“representative burn-in keeps package temperatures below alert thresholds” addresses cooling design underestimates real application heat dissipation, but the current hardware or infrastructure issue is site-level efficiency metrics cannot replace device-level thermal validation. The option “per-node thermal histories satisfy load-specific thresholds despite aggregate PUE” is more directly tied to the required outcome.

“combined power and benchmark traces satisfy both electrical and throughput bounds” addresses power throttling policy conflicts with approved application requirements, but the current hardware or infrastructure issue is site-level efficiency metrics cannot replace device-level thermal validation. The option “per-node thermal histories satisfy load-specific thresholds despite aggregate PUE” is more directly tied to the required outcome.

“measured full-load draw remains under all configured branch and PDU limits” addresses peak electrical demand exceeds the approved power delivery envelope, but the current hardware or infrastructure issue is site-level efficiency metrics cannot replace device-level thermal validation. The option “per-node thermal histories satisfy load-specific thresholds despite aggregate PUE” is more directly tied to the required outcome.

“validate installed CUDA runtime and application compatibility against the driver” addresses driver visibility was mistaken for full application-runtime readiness, but the current hardware or infrastructure issue is host driver registration has not established visibility of installed GPUs. The option “verify the supported driver kernel modules and device enumeration” is more directly tied to the required outcome.

Useful for interconnect readiness is not captured in GPU bring-up acceptance, but it does not remove the specific condition: host driver registration has not established visibility of installed GPUs.

“configure supported GPU persistence behavior for the environment” addresses device persistence settings do not match the intended repeated-job operating profile, but the current hardware or infrastructure issue is host driver registration has not established visibility of installed GPUs. The option “verify the supported driver kernel modules and device enumeration” is more directly tied to the required outcome.

“a sample CUDA application initializes and reports the expected device” addresses driver visibility was mistaken for full application-runtime readiness, but the current hardware or infrastructure issue is host driver registration has not established visibility of installed GPUs. The option “nvidia-smi lists the intended GPU count with no driver initialization errors” is more directly tied to the required outcome.

“repeated short-job startup measurements show reduced initialization overhead” addresses device persistence settings do not match the intended repeated-job operating profile, but the current hardware or infrastructure issue is host driver registration has not established visibility of installed GPUs. The option “nvidia-smi lists the intended GPU count with no driver initialization errors” is more directly tied to the required outcome.

“topology display and local bandwidth checks match the approved HGX design” addresses interconnect readiness is not captured in GPU bring-up acceptance, but the current hardware or infrastructure issue is host driver registration has not established visibility of installed GPUs. The option “nvidia-smi lists the intended GPU count with no driver initialization errors” is more directly tied to the required outcome.

Useful for ephemeral GPU index positions are unsuitable as maintenance identities, but it does not remove the specific condition: accelerator identity lacks a stable mapping to the physical PCIe slot.

“check GPU seating device power and driver health before re-adding node to scheduling” addresses a physical or driver-level enumeration fault affects one accelerator, but the current hardware or infrastructure issue is accelerator identity lacks a stable mapping to the physical PCIe slot. The option “correlate nvidia-smi PCI bus identifiers with chassis service labels” is more directly tied to the required outcome.

“run a bounded representative CUDA workload while monitoring nvidia-smi” addresses an idle utilization sample does not test compute functionality, but the current hardware or infrastructure issue is accelerator identity lacks a stable mapping to the physical PCIe slot. The option “correlate nvidia-smi PCI bus identifiers with chassis service labels” is more directly tied to the required outcome.

“a sample CUDA application initializes and reports the expected device” addresses driver visibility was mistaken for full application-runtime readiness, but the current hardware or infrastructure issue is accelerator identity lacks a stable mapping to the physical PCIe slot. The option “an inventory sample maps every reported GPU ID to the expected physical slot” is more directly tied to the required outcome.

“device inventory meets requirements and the test allocates memory without failure” addresses scheduling readiness did not confirm required device memory and capabilities, but the current hardware or infrastructure issue is accelerator identity lacks a stable mapping to the physical PCIe slot. The option “an inventory sample maps every reported GPU ID to the expected physical slot” is more directly tied to the required outcome.

“nvidia-smi and platform logs show all eight devices healthy” addresses a physical or driver-level enumeration fault affects one accelerator, but the current hardware or infrastructure issue is accelerator identity lacks a stable mapping to the physical PCIe slot. The option “an inventory sample maps every reported GPU ID to the expected physical slot” is more directly tied to the required outcome.

“verify the supported driver kernel modules and device enumeration” addresses host driver registration has not established visibility of installed GPUs, but the current hardware or infrastructure issue is a physical or driver-level enumeration fault affects one accelerator. The option “check GPU seating device power and driver health before re-adding node to scheduling” is more directly tied to the required outcome.

