Specifications
Module Type
MRDIMM (Multiplexed Rank DIMM / MCRDIMM)
Capacity
128 GB
Data Rate
8800 MT/s
Platform
Intel Xeon 6 processors
Bandwidth Gain
Up to 39% over DDR5 RDIMM
Latency
Up to 40% lower than RDIMM
Bus Efficiency
+15% versus RDIMM
On-Module Logic
MRCD register clock driver + MDB data buffer chips
Rank Configuration
Multiple DDR5 ranks combined into one module
Capacity Range
32 GB / 64 GB / 96 GB / 128 GB / 256 GB
Power Profile
Predictable module power; on-module devices run at half the host rate
DIMM Slots
Each DIMM carries the multiplexing overhead of a full RDIMM slot
Use Case
Bandwidth-sensitive AI inference and HPC nodes
Generation
First-generation MRDIMM, 8800 MT/s
Overview
The 128 GB DDR5 MRDIMM sits in the middle of Micron's MRDIMM capacity range and is the most commonly deployed capacity point for Xeon 6 AI nodes — enough capacity per module to populate a 12-channel server densely without paying for the 256 GB top SKU.
Like every MRDIMM, it carries an MRCD multiplexer register clock driver and MDB data buffers on the module. The host interface is driven at 8800 MT/s while the DRAM ranks themselves run at half rate, which is what keeps module power inside a predictable budget while still producing up to 39% more bandwidth than an equivalent RDIMM.
Because MRDIMM multiplexes ranks rather than overclocking the DRAM, the latency penalty that usually accompanies higher memory clock rates does not appear — Micron quotes up to 40% lower latency than RDIMM, which matters for the irregular access patterns in AI inference.
Key Benefits
128 GB per module — the mainstream MRDIMM capacity for dense Xeon 6 servers. Full MRDIMM bandwidth gain — 39% over RDIMM at the same slot count. 15% better bus efficiency versus RDIMM.
Applications
Intel Xeon 6 AI inference servers, memory-bandwidth-bound HPC applications, virtualisation hosts, and scale-out database nodes that need more bandwidth per core than RDIMM can deliver.
Request a Quote — MICRON 128GB DDR5 MRDIMM 8800 MT/S — HIGH-BANDWIDTH SERVER MEMORY
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →