster-networking-with-nvidia-cc/370078)
**Author:** [@amean](https://forums.developer.nvidia.com/u/amean)\ **Replies:** 0\ **Last updated:** [May 14, 2026, 12:02am UTC](https://forums.developer.nvidia.com/t/cluster-networking-with-nvidia-cc/370078 "2026-05-14T00:02:59Z")
Based on the current Secure AI Guide, GPUDirect RDMA is disabled which means that in CC mode any RDMA based cluster networking is not possible. Curious if with TEE I/O or the upcoming Vera Rubin generation if RDMA will … --- ## [cuFFT (libcufft) crashes on H100 in Confidential Computing (CC) mode](https://forums.developer.nvidia.com/t/cufft-libcufft-crashes-on-h100-in-confidential-computing-cc-mode/368342)
**Author:** [@asder4444](https://forums.developer.nvidia.com/u/asder4444)\ **Replies:** 1\ **Last updated:** [April 29, 2026, 12:53am UTC](https://forums.developer.nvidia.com/t/cufft-libcufft-crashes-on-h100-in-confidential-computing-cc-mode/368342 "2026-04-29T00:53:24Z")
Has anyone seen the following issue? ## Summary \`cufftEstimate1d(nx=512, type=CUFFT\_R2C, batch=1, &workSize)\` — documented as a pure host-side workspace-size lookup that does \*\*not\*\* touch the device — returns \`CUFFT… --- ## [MIG + Confidential Computing](https://forums.developer.nvidia.com/t/mig-confidential-computing/360895)
**Author:** [@kayondo](https://forums.developer.nvidia.com/u/kayondo)\ **Replies:** 3\ **Last updated:** [April 26, 2026, 4:47am UTC](https://forums.developer.nvidia.com/t/mig-confidential-computing/360895 "2026-04-26T04:47:46Z")
Hi all, I am wondering if anyone has succeeded in running MIG when confidential computing is enabled. NVIDIA advertises this as a possible setting, but trying to enable MIG when CC is on says MIG is unsupported, despit… --- ## [Does MIG/TEE isolation boundary fail upon host kernel compromise](https://forums.developer.nvidia.com/t/does-mig-tee-isolation-boundary-fail-upon-host-kernel-compromise/366923)
**Author:** [@ryan353](https://forums.developer.nvidia.com/u/ryan353)\ **Replies:** 0\ **Last updated:** [April 17, 2026, 8:12pm UTC](https://forums.developer.nvidia.com/t/does-mig-tee-isolation-boundary-fail-upon-host-kernel-compromise/366923 "2026-04-17T20:12:08Z")
​I am architecting a high-security local AI deployment using Multi-Instance GPU (MIG) and Trusted Execution Environments (TEEs) to isolate an untrusted, self-evolving LLM from the host system. ​While MIG partitions the … --- ## [Multi-Tenant Support in NVIDIA Confidential Computing (CC) on a Single GPU](https://forums.developer.nvidia.com/t/multi-tenant-support-in-nvidia-confidential-computing-cc-on-a-single-gpu/363235)
**Author:** [@hojun.choe](https://forums.developer.nvidia.com/u/hojun.choe)\ **Replies:** 3\ **Last updated:** [March 14, 2026, 10:25am UTC](https://forums.developer.nvidia.com/t/multi-tenant-support-in-nvidia-confidential-computing-cc-on-a-single-gpu/363235 "2026-03-14T10:25:43Z")
Hi, I have a question regarding whether NVIDIA Confidential Computing (CC) currently provides protection for multiple VMs in a multi-tenant environment on a single GPU. In Figure 5 of the initial NVIDIA CC whitepaper, … --- ## [Questions from a newcomer — GPU Confidential Computing for multi-tenant inference](https://forums.developer.nvidia.com/t/questions-from-a-newcomer-gpu-confidential-computing-for-multi-tenant-inference/362489)
**Author:** [@xichen0425](https://forums.developer.nvidia.com/u/xichen0425)\ **Replies:** 1\ **Last updated:** [March 6, 2026, 2:04pm UTC](https://forums.developer.nvidia.com/t/questions-from-a-newcomer-gpu-confidential-computing-for-multi-tenant-inference/362489 "2026-03-06T14:04:00Z")
Body: Hi all, I’m fairly new to the GPU world and have been spending time researching confidential inference — specifically protecting proprietary model weights on multi-tenant H100/B200 infrastructure. I’ve been readin… --- ## [Cryptographically sealing AI intent payloads (TEEs) while awaiting physical GPIO actuation](https://forums.developer.nvidia.com/t/cryptographically-sealing-ai-intent-payloads-tees-while-awaiting-physical-gpio-actuation/361946)
