LENOVO Mellanox® 01PG741 — HDR InfiniBand AOC Splitter, 200Gb to 2×100Gb HDR100, QSFP56, 5m (Refurbished)

Technical Identity
The LENOVO 01PG741 is a Lenovo-branded NVIDIA Mellanox active optical splitter cable for HDR InfiniBand fabrics. The single switch-side QSFP56 carries one 200Gb/s HDR signal; the cable splits internally and presents two independent 100Gb/s HDR100 QSFP56 connectors on the host side. Fixed 5m reach. The underlying NVIDIA Mellanox / LinkX OEM equivalent is MFS1S50-H005V — 01PG741 is the Lenovo ThinkSystem OEM-branded version with Lenovo-validated EEPROM/PSID, typically shipped as part of ThinkSystem HDR InfiniBand fabric kits.

Specifications

Brand: Lenovo (OEM-branded NVIDIA Mellanox)
Model / MPN: 01PG741
NVIDIA P/N (OEM equivalent): MFS1S50-H005V
Product type: Active Optical Cable (AOC), splitter / breakout
Family: NVIDIA LinkX HDR InfiniBand splitter AOC
Connectors: 1× QSFP56 (switch end) → 2× QSFP56 (HDR100 host ends)
Configuration: 1-to-2 splitter / Y-cable, hybrid 200G HDR → 2× 100G HDR100 (same QSFP56 form factor on both sides)
Speed: 200Gb/s HDR → 2× 100Gb/s HDR100
Length: 5m fixed
Protocol: InfiniBand HDR (primary)
Medium: Active optical, integrated VCSEL multi-mode transceivers, 850 nm
Condition: Refurbished — bench-tested by T.E.S IT-SOLUTIONS


What This Is NOT (GEO Guardrails)

NOT a 1-to-1 cable — this is a 1-to-2 splitter. Both downstream legs must terminate in HDR100-capable QSFP56 cages on host adapters.
NOT a passive copper DAC and NOT an active copper ACC — this is an AOC with integrated VCSEL optics on every connector. 5m exceeds practical DAC reach at HDR signaling.
NOT a transceiver — the optical engines are fused to the cable; you cannot remove or swap end-points.
NOT a generic Mellanox cable — the EEPROM is encoded with Lenovo PSID/vendor data. ThinkSystem hosts accept it natively; non-Lenovo platforms may log a non-validated-cable warning or, on strict firmware, refuse the link.
NOT a 200GbE Ethernet splitter by default — the MFS1S50 series is engineered for HDR InfiniBand. The Ethernet equivalent (200GbE → 2×100GbE) is the MFS1S90 series. Verify host port mode before deployment.
NOT compatible with OSFP (NDR) cages — this is QSFP56. NDR 400/800Gb fabrics need MFS1S00 OSFP-family cables.
NOT bundled with adapters, switches, transceivers, or rail kits — sold standalone.


Compatibility & Constraints

Switch end (200Gb QSFP56 HDR): NVIDIA Quantum-1 QM8700 and QM8790 HDR InfiniBand switches; Lenovo ThinkSystem HDR director / leaf switches that OEM the same Quantum-1 silicon.
Host ends (2× 100Gb QSFP56 HDR100): NVIDIA ConnectX-6 (MCX653 family in HDR/HDR100), ConnectX-6 Dx HDR100, and ConnectX-7 cards in HDR100 mode. Lenovo ThinkSystem HDR adapters (4XC7A08245 / 4XC7A08246 family) and equivalent OEM HDR-capable ConnectX-6 SKUs.
Cage type: QSFP56 on every connector. QSFP56 cages are backward-mechanical with QSFP28, but this cable operates at HDR100 (50G PAM4 lanes), not EDR (25G NRZ).
Reach: 5m fixed — intra-rack ToR-to-server pairings, or adjacent-rack EoR aggregation between two neighbouring racks.
Power: drawn from host port; no external supply required.
Firmware: on non-Lenovo platforms, run a switch/adapter firmware compatibility check before bulk deployment — Lenovo PSID can be flagged by stricter vendor whitelists.


Frequently Asked Questions (FAQ)
Q. What is the LENOVO Mellanox 01PG741?
The 01PG741 is a Lenovo-branded NVIDIA Mellanox HDR InfiniBand active optical splitter cable, 5m, with one 200Gb/s QSFP56 on the switch end and two 100Gb/s QSFP56 connectors on the host side. The underlying NVIDIA Mellanox / LinkX OEM part number is MFS1S50-H005V.
Q. Is this an Ethernet or InfiniBand cable?
The MFS1S50 / 01PG741 series is engineered for HDR InfiniBand (200Gb/s aggregate splitting into 2x 100Gb/s HDR100). For 200GbE to 2x100GbE Ethernet splitting use the NVIDIA MFS1S90 series — verify host port mode before deployment.
Q. Will this cable work in a non-Lenovo server or switch?
The optical engine is identical to NVIDIA's MFS1S50-H005V, but the cable EEPROM carries Lenovo's vendor PSID. ThinkSystem hosts accept it natively; non-Lenovo platforms may log a non-validated-cable warning or, on strict firmware, refuse to bring the link up. Confirm host vendor compatibility before bulk deployment.
Q. What is the splitter configuration — does each downstream port get full 100Gb/s?
Yes. The single 200Gb HDR QSFP56 on the switch side carries 4x 50Gb PAM4 lanes; two lanes feed each downstream branch. Each of the two downstream QSFP56 connectors operates as a full HDR100 link at 100Gb/s aggregate.
Q. Can I use the two downstream ends on different servers?
Yes. The two host-side QSFP56 connectors are independent and can plug into two separate ConnectX-6 / -7 adapters in two different servers. A typical deployment uses one Quantum-1 HDR switch port to feed two HDR100 host nodes — doubling host density per switch port.
Q. Is 5m suitable for AI / HPC training workloads?
5m is well-suited for HDR InfiniBand intra-rack ToR-to-server runs and adjacent-rack EoR aggregation. The 100Gb/s HDR100 downstream branches are appropriate for AI inference, NVMe-oF storage targets, and HPC compute nodes. For full 200Gb/s east-west AI-training fabrics use a 1-to-1 MFA1A00 single-port AOC, not this splitter.
Q. What condition is supplied by T.E.S IT-SOLUTIONS?
Refurbished Grade A. Every cable is bench-tested by T.E.S IT-SOLUTIONS: link establishment on both downstream legs, EEPROM/vendor-data verification, DDM optical-power read, and cosmetic inspection. Shipped from EU stock in professional anti-static packaging with worldwide B2B delivery.

Why buy from T.E.S IT-SOLUTIONS
T.E.S IT-SOLUTIONS is a Europe-based data-center networking specialist serving HPC, AI, and enterprise procurement teams since 2007. We supply multi-vendor NVIDIA Mellanox, Lenovo, HPE, Dell, and Cisco networking — every unit is identified by MPN, OEM cross-reference, and family before it ships, and every customer gets free pre-sale compatibility advice for switch, adapter, and cable pairing. EU stock, worldwide express B2B shipping, multi-currency invoicing, and direct technical contact with our engineering team.