Help Me Choose: Memory

PowerEdge 13th-generation servers are very modular by nature, and you can now choose the exact quantity and type of DIMMs for your system.

To configure your system’s memory for optimal performance, keep these best practices in mind:

1. Populate all memory channels evenly with the same rank, size, and speed of DIMMs. For situations that demand mixing DIMM sizes, all memory should be chosen based on identical rank structure, and all memory channels should be populated with an identical mix of DIMM types.
2. Choose the appropriate processor to support the expected memory frequency. For example, if your workload requires a memory speed of 2133MT/s, select a processor model that supports this speed.
3. For most system configurations, 1 DIMM per channel (DPC) and 2 DPC memory configurations perform better than 3 DPC because that memory frequency will be reduced when all slots in a given memory channel are populated.
4. Choose the Optimizer memory operating mode, which is part of the system profile, instead of Advanced ECC due to the significant performance penalties implicit in running the memory in this RAS mode.
5. Choose LRDIMMs when the greatest capacity and performance of memory is desired. Choose RDIMMs for a balanced approach to memory capacity, performance and cost. Choose UDIMMs when using less power is a premium over capacity.
Memory Configuration Type
Reduce data loss and keep your system online and operational with reliability, availability and serviceability (RAS) features for Dell servers and storage, including:

High-reliability profile, which helps increase memory reliability through a selectable profile that adjusts parameters to reduce faults associated with speed, temperature, voltage and refresh rates. High reliability profile has a minor impact on power consumption and performance.

Memory mirroring has two adjacent memory channels configured to write the same data to each channel. If one memory channel fails or encounters an error, the other channel continues to transmit data. It’s an excellent safeguard for systems requiring uninterrupted operation, though it cuts memory capacity in half, doubles the cost per gigabyte, and can increase power consumption.

Memory sparing can reduce downtime from correctable errors by allocating one rank (64-bit wide data area on a DIMM) per channel as a spare. If a correctable error occurs in a rank or channel, it’s moved to the spare rank while the OS is running. This prevents the error from causing a failure. Memory sparing reduces memory capacity by one rank per channel and increases the cost per gigabyte.

Choose from two different dual in-line memory module types or Dimm types that can meet your needs depending on how your prioritize R.A.S. (reliability, availability, and serviceability) and power consumption.

Dell is supporting three DIMM types on the twelfth-generation servers:
  • RDIMM: Registered DIMM – Provides for higher capacity options and advanced RAS features.
  • LRDIMM: Load Reduced DIMM – Provides maximum capacity but higher power consumption.

RDIMM, or registered memory, is the most commonly used DIMM type, and offers the best mix of frequency, capacity, and rank structure choices. It provides high signal integrity — performing parity checking to detect improper addresses or commands — and increased performance for heavy workloads.

(Single Rank and Dual Rank)

  • Maximum frequency of 2133 MT/s
  • Maximum frequency using 3 DIMMs per channel of 1866 MT/s
  • Maximum capacity of 16GB per DIMM
  • Maximum system capacity of 384 GB

, or load-reduced memory, uses a buffer to reduce memory loading to a single load on all DDR signals, allowing for greater density.


  • Maximum frequency of 2133MT/s
  • Maximum frequency using 3 DIMMs per channel of 1866MT/s
  • Maximum capacity of 32GB per DIMM
  • Maximum system capacity of 768GB

DIMM Speed

There is only one speed option for DDR4
  • 2133MT/s – DIMMS will run at 1.2 volt

1866MT/s – DIMMS will run at 1.5 volt and are supported with Intel® Xeon® 2600 v2 product family only

RAS Feature

RAS feature options offered on the twelfth-generation servers:
  • Memory Optimized – Baseline RAS features for highest performance optimization
  • Advanced ECC - RAS Feature that extends multi-bit single device error correction to x8 DRAMs
  • Memory Sparing- RAS feature that allocates one rank per channel as a spare to reduce the likelihood of correctable errors from becoming uncorrectable errors but available memory capacity is reduced
  • Memory Mirroring – RAS Feature with paired memory provides data and transmission fault-tolerance but available memory capacity is reduced by 50%
  • Dell Fault Resilient Memory (FRM) - This Dell-patented technology enables PowerEdge 12th generation server customers supporting VMware vSphere 5.5 to maximize available server memory while increasing protection against memory faults for the hypervisor and mission critical applications.
For more information on memory, please refer to our memory page at 

The table below lists the comparison between different DIMM types.

DIMM Comparison

Registered/Buffered Address only Address and data
Frequency/Speed 2133 MT/s 2133 MT/s
Ranks Supported 1 or 24 only
Capacity per DIMM 4, 8 and 16 GB 32GB
Maximum DIMM per Channel 3 3
Dynamic RAM (DRAM) Technology x4 and x8 x4
Temperature Sensor YesYes
Error-Correction Code (ECC) YesYes
Single Device Data Correction (SDDC) Yes
(x8 requires Advanced ECC mode)
Advanced ECCYesYes
Address Parity YesYes

For more information on memory, please refer to our memory page at

DIMM Speed and Frequency

The memory offerings for Dell PowerEdge 13th-generation servers are based on the double-data rate type 4 (DDR4) which operates at 1.2 volts.

PowerEdge memory speeds by type and loading

DIMM Type DIMM Ranking
DIMM Rated Voltage, Speed12th Generation Memory (EP, 3DPC Platforms) *DPC Means DIMMs per channel
RDIMM1R/2R4GB, 8GB, 16GB DDR4 (1.2V), 2133213321331866
4R32GB DDR4 (1.2V), 2133 213321331866

For more information on memory, please refer to our memory page at