Product Overview
The new ANNA Ampere L1 server features the power of NVIDIA A100 Tensor Core GPUs and the HGX A100 4-GPU baseboard. The system supports PCI-E Gen 4 for fast CPU-GPU connection and high-speed networking expansion cards.
The new ANNA Ampere L1 server features the power of NVIDIA A100 Tensor Core GPUs and the HGX A100 4-GPU baseboard. The system supports PCI-E Gen 4 for fast CPU-GPU connection and high-speed networking expansion cards.
With GPU Direct RDMA and 1 to 1 mapping between network interconnects and GPU. It uses NVLink to provide GPU to GPU communications in a mesh system at speeds up to 200GB per second.
With 1 + 1 power redundancy it makes the system ideal for HPC and AI workloads. Designed with Speech Recognition, Computer vision and Data science in mind, the ANNA Ampere L1 would be the perfect choice.
Certified by Nvidia for use with Nvidia AI Enterprise suite and with access to NGC.
Key Features
2U Rackmount Chassis Dual AMD EPYC™ 7003/7002 Series Processor Up to 8TB Registered ECC DDR4 3200MHz SDRAM in 32 DIMMs NVIDIA® 4-GPU HGX A100 (SXM4) with NVLink 4 front Hot-swap 2.5" drive bays
-
Chipset
System-On-Chip (SoC)
-
Drive Bays
4 x 2.5" Hot-swap drive bays
-
Drive Support
SATA/NVMe Hybrid or SAS with optional HBA
-
Expansion Slots
1 PCIe 4.0 x8 LP
4 x PCI-E Gen 4 x16 (LP) slots
-
Form Factor
2U Rackmount
-
GPU Manufacturer
NVIDIA
-
GPU Quantity
NVIDIA 4-GPU HGX-A100 (80GB) - SXM4
-
Manufacturer
Supermicro
-
Memory (Maximum)
Up to 8TB 3DS ECC DDR4-3200MH SDRAM
-
Memory Slots
32 DIMM Slots
-
Memory Type
3200MHz ECC DDR4 SDRAM
-
Network Connectivity
Dual RJ45 10GbE-aggregate host LAN
-
Power Supply
2200W Redundant Power Supplies with PMBus (Platinum)