We take cryptocurrency — so the coins you hold can buy real hardware.Pay in crypto, spend it on real hardware. Free insured shipping worldwide over $399. Plain, unbranded boxes, sent from Germany.Free insured shipping over $399. Your rate is locked for 30 minutes once checkout opens.Rate locked 30 minutes. Every item sealed, serial-checked and covered by a 24-month warranty.Sealed, 24-month warranty.

Saved items Sign in

SMUGDESK

SMUGDESK Mellanox ConnectX-3 VPI Dual-Port QSFP+ 40/56GbE NIC network card

SKU CH-F43A-JMYXZ

Dual-port QSFP+ adapter delivers 40/56GbE Ethernet and FDR InfiniBand connectivity over PCIe 3.0 x8 with full hardware offload and RoCE RDMA

  • Dual-protocol ports auto-switch between 40/56GbE and FDR InfiniBand for flexible interconnect
  • PCIe 3.0 x8 bus provides high throughput with backward compatibility to 2.0/1.1
  • Full hardware offload including RoCE v1/v2 RDMA slashes host CPU load under heavy traffic
  • SR-IOV virtualization plus jumbo frames up to 9KB and PXE boot for server deployments
  • Broad OS support across Windows Server, Linux, VMware ESXi and FreeBSD for HPC and storage clusters

Dual-protocol QSFP+ NIC for HPC clusters

The SMUGDESK Mellanox ConnectX-3 VPI adapter provides two QSFP+ ports that auto-negotiate between 10, 40 and 56GbE Ethernet or FDR InfiniBand. It targets administrators building storage clusters, virtualization hosts or low-latency trading platforms who need a single card for both fabric types.

Hardware offload and SR-IOV reduce host CPU load

Full TCP/UDP checksum, TSO, LSO and VXLAN offload plus RoCE v1 and v2 RDMA move data-plane work onto the adapter. SR-IOV and 9KB jumbo frames let hypervisors present near-native throughput to guest workloads without software bottlenecks.

PCIe 3.0 x8 install with dual brackets supplied

The card fits a standard PCIe 3.0 x8 slot and falls back to 2.0 or 1.1 speeds if required. Both full-height and low-profile brackets are included so the same unit fits 1U or full-height chassis. Buyers who need SFP+ cages or single-port density should select a different model.

Highlights

  • Dual-protocol ports auto-switch between 40/56GbE and FDR InfiniBand for flexible interconnect
  • PCIe 3.0 x8 bus provides high throughput with backward compatibility to 2.0/1.1
  • Full hardware offload including RoCE v1/v2 RDMA slashes host CPU load under heavy traffic
  • SR-IOV virtualization plus jumbo frames up to 9KB and PXE boot for server deployments
  • Broad OS support across Windows Server, Linux, VMware ESXi and FreeBSD for HPC and storage clusters

Specifications

No published specifications for this item yet.

Questions about this item

Will the QSFP+ transceivers or DAC cables need periodic replacement?

The dual QSFP+ ports accept standard DAC copper or fiber modules; the card itself has no moving parts or consumable components that require scheduled replacement.

What limits sustained throughput on the PCIe 3.0 x8 bus?

The 8 GT/s PCIe 3.0 x8 interface provides the ceiling for combined 40/56 GbE traffic; once aggregate bandwidth approaches that link rate the bus becomes the bottleneck.

How much heat does the adapter produce when running both ports at 56 GbE all day?

No power-draw or thermal figures are supplied for continuous dual-port 56 GbE operation.

Also in this aisle

People compared these

Same shelf, same checkout — eight coins and a 30-minute rate lock.