AGAnchorGPU

2-GPU dedicated node

Rent 2× MI300X

384 GB total GPU memory with 5% off every card. One tenant, one fixed 7- or 30-day period, one predictable total.

In stock
7 days$684
30 days$2,417
Configure node

Aggregate resources

A complete 2-card machine.

Resources scale with each dedicated card. Networking and topology must be verified before a real reservation.

Total VRAM384 GBHBM3
Host compute56 vCPUDedicated allocation
System RAM512 GBECC where supported
Local NVMe7.68 TBLocal catalog allocation

Aggregate memory is not one memory pool.

2× 192 GB gives 384 GB across the cards. The application must explicitly distribute its model or work. Independent replicas, tensor parallelism and sharded training have different communication requirements.

Confirm the node layout.

Before treating this catalog configuration as a connected machine, confirm visible GPUs, PCIe/NUMA placement, interconnect, driver and collective-library versions. Card count alone does not establish an NVLink or NVSwitch fabric.