Hardware Management Console
The HMC is the console an IBM i shop uses to create the partition an AI model runs in, then hand that partition processors, memory and the PCIe slot holding an accelerator card while production keeps running. IBM Power 11 servers need HMC release V11R1M1110 or later.
What does the HMC actually do?
Picture your Power server as an apartment building. The HMC is the manager's office. It decides how many apartments exist, how big each one is, and which apartment gets the parking space that holds the AI card. IBM's Power S1112 technical overview calls it the central point of control for one or more managed systems.
- Create, change and delete logical partitions, which IBM calls LPARs.
- Move processors, memory and card slots between partitions.
- Apply firmware updates without taking the server down.
- Manage Capacity on Demand, which switches on hardware you already own.
- Power the server on and off, collect dumps, and call IBM for service.
Do I need one just to run an AI partition?
A new Power server arrives in what IBM calls the manufacturing default configuration. That is one partition using the whole machine. IBM's Power 11 documentation says you do not have to manage the system with an HMC in that state. The moment you use the HMC to create or change a partition, the system is in partition mode, and from then on you must use the HMC to manage it. Running an AI model in its own Linux partition beside IBM i means carving the machine up, so that is the moment the console stops being optional.
That wording comes from the documentation for one Power 11 scale-out machine type. Check the documentation for your exact model before you plan a purchase.
How does an AI card get into a running partition?
Handing a PCIe accelerator card to a live partition
- The card goes in a PCIe slotIBM Research describes the Spyre Accelerator as a single slot PCIe card that needs no extra power cable.
- The HMC adds the slot to the partitionIBM's Power 11 documentation says the HMC can dynamically add a physical I/O slot, plus the adapter and devices attached to it, to a logical partition that is already running.
- The Linux side has to be readyThe partition needs a Linux release that supports dynamic partitioning, with the DynamicRM tool package installed inside it.
IBM lists this task for HMC Version 10.2.1030 or later and says it cannot be used with partitions that use shared memory.
A 7063-CR2 box or a virtual HMC?
The appliance and the virtual console side by side
| Item | 7063-CR2 appliance | Virtual HMC |
|---|---|---|
| Processor | 6-core Power 9 processor, 130 W | Four virtual processors |
| Memory | 64 GB or 128 GB | 16 GB |
| Storage | 1.8 TB, mirrored with RAID 1 | One disk with 500 GB free |
| Runs on | Its own rack box, 1.7 inches tall, 17.6 kg | x86 KVM, Xen or VMware, or a PowerVM partition |
IBM recommends two consoles for availability. That can be two boxes, one box and one virtual console, or two virtual consoles, and both must run the same version level.
Which HMC release does Power 11 need?
IBM's Power S1112 technical overview gives one floor: release V11R1M1110, generally available in July 2025, is the lowest level that supports the Power 11 platform, including the Power S1122 and the Power S1124. Power 11 systems do not support the older 7063-CR1 console at all, so an aging appliance is a replacement rather than an upgrade. IBM has since shipped V11 R2 M1120 for the 7063-CR2 and for virtual consoles, and that media upgrades from a minimum level of V10 R3.
Related Terms
Dynamic Logical Partitioning
Dynamic logical partitioning, or DLPAR, moves processors, memory and card slots in and out of a running IBM Power partition without restarting it. IBM documents it for the Spyre Accelerator, so an AI card can join or leave a Linux inference partition while the IBM i production partition keeps working.
AI Inference Partition
A separate supported operating-system partition used to execute an AI model while an IBM i partition may remain the production application and data source.
AI Workload Placement
The decision about which system, operating environment, and partition executes an AI model or agent, separately from the IBM i application supplying its data.
Spyre Accelerator
IBM's PCIe AI inference card. On IBM Power 11 it carries 32 cores, 128 GB of memory, and more than 300 TOPS inside a 75 watt envelope.
IBM Power S1124
The Power S1124 is the largest scale-out IBM Power 11 server, machine type 9824-42A, and IBM lists it for 8, 12, 16 or 24 Spyre AI accelerator cards ... but only when both processor modules are installed. An IBM i shop can keep production on IBM i 7.4 TR12 or later and run model inference in a Red Hat Enterprise Linux partition on the same 4U machine.