NVIDIA Mellanox MCX512A-ACAT ConnectX®-5 EN Network
Interface Card, 10/25GbE Dual-Port SFP28, PCIe 3.0 x8, Tall Bracket
ConnectX-5 MCX512A-ACAT Ethernet network interface card
provides high performance and flexible solutions with up to two ports of 25GbE
connectivity, 750ns latency, up to 75 million messages per second (Mpps). For
storage workloads, the 25 Gigabit Ethernet adapter card delivers a range of
innovative accelerations, such as Signature Handover (T10-DIF) in hardware, an
embedded PCIe Switch, and NVMe over Fabric target offloads.
Product Highlights
- PCIe 3.0 x 8 Host Interface and 25GbE Dual-Port
Transmission
- Industry Leading RoCE v1/v2 delivers high throughput
and low-latency connectivity
- Efficient I/O consolidation, lowering data center costs
and complexity
- Advanced storage capabilities including NVMe over
Fabric offloads
- Cutting-edge performance in virtualized networks including NFV
- Support Windows, RHEL/CentOS, VMware, FreeBSD, OFED and
WinOF-2
Specifications
Manufacturer
MELLANOX TECHNOLOGIES INC
Controller
Mellanox Technologies MT28800 Family ConnectX-5
Ports
Dual
Technology
Ethernet
Supported Ethernet Speeds
25/10/1GbE
Interface Type
SFP28
Host Interface
PCIe 3.0 x8
Link Rate
8.0 GT/s
OS Support
Windows, RHEL/CentOS, VMware, FreeBSD, OFED and WinOF-2
Cabling Type
SFP28 Transceiver, DAC and AOC
RDMA
Yes
RoHS R6 Compliant
Yes
Controller | ConnectX®-4 Lx | ConnectX®-5 | ConnectX®-5 | ConnectX®-5 | ConnectX®-6 Dx |
Ports | Dual-Port | Dual-Port | Single-Port | Dual-Port | Dual-Port |
Supported Ethernet Speeds | 25/10/1GbE | 25/10/1GbE | 100/50/40/ 25/10/1GbE | 100/50/40/ 25/10/1GbE | 100/50/40 25/10/1GbE |
Systems Interface Type | PCIe 3.0 x8 (8.0 GT/s) | PCIe 3.0 x8 (8.0 GT/s) | PCIe 3.0 x16 (8.0 GT/s) | PCIe 3.0 x16 (8.0 GT/s) | PCIe 4.0 x16 (16.0 GT/s) |
RDMA over Converged Ethernet (RoCE) | |||||
Zero Touch RoCE (ZTR) | |||||
ASAP2 (Virtual Switch offload) | — | ||||
SR-IOV | |||||
Hardware-based I/O Virtualization | |||||
NVMe-oF Target Offload | — | ||||
Burst Buffer Offloads | — | ||||
Hardware Offload of VXLAN, NVGRE, and GENEVE Overlay Networks | — |