Useful for full installed-GPU visibility does not prove healthy load behavior, but it does not remove the specific condition: a physical or driver-level enumeration fault affects one accelerator.

“validate installed CUDA runtime and application compatibility against the driver” addresses driver visibility was mistaken for full application-runtime readiness, but the current hardware or infrastructure issue is a physical or driver-level enumeration fault affects one accelerator. The option “check GPU seating device power and driver health before re-adding node to scheduling” is more directly tied to the required outcome.

“an inventory sample maps every reported GPU ID to the expected physical slot” addresses accelerator identity lacks a stable mapping to the physical PCIe slot, but the current hardware or infrastructure issue is a physical or driver-level enumeration fault affects one accelerator. The option “nvidia-smi and platform logs show all eight devices healthy” is more directly tied to the required outcome.

“sustained workload metrics show no unexpected GPU throttling flags” addresses full installed-GPU visibility does not prove healthy load behavior, but the current hardware or infrastructure issue is a physical or driver-level enumeration fault affects one accelerator. The option “nvidia-smi and platform logs show all eight devices healthy” is more directly tied to the required outcome.

“repeated short-job startup measurements show reduced initialization overhead” addresses device persistence settings do not match the intended repeated-job operating profile, but the current hardware or infrastructure issue is a physical or driver-level enumeration fault affects one accelerator. The option “nvidia-smi and platform logs show all eight devices healthy” is more directly tied to the required outcome.

Useful for scheduling readiness did not confirm required device memory and capabilities, but it does not remove the specific condition: an idle utilization sample does not test compute functionality.

“investigate GPU throttle reasons and restore thermal or power headroom” addresses full installed-GPU visibility does not prove healthy load behavior, but the current hardware or infrastructure issue is an idle utilization sample does not test compute functionality. The option “run a bounded representative CUDA workload while monitoring nvidia-smi” is more directly tied to the required outcome.

“correlate nvidia-smi PCI bus identifiers with chassis service labels” addresses accelerator identity lacks a stable mapping to the physical PCIe slot, but the current hardware or infrastructure issue is an idle utilization sample does not test compute functionality. The option “run a bounded representative CUDA workload while monitoring nvidia-smi” is more directly tied to the required outcome.

“device inventory meets requirements and the test allocates memory without failure” addresses scheduling readiness did not confirm required device memory and capabilities, but the current hardware or infrastructure issue is an idle utilization sample does not test compute functionality. The option “utilization rises during the job and diagnostic compute completes successfully” is more directly tied to the required outcome.

“topology display and local bandwidth checks match the approved HGX design” addresses interconnect readiness is not captured in GPU bring-up acceptance, but the current hardware or infrastructure issue is an idle utilization sample does not test compute functionality. The option “utilization rises during the job and diagnostic compute completes successfully” is more directly tied to the required outcome.

“a sample CUDA application initializes and reports the expected device” addresses driver visibility was mistaken for full application-runtime readiness, but the current hardware or infrastructure issue is an idle utilization sample does not test compute functionality. The option “utilization rises during the job and diagnostic compute completes successfully” is more directly tied to the required outcome.

“verify the supported driver kernel modules and device enumeration” addresses host driver registration has not established visibility of installed GPUs, but the current hardware or infrastructure issue is device persistence settings do not match the intended repeated-job operating profile. The option “configure supported GPU persistence behavior for the environment” is more directly tied to the required outcome.

Useful for interconnect readiness is not captured in GPU bring-up acceptance, but it does not remove the specific condition: device persistence settings do not match the intended repeated-job operating profile.

“inspect GPU memory and compute capability and run the approved workload smoke test” addresses scheduling readiness did not confirm required device memory and capabilities, but the current hardware or infrastructure issue is device persistence settings do not match the intended repeated-job operating profile. The option “configure supported GPU persistence behavior for the environment” is more directly tied to the required outcome.

“sustained workload metrics show no unexpected GPU throttling flags” addresses full installed-GPU visibility does not prove healthy load behavior, but the current hardware or infrastructure issue is device persistence settings do not match the intended repeated-job operating profile. The option “repeated short-job startup measurements show reduced initialization overhead” is more directly tied to the required outcome.

“nvidia-smi lists the intended GPU count with no driver initialization errors” addresses host driver registration has not established visibility of installed GPUs, but the current hardware or infrastructure issue is device persistence settings do not match the intended repeated-job operating profile. The option “repeated short-job startup measurements show reduced initialization overhead” is more directly tied to the required outcome.

“nvidia-smi and platform logs show all eight devices healthy” addresses a physical or driver-level enumeration fault affects one accelerator, but the current hardware or infrastructure issue is device persistence settings do not match the intended repeated-job operating profile. The option “repeated short-job startup measurements show reduced initialization overhead” is more directly tied to the required outcome.