**Author:** [@johngreetme](https://forums.developer.nvidia.com/u/johngreetme)\ **Replies:** 0\ **Last updated:** [February 27, 2026, 11:52pm UTC](https://forums.developer.nvidia.com/t/cryptographically-sealing-ai-intent-payloads-tees-while-awaiting-physical-gpio-actuation/361946 "2026-02-27T23:52:51Z")
We are deploying a decentralized compute network where nodes generate physical/network actions locally. We utilize a “State-Locked Protocol” (SLP). When the AI generates an action payload, it enters an “Amber Zone”, a ho… --- ## [Inquiry on Checkpointing on Confidential Computing](https://forums.developer.nvidia.com/t/inquiry-on-checkpointing-on-confidential-computing/361667)
**Author:** [@leejongh](https://forums.developer.nvidia.com/u/leejongh)\ **Replies:** 3\ **Last updated:** [February 26, 2026, 11:46pm UTC](https://forums.developer.nvidia.com/t/inquiry-on-checkpointing-on-confidential-computing/361667 "2026-02-26T23:46:09Z")
I have been trying to run checkpointing (github cuda-checkpoint & CUDA Driver API) on Confidential Computing but I have been getting the error message “operation not supported”. Does Confidential Computing by design not … --- ## [Inquiry About Compatible Commit on sev-snp-devel Branch (Ubuntu 22.04)](https://forums.developer.nvidia.com/t/inquiry-about-compatible-commit-on-sev-snp-devel-branch-ubuntu-22-04/360355)
**Author:** [@esp](https://forums.developer.nvidia.com/u/esp)\ **Replies:** 7\ **Last updated:** [February 19, 2026, 3:34am UTC](https://forums.developer.nvidia.com/t/inquiry-about-compatible-commit-on-sev-snp-devel-branch-ubuntu-22-04/360355 "2026-02-19T03:34:03Z")
Hi, I am currently trying to set up a CC environment on Ubuntu 22.04 referencing the old version of the deployment guide. However, I noticed that there have been substantial code changes in the AMDSEV repository, when … --- ## [NVIDIA B200: NCCL WARN Cuda failure 700 'an illegal memory access was encountered'](https://forums.developer.nvidia.com/t/nvidia-b200-nccl-warn-cuda-failure-700-an-illegal-memory-access-was-encountered/360195)
**Author:** [@venishpatidar](https://forums.developer.nvidia.com/u/venishpatidar)\ **Replies:** 5\ **Last updated:** [February 19, 2026, 3:32am UTC](https://forums.developer.nvidia.com/t/nvidia-b200-nccl-warn-cuda-failure-700-an-illegal-memory-access-was-encountered/360195 "2026-02-19T03:32:07Z")
Multi-GPU NCCL 700 'an illegal memory access Failure on NVIDIA B200 GPUs I am encountering NCCL WARN Cuda failure 700 'an illegal memory access was encountered' errors when running both nccl-tests and vLLM with multi-GPU… --- ## [GPUDirect RDMA in CC mode](https://forums.developer.nvidia.com/t/gpudirect-rdma-in-cc-mode/359783)
**Author:** [@teofil.bodea](https://forums.developer.nvidia.com/u/teofil.bodea)\ **Replies:** 0\ **Last updated:** [February 5, 2026, 3:14pm UTC](https://forums.developer.nvidia.com/t/gpudirect-rdma-in-cc-mode/359783 "2026-02-05T15:14:39Z")
Hello everyone, this document \[1\] mentions that GPUDirect RDMA is not available in CC mode. I was wondering if there are any plans to introduce it in the future, and, if not, what would be the main challenges preventing… --- ## [How memory OOM concern is fixed in CC mode](https://forums.developer.nvidia.com/t/how-memory-oom-concern-is-fixed-in-cc-mode/357229)
**Author:** [@springchensjtu](https://forums.developer.nvidia.com/u/springchensjtu)\ **Replies:** 4\ **Last updated:** [January 15, 2026, 3:55am UTC](https://forums.developer.nvidia.com/t/how-memory-oom-concern-is-fixed-in-cc-mode/357229 "2026-01-15T03:55:46Z")
Hi, I want to raise a general question that How the OOM risk is handled in Hopper confidential computing mode. Regarding to H100 CC white pape, cudaMallocHost memory will be handled with UVM paging mechanism like the m… --- ## [NVIDIA H100 driver not loading on Ubuntu 25.x (kernel 6.14) nvidia-smi fails](https://forums.developer.nvidia.com/t/nvidia-h100-driver-not-loading-on-ubuntu-25-x-kernel-6-14-nvidia-smi-fails/357337)
**Author:** [@Anubhav\_Gupta](https://forums.developer.nvidia.com/u/Anubhav_Gupta)\ **Replies:** 1\ **Last updated:** [January 12, 2026, 1:02pm UTC](https://forums.developer.nvidia.com/t/nvidia-h100-driver-not-loading-on-ubuntu-25-x-kernel-6-14-nvidia-smi-fails/357337 "2026-01-12T13:02:56Z")
Hello NVIDIA team, I am trying to install and use NVIDIA drivers on a system with an NVIDIA H100 (GH100 PCIe) GPU, but I am unable to get nvidia-smi working. System details GPU: NVIDIA H100 (GH100 PCIe) OS: Ubuntu… [Next page](https://forums.developer.nvidia.com/c/accelerated-computing/confidential-computing/663.md?page=1)